Compare commits

..

1 Commits

Author SHA1 Message Date
98df04527d WIP 2023-11-14 21:12:13 -08:00
114 changed files with 2876 additions and 21410 deletions

View File

@@ -1,9 +0,0 @@
[build]
rustflags = [ "--cfg=web_sys_unstable_apis" ]
[registry]
global-credential-providers = ["cargo:token"]
[registries.xinu]
index = "sparse+https://git.z.xinu.tv/api/packages/wathiede/cargo/"

10
.envrc
View File

@@ -1,10 +0,0 @@
source_up
export DATABASE_USER="newsreader";
export DATABASE_NAME="newsreader";
export DATABASE_HOST="nixos-07.h.xinu.tv";
export DATABASE_URL="postgres://${DATABASE_USER}@${DATABASE_HOST}/${DATABASE_NAME}";
export PROD_DATABASE_USER="newsreader";
export PROD_DATABASE_NAME="newsreader";
export PROD_DATABASE_HOST="postgres.h.xinu.tv";
export PROD_DATABASE_URL="postgres://${PROD_DATABASE_USER}@${PROD_DATABASE_HOST}/${PROD_DATABASE_NAME}";

View File

@@ -1,67 +0,0 @@
on: [push]
name: Continuous integration
jobs:
check:
name: Check
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: actions-rust-lang/setup-rust-toolchain@v1
- run: cargo check
test:
name: Test Suite
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: actions-rust-lang/setup-rust-toolchain@v1
- run: cargo test
trunk:
name: Trunk
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: actions-rust-lang/setup-rust-toolchain@v1
with:
toolchain: nightly
target: wasm32-unknown-unknown
- run: cargo install trunk
- run: cd web; trunk build
fmt:
name: Rustfmt
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: actions-rust-lang/setup-rust-toolchain@v1
with:
components: rustfmt
- name: Rustfmt Check
uses: actions-rust-lang/rustfmt@v1
build:
name: build
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: actions-rust-lang/setup-rust-toolchain@v1
- run: cargo build
udeps:
name: Disallow unused dependencies
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: actions-rust-lang/setup-rust-toolchain@v1
with:
toolchain: nightly
- name: Run cargo-udeps
uses: aig787/cargo-udeps-action@v1
with:
version: 'latest'
args: '--all-targets'

7267
Cargo.lock generated

File diff suppressed because it is too large Load Diff

View File

@@ -1,18 +1,12 @@
[workspace]
resolver = "2"
default-members = ["server"]
members = ["web", "server", "notmuch", "procmail2notmuch", "shared"]
[workspace.package]
authors = ["Bill Thiede <git@xinu.tv>"]
edition = "2021"
license = "UNLICENSED"
publish = ["xinu"]
version = "0.17.32"
repository = "https://git.z.xinu.tv/wathiede/letterbox"
[profile.dev]
opt-level = 1
members = [
"web",
"server",
"notmuch",
"procmail2notmuch",
"shared"
]
[profile.release]
lto = true

View File

@@ -1,19 +0,0 @@
export CARGO_INCREMENTAL := "0"
export RUSTFLAGS := "-D warnings"
default:
@echo "Run: just patch|minor|major"
major: (_release "major")
minor: (_release "minor")
patch: (_release "patch")
sqlx-prepare:
cd server; cargo sqlx prepare && git add .sqlx; git commit -m "cargo sqlx prepare" .sqlx || true
pull:
git pull
_release level: pull sqlx-prepare
cargo-release release -x {{ level }} --workspace --no-confirm --registry=xinu

View File

@@ -1,7 +0,0 @@
.PHONEY: release
release:
(cd server; cargo sqlx prepare && git add .sqlx; git commit -m "cargo sqlx prepare" .sqlx || true)
bash scripts/update-crate-version.sh
git push
all: release

4
dev.sh
View File

@@ -1,7 +1,7 @@
cd -- "$( dirname -- "${BASH_SOURCE[0]}" )"
tmux new-session -d -s letterbox-dev
tmux rename-window web
tmux send-keys "cd web; trunk serve -w ../.git -w ../shared -w ../notmuch -w ./" C-m
tmux send-keys "cd web; trunk serve -w ../shared -w ../notmuch -w ./" C-m
tmux new-window -n server
tmux send-keys "cd server; cargo watch -c -w ../.git -w ../shared -w ../notmuch -w ./ -x 'run postgres://newsreader@nixos-07.h.xinu.tv/newsreader ../target/database/newsreader /tmp/letterbox/slurp'" C-m
tmux send-keys "cd server; cargo watch -x run -w ../shared -w ../notmuch -w ./" C-m
tmux attach -d -t letterbox-dev

View File

@@ -1,24 +1,19 @@
[package]
name = "letterbox-notmuch"
exclude = ["/testdata"]
description = "Wrapper for calling notmuch cli"
authors.workspace = true
edition.workspace = true
license.workspace = true
publish.workspace = true
repository.workspace = true
version.workspace = true
name = "notmuch"
version = "0.1.0"
edition = "2021"
# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
[dependencies]
log = "0.4.27"
mailparse = "0.16.1"
anyhow = "1.0.75"
clap = { version = "4.4.7", features = ["derive"] }
log = "0.4.14"
serde = { version = "1.0", features = ["derive"] }
serde_json = { version = "1.0", features = ["unbounded_depth"] }
thiserror = "2.0.12"
tracing = "0.1.41"
thiserror = "1.0.30"
[dev-dependencies]
itertools = "0.14.0"
itertools = "0.10.1"
pretty_assertions = "1"
rayon = "1.10"
rayon = "1.5"

View File

@@ -207,15 +207,14 @@
//! ```
use std::{
collections::HashMap,
ffi::OsStr,
io::{self},
io::{self, BufRead, BufReader, Lines},
path::{Path, PathBuf},
process::Command,
process::{Child, ChildStdout, Command, Stdio},
};
use log::info;
use serde::{Deserialize, Serialize};
use tracing::{error, info, instrument, warn};
/// # Number of seconds since the Epoch
pub type UnixTime = isize;
@@ -270,12 +269,6 @@ pub struct Headers {
#[serde(skip_serializing_if = "Option::is_none")]
pub bcc: Option<String>,
#[serde(skip_serializing_if = "Option::is_none")]
#[serde(alias = "Delivered-To")]
pub delivered_to: Option<String>,
#[serde(skip_serializing_if = "Option::is_none")]
#[serde(alias = "X-Original-To")]
pub x_original_to: Option<String>,
#[serde(skip_serializing_if = "Option::is_none")]
pub reply_to: Option<String>,
pub date: String,
}
@@ -465,17 +458,13 @@ pub enum NotmuchError {
StringUtf8Error(#[from] std::string::FromUtf8Error),
#[error("failed to parse str as int")]
ParseIntError(#[from] std::num::ParseIntError),
#[error("failed to parse mail: {0}")]
MailParseError(#[from] mailparse::MailParseError),
}
#[derive(Clone, Default)]
#[derive(Default)]
pub struct Notmuch {
config_path: Option<PathBuf>,
}
// TODO: rewrite to use tokio::process::Command and make everything async to see if that helps with
// concurrency being more parallel.
impl Notmuch {
pub fn with_config<P: AsRef<Path>>(config_path: P) -> Notmuch {
Notmuch {
@@ -483,7 +472,6 @@ impl Notmuch {
}
}
#[instrument(skip_all)]
pub fn new(&self) -> Result<Vec<u8>, NotmuchError> {
self.run_notmuch(["new"])
}
@@ -492,88 +480,38 @@ impl Notmuch {
self.run_notmuch(std::iter::empty::<&str>())
}
#[instrument(skip_all, fields(query=query))]
pub fn tags_for_query(&self, query: &str) -> Result<Vec<String>, NotmuchError> {
let res = self.run_notmuch(["search", "--format=json", "--output=tags", query])?;
Ok(serde_json::from_slice(&res)?)
}
pub fn tags(&self) -> Result<Vec<String>, NotmuchError> {
self.tags_for_query("*")
}
pub fn tag_add(&self, tag: &str, search_term: &str) -> Result<(), NotmuchError> {
self.tags_add(tag, &[search_term])
}
#[instrument(skip_all, fields(tag=tag,search_term=?search_term))]
pub fn tags_add(&self, tag: &str, search_term: &[&str]) -> Result<(), NotmuchError> {
let tag = format!("+{tag}");
let mut args = vec!["tag", &tag];
args.extend(search_term);
self.run_notmuch(&args)?;
Ok(())
}
pub fn tag_remove(&self, tag: &str, search_term: &str) -> Result<(), NotmuchError> {
self.tags_remove(tag, &[search_term])
}
#[instrument(skip_all, fields(tag=tag,search_term=?search_term))]
pub fn tags_remove(&self, tag: &str, search_term: &[&str]) -> Result<(), NotmuchError> {
let tag = format!("-{tag}");
let mut args = vec!["tag", &tag];
args.extend(search_term);
self.run_notmuch(&args)?;
Ok(())
}
#[instrument(skip_all, fields(query=query,offset=offset,limit=limit))]
pub fn search(
&self,
query: &str,
offset: usize,
limit: usize,
) -> Result<SearchSummary, NotmuchError> {
let query = if query.is_empty() { "*" } else { query };
let res = self
.run_notmuch([
"search",
"--format=json",
&format!("--offset={offset}"),
&format!("--limit={limit}"),
query,
])
.inspect_err(|err| error!("failed to notmuch search for query '{query}': {err}"))?;
Ok(serde_json::from_slice(&res).unwrap_or_else(|err| {
error!("failed to decode search result for query '{query}': {err}");
SearchSummary(Vec::new())
}))
let res = self.run_notmuch([
"search",
"--format=json",
&format!("--offset={offset}"),
&format!("--limit={limit}"),
query,
])?;
Ok(serde_json::from_slice(&res)?)
}
#[instrument(skip_all, fields(query=query))]
pub fn count(&self, query: &str) -> Result<usize, NotmuchError> {
// NOTE: --output=threads is technically more correct, but really slow
// TODO: find a fast thread count path
// let res = self.run_notmuch(["count", "--output=threads", query])?;
let res = self.run_notmuch(["count", query])?;
// Strip '\n' from res.
let s = std::str::from_utf8(&res)?.trim();
Ok(s.parse()
.inspect_err(|err| error!("failed to parse count for query '{query}': {err}"))
.unwrap_or(0))
let s = std::str::from_utf8(&res[..res.len() - 1])?;
Ok(s.parse()?)
}
#[instrument(skip_all, fields(query=query))]
pub fn show(&self, query: &str) -> Result<ThreadSet, NotmuchError> {
let slice = self.run_notmuch([
"show",
"--include-html=true",
"--entire-thread=false",
"--entire-thread=true",
"--format=json",
query,
])?;
// Notmuch returns JSON with invalid unicode. So we lossy convert it to a string here and
// Notmuch returns JSON with invalid unicode. So we lossy convert it to a string here an
// use that for parsing in rust.
let s = String::from_utf8_lossy(&slice);
let mut deserializer = serde_json::Deserializer::from_str(&s);
@@ -583,7 +521,6 @@ impl Notmuch {
Ok(val)
}
#[instrument(skip_all, fields(query=query,part=part))]
pub fn show_part(&self, query: &str, part: usize) -> Result<Part, NotmuchError> {
let slice = self.run_notmuch([
"show",
@@ -593,7 +530,7 @@ impl Notmuch {
&format!("--part={}", part),
query,
])?;
// Notmuch returns JSON with invalid unicode. So we lossy convert it to a string here and
// Notmuch returns JSON with invalid unicode. So we lossy convert it to a string here an
// use that for parsing in rust.
let s = String::from_utf8_lossy(&slice);
let mut deserializer = serde_json::Deserializer::from_str(&s);
@@ -603,107 +540,21 @@ impl Notmuch {
Ok(val)
}
#[instrument(skip_all, fields(id=id))]
pub fn show_original(&self, id: &MessageId) -> Result<Vec<u8>, NotmuchError> {
self.show_original_part(id, 0)
}
#[instrument(skip_all, fields(id=id,part=part))]
pub fn show_original_part(&self, id: &MessageId, part: usize) -> Result<Vec<u8>, NotmuchError> {
let id = if id.starts_with("id:") {
id
} else {
&format!("id:{id}")
};
let res = self.run_notmuch(["show", "--part", &part.to_string(), id])?;
Ok(res)
}
#[instrument(skip_all, fields(query=query))]
pub fn message_ids(&self, query: &str) -> Result<Vec<String>, NotmuchError> {
let res = self.run_notmuch(["search", "--output=messages", "--format=json", query])?;
Ok(serde_json::from_slice(&res)?)
pub fn message_ids(&self, query: &str) -> Result<Lines<BufReader<ChildStdout>>, NotmuchError> {
let mut child = self.run_notmuch_pipe(["search", "--output=messages", query])?;
Ok(BufReader::new(child.stdout.take().unwrap()).lines())
}
#[instrument(skip_all, fields(query=query))]
pub fn files(&self, query: &str) -> Result<Vec<String>, NotmuchError> {
let res = self.run_notmuch(["search", "--output=files", "--format=json", query])?;
Ok(serde_json::from_slice(&res)?)
}
#[instrument(skip_all)]
pub fn unread_recipients(&self) -> Result<HashMap<String, usize>, NotmuchError> {
let slice = self.run_notmuch([
"show",
"--include-html=false",
"--entire-thread=false",
"--body=false",
"--format=json",
// Arbitrary limit to prevent too much work
"--limit=1000",
"is:unread",
])?;
// Notmuch returns JSON with invalid unicode. So we lossy convert it to a string here and
// use that for parsing in rust.
let s = String::from_utf8_lossy(&slice);
let mut deserializer = serde_json::Deserializer::from_str(&s);
deserializer.disable_recursion_limit();
let ts: ThreadSet = serde::de::Deserialize::deserialize(&mut deserializer)?;
deserializer.end()?;
let mut r = HashMap::new();
fn collect_from_thread_node(
r: &mut HashMap<String, usize>,
tn: &ThreadNode,
) -> Result<(), NotmuchError> {
let Some(msg) = &tn.0 else {
return Ok(());
};
let mut addrs = vec![];
let hdr = &msg.headers.to;
if let Some(to) = hdr {
addrs.push(to);
} else {
let hdr = &msg.headers.x_original_to;
if let Some(to) = hdr {
addrs.push(to);
} else {
let hdr = &msg.headers.delivered_to;
if let Some(to) = hdr {
addrs.push(to);
};
};
};
let hdr = &msg.headers.cc;
if let Some(cc) = hdr {
addrs.push(cc);
};
for recipient in addrs {
mailparse::addrparse(&recipient)?
.into_inner()
.iter()
.for_each(|a| {
let mailparse::MailAddr::Single(si) = a else {
return;
};
let addr = &si.addr;
if addr == "couchmoney@gmail.com" || addr.ends_with("@xinu.tv") {
*r.entry(addr.to_lowercase()).or_default() += 1;
}
});
}
Ok(())
}
for t in ts.0 {
for tn in t.0 {
collect_from_thread_node(&mut r, &tn)?;
for sub_tn in tn.1 {
collect_from_thread_node(&mut r, &sub_tn)?;
}
}
}
Ok(r)
}
// TODO(wathiede): implement tags() based on "notmuch search --output=tags '*'"
fn run_notmuch<I, S>(&self, args: I) -> Result<Vec<u8>, NotmuchError>
where
@@ -717,15 +568,23 @@ impl Notmuch {
cmd.args(args);
info!("{:?}", &cmd);
let out = cmd.output()?;
if !out.stderr.is_empty() {
warn!(
"{:?}: STDERR:\n{}",
&cmd,
String::from_utf8_lossy(&out.stderr)
);
}
Ok(out.stdout)
}
fn run_notmuch_pipe<I, S>(&self, args: I) -> Result<Child, NotmuchError>
where
I: IntoIterator<Item = S>,
S: AsRef<OsStr>,
{
let mut cmd = Command::new("notmuch");
if let Some(config_path) = &self.config_path {
cmd.arg("--config").arg(config_path);
}
cmd.args(args);
info!("{:?}", &cmd);
let child = cmd.stdout(Stdio::piped()).spawn()?;
Ok(child)
}
}
#[cfg(test)]

46
notmuch/src/main.rs Normal file
View File

@@ -0,0 +1,46 @@
use std::path::PathBuf;
use clap::{Parser, Subcommand};
use notmuch::Notmuch;
#[derive(Parser)]
#[command(author, version, about, long_about = None)]
#[command(propagate_version = true)]
struct Cli {
/// Optional notmuch config file
#[arg(short, long)]
config: Option<PathBuf>,
#[command(subcommand)]
command: Commands,
}
#[derive(Subcommand)]
enum Commands {
/// Show given search term
Show { search_term: String },
/// Search for given search term
Search { search_term: String },
}
fn main() -> anyhow::Result<()> {
let cli = Cli::parse();
let nm = if let Some(config_path) = cli.config {
Notmuch::with_config(config_path)
} else {
Notmuch::default()
};
// You can check for the existence of subcommands, and if found use their
// matches just as you would the top level cmd
match &cli.command {
Commands::Search { search_term } => {
println!("{:#?}", nm.search(&search_term, 0, 10)?);
}
Commands::Show { search_term } => {
println!("{:#?}", nm.show(&search_term)?);
}
}
Ok(())
}

View File

@@ -1,10 +1,11 @@
use std::{
error::Error,
io::{stdout, Write},
time::Instant,
time::{Duration, Instant},
};
use letterbox_notmuch::Notmuch;
use itertools::Itertools;
use notmuch::{Notmuch, NotmuchError, SearchSummary, ThreadSet};
use rayon::iter::{ParallelBridge, ParallelIterator};
#[test]
@@ -22,11 +23,11 @@ fn parse_one() -> Result<(), Box<dyn Error>> {
let total = nm.count("*")? as f32;
let start = Instant::now();
nm.message_ids("*")?
.iter()
.enumerate()
.par_bridge()
.for_each(|(i, msg)| {
let _ts = nm
let msg = msg.expect("failed to unwrap msg");
let ts = nm
.show(&msg)
.expect(&format!("failed to show msg: {}", msg));
//println!("{:?}", ts);
@@ -76,9 +77,11 @@ fn parse_bulk() -> Result<(), Box<dyn Error>> {
.into_iter()
.enumerate()
//.par_bridge()
.for_each(|(i, msgs)| {
.for_each(|(i, chunk)| {
let msgs: Result<Vec<_>, _> = chunk.collect();
let msgs = msgs.expect("failed to unwrap msg");
let query = msgs.join(" OR ");
let _ts = nm
let ts = nm
.show(&query)
.expect(&format!("failed to show msgs: {}", query));
//println!("{:?}", ts);

View File

@@ -1,20 +1,9 @@
[package]
name = "letterbox-procmail2notmuch"
description = "Tool for generating notmuch rules from procmail"
authors.workspace = true
edition.workspace = true
license.workspace = true
publish.workspace = true
repository.workspace = true
version.workspace = true
name = "procmail2notmuch"
version = "0.1.0"
edition = "2021"
# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
[dependencies]
anyhow = "1.0.98"
clap = { version = "4.5.37", features = ["derive", "env"] }
letterbox-notmuch = { version = "0.17.9", registry = "xinu" }
letterbox-shared = { version = "0.17.9", registry = "xinu" }
serde = { version = "1.0.219", features = ["derive"] }
sqlx = { version = "0.8.5", features = ["postgres", "runtime-tokio"] }
tokio = { version = "1.44.2", features = ["rt", "macros", "rt-multi-thread"] }
anyhow = "1.0.69"

View File

@@ -1,36 +1,210 @@
use std::{collections::HashMap, io::Write};
use std::{convert::Infallible, io::Write, str::FromStr};
use clap::{Parser, Subcommand};
use letterbox_shared::{cleanup_match, Match, MatchType, Rule};
use sqlx::{types::Json, PgPool};
#[derive(Debug, Subcommand)]
enum Mode {
Debug,
Notmuchrc,
LoadSql {
#[arg(short, long)]
dsn: String,
},
#[derive(Debug, Default)]
enum MatchType {
From,
Sender,
To,
Cc,
Subject,
List,
DeliveredTo,
XForwardedTo,
ReplyTo,
XOriginalTo,
XSpam,
Body,
#[default]
Unknown,
}
#[derive(Debug, Default)]
struct Match {
match_type: MatchType,
needle: String,
}
/// Simple program to greet a person
#[derive(Parser, Debug)]
#[command(version, about, long_about = None)]
struct Args {
#[arg(short, long, default_value = "/home/wathiede/dotfiles/procmailrc")]
input: String,
#[command(subcommand)]
mode: Mode,
#[derive(Debug, Default)]
struct Rule {
matches: Vec<Match>,
tags: Vec<String>,
}
#[tokio::main]
async fn main() -> anyhow::Result<()> {
let args = Args::parse();
fn unescape(s: &str) -> String {
s.replace('\\', "")
}
fn cleanup_match(prefix: &str, s: &str) -> String {
unescape(&s[prefix.len()..]).replace(".*", "")
}
mod matches {
pub const TO: &'static str = "TO";
pub const CC: &'static str = "Cc";
pub const TOCC: &'static str = "(TO|Cc)";
pub const FROM: &'static str = "From";
pub const SENDER: &'static str = "Sender";
pub const SUBJECT: &'static str = "Subject";
pub const DELIVERED_TO: &'static str = "Delivered-To";
pub const X_FORWARDED_TO: &'static str = "X-Forwarded-To";
pub const REPLY_TO: &'static str = "Reply-To";
pub const X_ORIGINAL_TO: &'static str = "X-Original-To";
pub const LIST_ID: &'static str = "List-ID";
pub const X_SPAM: &'static str = "X-Spam";
pub const X_SPAM_FLAG: &'static str = "X-Spam-Flag";
}
impl FromStr for Match {
type Err = Infallible;
fn from_str(s: &str) -> Result<Self, Self::Err> {
// Examples:
// "* 1^0 ^TOsonyrewards.com@xinu.tv"
// "* ^TOsonyrewards.com@xinu.tv"
let mut it = s.split_whitespace().skip(1);
let mut needle = it.next().unwrap();
if needle == "1^0" {
needle = it.next().unwrap();
}
let mut needle = vec![needle];
needle.extend(it);
let needle = needle.join(" ");
let first = needle.chars().nth(0).unwrap_or(' ');
use matches::*;
if first == '^' {
let needle = &needle[1..];
if needle.starts_with(TO) {
return Ok(Match {
match_type: MatchType::To,
needle: cleanup_match(TO, needle),
});
} else if needle.starts_with(FROM) {
return Ok(Match {
match_type: MatchType::From,
needle: cleanup_match(FROM, needle),
});
} else if needle.starts_with(CC) {
return Ok(Match {
match_type: MatchType::Cc,
needle: cleanup_match(CC, needle),
});
} else if needle.starts_with(TOCC) {
return Ok(Match {
match_type: MatchType::To,
needle: cleanup_match(TOCC, needle),
});
} else if needle.starts_with(SENDER) {
return Ok(Match {
match_type: MatchType::Sender,
needle: cleanup_match(SENDER, needle),
});
} else if needle.starts_with(SUBJECT) {
return Ok(Match {
match_type: MatchType::Subject,
needle: cleanup_match(SUBJECT, needle),
});
} else if needle.starts_with(X_ORIGINAL_TO) {
return Ok(Match {
match_type: MatchType::XOriginalTo,
needle: cleanup_match(X_ORIGINAL_TO, needle),
});
} else if needle.starts_with(LIST_ID) {
return Ok(Match {
match_type: MatchType::List,
needle: cleanup_match(LIST_ID, needle),
});
} else if needle.starts_with(REPLY_TO) {
return Ok(Match {
match_type: MatchType::ReplyTo,
needle: cleanup_match(REPLY_TO, needle),
});
} else if needle.starts_with(X_SPAM_FLAG) {
return Ok(Match {
match_type: MatchType::XSpam,
needle: '*'.to_string(),
});
} else if needle.starts_with(X_SPAM) {
return Ok(Match {
match_type: MatchType::XSpam,
needle: '*'.to_string(),
});
} else if needle.starts_with(DELIVERED_TO) {
return Ok(Match {
match_type: MatchType::DeliveredTo,
needle: cleanup_match(DELIVERED_TO, needle),
});
} else if needle.starts_with(X_FORWARDED_TO) {
return Ok(Match {
match_type: MatchType::XForwardedTo,
needle: cleanup_match(X_FORWARDED_TO, needle),
});
} else {
unreachable!("needle: '{needle}'")
}
} else {
return Ok(Match {
match_type: MatchType::Body,
needle: cleanup_match("", &needle),
});
}
}
}
fn notmuch_from_rules<W: Write>(mut w: W, rules: &[Rule]) -> anyhow::Result<()> {
// TODO(wathiede): if reindexing this many tags is too slow, see if combining rules per tag is
// faster.
let mut lines = Vec::new();
for r in rules {
for m in &r.matches {
for t in &r.tags {
if let MatchType::Unknown = m.match_type {
eprintln!("rule has unknown match {:?}", r);
continue;
}
let rule = match m.match_type {
MatchType::From => "from:",
// TODO(wathiede): something more specific?
MatchType::Sender => "from:",
MatchType::To => "to:",
MatchType::Cc => "to:",
MatchType::Subject => "subject:",
MatchType::List => "List-ID:",
MatchType::Body => "",
// TODO(wathiede): these will probably require adding fields to notmuch
// index. Handle them later.
MatchType::DeliveredTo
| MatchType::XForwardedTo
| MatchType::ReplyTo
| MatchType::XOriginalTo
| MatchType::XSpam => continue,
MatchType::Unknown => unreachable!(),
};
// Preserve unread status if run with --remove-all
lines.push(format!(
r#"-unprocessed +{} +unread -- is:unread tag:unprocessed {}"{}""#,
t, rule, m.needle
));
lines.push(format!(
// TODO(wathiede): this assumes `notmuch new` is configured to add
// `tag:unprocessed` to all new mail.
r#"-unprocessed +{} -- tag:unprocessed {}"{}""#,
t, rule, m.needle
));
}
}
}
lines.sort();
for l in lines {
writeln!(w, "{l}")?;
}
Ok(())
}
fn main() -> anyhow::Result<()> {
let input = "/home/wathiede/dotfiles/procmailrc";
let mut rules = Vec::new();
let mut cur_rule = Rule::default();
for l in std::fs::read_to_string(args.input)?.lines() {
for l in std::fs::read_to_string(input)?.lines() {
let l = if let Some(idx) = l.find('#') {
&l[..idx]
} else {
@@ -48,9 +222,6 @@ async fn main() -> anyhow::Result<()> {
match first {
':' => {
// start of rule
// If carbon-copy flag present, don't stop on match
cur_rule.stop_on_match = !l.contains('c');
}
'*' => {
// add to current rule
@@ -59,119 +230,26 @@ async fn main() -> anyhow::Result<()> {
}
'.' => {
// delivery to folder
cur_rule.tag = cleanup_match(
cur_rule.tags.push(cleanup_match(
"",
&l.replace('.', "/")
.replace(' ', "")
.trim_matches('/')
.to_string(),
);
));
rules.push(cur_rule);
cur_rule = Rule::default();
}
'/' => cur_rule = Rule::default(), // Ex. /dev/null
'|' => cur_rule = Rule::default(), // external command
'$' => {
// TODO(wathiede): tag messages with no other tag as 'inbox'
cur_rule.tag = cleanup_match("", "inbox");
cur_rule.tags.push(cleanup_match("", "inbox"));
rules.push(cur_rule);
cur_rule = Rule::default();
} // variable, should only be $DEFAULT in my config
_ => panic!("Unhandled first character '{}'\nLine: {}", first, l),
_ => panic!("Unhandled first character '{}' {}", first, l),
}
}
match args.mode {
Mode::Debug => print_rules(&rules),
Mode::Notmuchrc => notmuch_from_rules(std::io::stdout(), &rules)?,
Mode::LoadSql { dsn } => load_sql(&dsn, &rules).await?,
}
Ok(())
}
fn print_rules(rules: &[Rule]) {
let mut tally = HashMap::new();
for r in rules {
for m in &r.matches {
*tally.entry(m.match_type).or_insert(0) += 1;
}
}
let mut sorted: Vec<_> = tally.iter().map(|(k, v)| (v, k)).collect();
sorted.sort();
sorted.reverse();
for (v, k) in sorted {
println!("{k:?}: {v}");
}
}
fn notmuch_from_rules<W: Write>(mut w: W, rules: &[Rule]) -> anyhow::Result<()> {
// TODO(wathiede): if reindexing this many tags is too slow, see if combining rules per tag is
// faster.
let mut lines = Vec::new();
for r in rules {
for m in &r.matches {
let t = &r.tag;
if let MatchType::Unknown = m.match_type {
eprintln!("rule has unknown match {:?}", r);
continue;
}
let rule = match m.match_type {
MatchType::From => "from:",
// TODO(wathiede): something more specific?
MatchType::Sender => "from:",
MatchType::To => "to:",
MatchType::Cc => "to:",
MatchType::Subject => "subject:",
MatchType::ListId => "List-ID:",
MatchType::Body => "",
// TODO(wathiede): these will probably require adding fields to notmuch
// index. Handle them later.
MatchType::DeliveredTo
| MatchType::XForwardedTo
| MatchType::ReplyTo
| MatchType::XOriginalTo
| MatchType::XSpam => continue,
MatchType::Unknown => unreachable!(),
};
// Preserve unread status if run with --remove-all
lines.push(format!(
r#"-unprocessed +{} +unread -- is:unread tag:unprocessed {}"{}""#,
t, rule, m.needle
));
lines.push(format!(
// TODO(wathiede): this assumes `notmuch new` is configured to add
// `tag:unprocessed` to all new mail.
r#"-unprocessed +{} -- tag:unprocessed {}"{}""#,
t, rule, m.needle
));
}
}
lines.sort();
for l in lines {
writeln!(w, "{l}")?;
}
Ok(())
}
async fn load_sql(dsn: &str, rules: &[Rule]) -> anyhow::Result<()> {
let pool = PgPool::connect(dsn).await?;
println!("clearing email_rule table");
sqlx::query!("DELETE FROM email_rule")
.execute(&pool)
.await?;
for (order, rule) in rules.iter().enumerate() {
println!("inserting {order}: {rule:?}");
sqlx::query!(
r#"
INSERT INTO email_rule (sort_order, rule)
VALUES ($1, $2)
"#,
order as i32,
Json(rule) as _
)
.execute(&pool)
.await?;
}
notmuch_from_rules(std::io::stdout(), &rules)?;
Ok(())
}

View File

@@ -1,10 +0,0 @@
set -e
cd ~/dotfiles
git diff
scp nasx:.procmailrc procmailrc
git diff
cd ~/src/xinu.tv/letterbox/procmail2notmuch
cargo run > /tmp/notmuch.tags
mv /tmp/notmuch.tags ~/dotfiles/notmuch.tags
cd ~/dotfiles
git diff

View File

@@ -1,6 +0,0 @@
{
"$schema": "https://docs.renovatebot.com/renovate-schema.json",
"extends": [
"config:recommended"
]
}

View File

@@ -1,5 +0,0 @@
#!env bash
set -e -x
cargo-set-version set-version --bump patch
VERSION="$(awk -F\" '/^version/ {print $2}' server/Cargo.toml)"
git commit Cargo.lock */Cargo.toml -m "Bumping version to ${VERSION:?}"

View File

@@ -1,32 +0,0 @@
{
"db_name": "PostgreSQL",
"query": "SELECT\n site,\n name,\n count (\n NOT is_read\n OR NULL\n ) unread\nFROM\n post AS p\n JOIN feed AS f ON p.site = f.slug --\n -- TODO: figure this out to make the query faster when only looking for unread\n --WHERE\n -- (\n -- NOT $1\n -- OR NOT is_read\n -- )\nGROUP BY\n 1,\n 2\nORDER BY\n site\n",
"describe": {
"columns": [
{
"ordinal": 0,
"name": "site",
"type_info": "Text"
},
{
"ordinal": 1,
"name": "name",
"type_info": "Text"
},
{
"ordinal": 2,
"name": "unread",
"type_info": "Int8"
}
],
"parameters": {
"Left": []
},
"nullable": [
true,
true,
null
]
},
"hash": "2dcbedef656e1b725c5ba4fb67d31ce7962d8714449b2fb630f49a7ed1acc270"
}

View File

@@ -1,70 +0,0 @@
{
"db_name": "PostgreSQL",
"query": "SELECT\n date,\n is_read,\n link,\n site,\n summary,\n clean_summary,\n title,\n name,\n homepage\nFROM\n post AS p\nINNER JOIN feed AS f ON p.site = f.slug\nWHERE\n uid = $1\n",
"describe": {
"columns": [
{
"ordinal": 0,
"name": "date",
"type_info": "Timestamp"
},
{
"ordinal": 1,
"name": "is_read",
"type_info": "Bool"
},
{
"ordinal": 2,
"name": "link",
"type_info": "Text"
},
{
"ordinal": 3,
"name": "site",
"type_info": "Text"
},
{
"ordinal": 4,
"name": "summary",
"type_info": "Text"
},
{
"ordinal": 5,
"name": "clean_summary",
"type_info": "Text"
},
{
"ordinal": 6,
"name": "title",
"type_info": "Text"
},
{
"ordinal": 7,
"name": "name",
"type_info": "Text"
},
{
"ordinal": 8,
"name": "homepage",
"type_info": "Text"
}
],
"parameters": {
"Left": [
"Text"
]
},
"nullable": [
true,
true,
false,
true,
true,
true,
true,
true,
true
]
},
"hash": "383221a94bc3746322ba78e41cde37994440ee67dc32e88d2394c51211bde6cd"
}

View File

@@ -1,32 +0,0 @@
{
"db_name": "PostgreSQL",
"query": "SELECT\n p.id,\n link,\n clean_summary\nFROM\n post AS p\nINNER JOIN feed AS f ON p.site = f.slug -- necessary to weed out nzb posts\nWHERE\n search_summary IS NULL\n -- TODO remove AND link ~ '^<'\nORDER BY\n ROW_NUMBER() OVER (PARTITION BY site ORDER BY date DESC)\nLIMIT 100;\n",
"describe": {
"columns": [
{
"ordinal": 0,
"name": "id",
"type_info": "Int4"
},
{
"ordinal": 1,
"name": "link",
"type_info": "Text"
},
{
"ordinal": 2,
"name": "clean_summary",
"type_info": "Text"
}
],
"parameters": {
"Left": []
},
"nullable": [
false,
false,
true
]
},
"hash": "3d271b404f06497a5dcde68cf6bf07291d70fa56058ea736ac24e91d33050c04"
}

View File

@@ -1,24 +0,0 @@
{
"db_name": "PostgreSQL",
"query": "SELECT COUNT(*) AS count\nFROM\n post\nWHERE\n (\n $1::text IS NULL\n OR site = $1\n )\n AND (\n NOT $2\n OR NOT is_read\n )\n AND (\n $3::text IS NULL\n OR TO_TSVECTOR('english', search_summary)\n @@ WEBSEARCH_TO_TSQUERY('english', $3)\n )\n",
"describe": {
"columns": [
{
"ordinal": 0,
"name": "count",
"type_info": "Int8"
}
],
"parameters": {
"Left": [
"Text",
"Bool",
"Text"
]
},
"nullable": [
null
]
},
"hash": "8c1b3c78649135e98b89092237750088433f7ff1b7c2ddeedec553406ea9f203"
}

View File

@@ -1,15 +0,0 @@
{
"db_name": "PostgreSQL",
"query": "UPDATE\n post\nSET\n is_read = $1\nWHERE\n uid = $2\n",
"describe": {
"columns": [],
"parameters": {
"Left": [
"Bool",
"Text"
]
},
"nullable": []
},
"hash": "b39147b9d06171cb742141eda4675688cb702fb284758b1224ed3aa2d7f3b3d9"
}

View File

@@ -1,15 +0,0 @@
{
"db_name": "PostgreSQL",
"query": "UPDATE post SET search_summary = $1 WHERE id = $2",
"describe": {
"columns": [],
"parameters": {
"Left": [
"Text",
"Int4"
]
},
"nullable": []
},
"hash": "ef8327f039dbfa8f4e59b7a77a6411252a346bf51cf940024a17d9fbb2df173c"
}

View File

@@ -1,56 +0,0 @@
{
"db_name": "PostgreSQL",
"query": "SELECT\n site,\n date,\n is_read,\n title,\n uid,\n name\nFROM\n post p\n JOIN feed f ON p.site = f.slug\nWHERE\n ($1::text IS NULL OR site = $1)\n AND (\n NOT $2\n OR NOT is_read\n )\n AND (\n $5 :: text IS NULL\n OR to_tsvector('english', search_summary) @@ websearch_to_tsquery('english', $5)\n )\nORDER BY\n date DESC,\n title OFFSET $3\nLIMIT\n $4\n",
"describe": {
"columns": [
{
"ordinal": 0,
"name": "site",
"type_info": "Text"
},
{
"ordinal": 1,
"name": "date",
"type_info": "Timestamp"
},
{
"ordinal": 2,
"name": "is_read",
"type_info": "Bool"
},
{
"ordinal": 3,
"name": "title",
"type_info": "Text"
},
{
"ordinal": 4,
"name": "uid",
"type_info": "Text"
},
{
"ordinal": 5,
"name": "name",
"type_info": "Text"
}
],
"parameters": {
"Left": [
"Text",
"Bool",
"Int8",
"Int8",
"Text"
]
},
"nullable": [
true,
true,
true,
true,
false,
true
]
},
"hash": "fc4607f02cc76a5f3a6629cce4507c74f52ae44820897b47365da3f339d1da06"
}

View File

@@ -1,64 +1,25 @@
[package]
name = "letterbox-server"
default-run = "letterbox-server"
description = "Backend for letterbox"
authors.workspace = true
edition.workspace = true
license.workspace = true
publish.workspace = true
repository.workspace = true
version.workspace = true
name = "server"
version = "0.1.0"
edition = "2021"
default-bin = "server"
# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
[dependencies]
ammonia = "4.1.0"
anyhow = "1.0.98"
askama = { version = "0.14.0", features = ["derive"] }
async-graphql = { version = "7", features = ["log"] }
async-graphql-axum = "7.0.16"
async-trait = "0.1.88"
axum = { version = "0.8.3", features = ["ws"] }
axum-macros = "0.5.0"
build-info = "0.0.41"
cacher = { version = "0.2.0", registry = "xinu" }
chrono = "0.4.40"
clap = { version = "4.5.37", features = ["derive"] }
css-inline = "0.17.0"
flate2 = "1.1.2"
futures = "0.3.31"
headers = "0.4.0"
html-escape = "0.2.13"
letterbox-notmuch = { path = "../notmuch", version = "0.17.32", registry = "xinu" }
letterbox-shared = { path = "../shared", version = "0.17.32", registry = "xinu" }
linkify = "0.10.0"
lol_html = "2.3.0"
mailparse = "0.16.1"
maplit = "1.0.2"
memmap = "0.7.0"
quick-xml = { version = "0.38.1", features = ["serialize"] }
regex = "1.11.1"
reqwest = { version = "0.12.15", features = ["blocking"] }
scraper = "0.23.1"
serde = { version = "1.0.219", features = ["derive"] }
serde_json = "1.0.140"
sqlx = { version = "0.8.5", features = ["postgres", "runtime-tokio", "time"] }
tantivy = { version = "0.24.1", optional = true }
thiserror = "2.0.12"
tokio = "1.44.2"
tower-http = { version = "0.6.2", features = ["trace"] }
tracing = "0.1.41"
url = "2.5.4"
rocket = { version = "0.5.0-rc.2", features = [ "json" ] }
rocket_cors = { git = "https://github.com/lawliet89/rocket_cors", branch = "master" }
notmuch = { path = "../notmuch" }
shared = { path = "../shared" }
serde_json = "1.0.87"
thiserror = "1.0.37"
serde = { version = "1.0.147", features = ["derive"] }
log = "0.4.17"
tokio = "1.26.0"
glog = "0.1.0"
urlencoding = "2.1.3"
#xtracing = { git = "http://git-private.h.xinu.tv/wathiede/xtracing.git" }
#xtracing = { path = "../../xtracing" }
xtracing = { version = "0.3.2", registry = "xinu" }
zip = "4.3.0"
[build-dependencies]
build-info-build = "0.0.41"
[features]
#default = [ "tantivy" ]
tantivy = ["dep:tantivy"]
[dependencies.rocket_contrib]
version = "0.4.11"
default-features = false
features = ["json"]

View File

@@ -1,13 +1,9 @@
[release]
address = "0.0.0.0"
port = 9345
newsreader_database_url = "postgres://newsreader@nixos-07.h.xinu.tv/newsreader"
newsreader_tantivy_db_path = "../target/database/newsreader"
[debug]
address = "0.0.0.0"
port = 9345
# Uncomment to make it production like.
#log_level = "critical"
newsreader_database_url = "postgres://newsreader@nixos-07.h.xinu.tv/newsreader"
newsreader_tantivy_db_path = "../target/database/newsreader"
slurp_cache_path = "/tmp/letterbox/slurp"

View File

@@ -1,6 +0,0 @@
fn main() {
// Calling `build_info_build::build_script` collects all data and makes it available to `build_info::build_info!`
// and `build_info::format!` in the main program.
build_info_build::build_script();
println!("cargo:rerun-if-changed=templates");
}

View File

@@ -1,3 +0,0 @@
DROP INDEX IF EXISTS post_summary_idx;
DROP INDEX IF EXISTS post_site_idx;
DROP INDEX IF EXISTS post_title_idx;

View File

@@ -1,3 +0,0 @@
CREATE INDEX post_summary_idx ON post USING GIN (to_tsvector('english', summary));
CREATE INDEX post_site_idx ON post USING GIN (to_tsvector('english', site));
CREATE INDEX post_title_idx ON post USING GIN (to_tsvector('english', title));

View File

@@ -1,24 +0,0 @@
BEGIN;
ALTER TABLE IF EXISTS public."Email" DROP CONSTRAINT IF EXISTS email_avatar_fkey;
ALTER TABLE IF EXISTS public."EmailDisplayName" DROP CONSTRAINT IF EXISTS email_id_fk;
ALTER TABLE IF EXISTS public."Message" DROP CONSTRAINT IF EXISTS message_to_fkey;
ALTER TABLE IF EXISTS public."Message" DROP CONSTRAINT IF EXISTS message_cc_fkey;
ALTER TABLE IF EXISTS public."Message" DROP CONSTRAINT IF EXISTS message_from_fkey;
ALTER TABLE IF EXISTS public."Message" DROP CONSTRAINT IF EXISTS message_header_fkey;
ALTER TABLE IF EXISTS public."Message" DROP CONSTRAINT IF EXISTS message_file_fkey;
ALTER TABLE IF EXISTS public."Message" DROP CONSTRAINT IF EXISTS message_body_id_fkey;
ALTER TABLE IF EXISTS public."Message" DROP CONSTRAINT IF EXISTS message_thread_fkey;
ALTER TABLE IF EXISTS public."Message" DROP CONSTRAINT IF EXISTS message_tag_fkey;
DROP TABLE IF EXISTS public."Email";
DROP TABLE IF EXISTS public."EmailDisplayName";
DROP TABLE IF EXISTS public."Message";
DROP TABLE IF EXISTS public."Header";
DROP TABLE IF EXISTS public."File";
DROP TABLE IF EXISTS public."Avatar";
DROP TABLE IF EXISTS public."Body";
DROP TABLE IF EXISTS public."Thread";
DROP TABLE IF EXISTS public."Tag";
END;

View File

@@ -1,174 +0,0 @@
-- This script was generated by the ERD tool in pgAdmin 4.
-- Please log an issue at https://github.com/pgadmin-org/pgadmin4/issues/new/choose if you find any bugs, including reproduction steps.
BEGIN;
ALTER TABLE IF EXISTS public."Email" DROP CONSTRAINT IF EXISTS email_avatar_fkey;
ALTER TABLE IF EXISTS public."EmailDisplayName" DROP CONSTRAINT IF EXISTS email_id_fk;
ALTER TABLE IF EXISTS public."Message" DROP CONSTRAINT IF EXISTS message_to_fkey;
ALTER TABLE IF EXISTS public."Message" DROP CONSTRAINT IF EXISTS message_cc_fkey;
ALTER TABLE IF EXISTS public."Message" DROP CONSTRAINT IF EXISTS message_from_fkey;
ALTER TABLE IF EXISTS public."Message" DROP CONSTRAINT IF EXISTS message_header_fkey;
ALTER TABLE IF EXISTS public."Message" DROP CONSTRAINT IF EXISTS message_file_fkey;
ALTER TABLE IF EXISTS public."Message" DROP CONSTRAINT IF EXISTS message_body_id_fkey;
ALTER TABLE IF EXISTS public."Message" DROP CONSTRAINT IF EXISTS message_thread_fkey;
ALTER TABLE IF EXISTS public."Message" DROP CONSTRAINT IF EXISTS message_tag_fkey;
CREATE TABLE IF NOT EXISTS public."Email"
(
id integer NOT NULL GENERATED ALWAYS AS IDENTITY,
address text NOT NULL,
avatar_id integer,
PRIMARY KEY (id),
CONSTRAINT avatar_id UNIQUE (avatar_id)
);
CREATE TABLE IF NOT EXISTS public."EmailDisplayName"
(
id integer NOT NULL GENERATED ALWAYS AS IDENTITY,
email_id integer NOT NULL,
PRIMARY KEY (id)
);
CREATE TABLE IF NOT EXISTS public."Message"
(
id integer NOT NULL GENERATED ALWAYS AS IDENTITY,
subject text,
"from" integer,
"to" integer,
cc integer,
header_id integer,
hash text NOT NULL,
file_id integer NOT NULL,
date timestamp with time zone NOT NULL,
unread boolean NOT NULL,
body_id integer NOT NULL,
thread_id integer NOT NULL,
tag_id integer,
CONSTRAINT body_id UNIQUE (body_id)
);
CREATE TABLE IF NOT EXISTS public."Header"
(
id integer NOT NULL GENERATED ALWAYS AS IDENTITY,
key text NOT NULL,
value text NOT NULL,
PRIMARY KEY (id)
);
CREATE TABLE IF NOT EXISTS public."File"
(
id integer NOT NULL GENERATED ALWAYS AS IDENTITY,
path text NOT NULL,
PRIMARY KEY (id)
);
CREATE TABLE IF NOT EXISTS public."Avatar"
(
id integer NOT NULL GENERATED ALWAYS AS IDENTITY,
url text NOT NULL,
PRIMARY KEY (id)
);
CREATE TABLE IF NOT EXISTS public."Body"
(
id integer NOT NULL GENERATED ALWAYS AS IDENTITY,
text text NOT NULL,
PRIMARY KEY (id)
);
CREATE TABLE IF NOT EXISTS public."Thread"
(
id integer NOT NULL GENERATED ALWAYS AS IDENTITY,
PRIMARY KEY (id)
);
CREATE TABLE IF NOT EXISTS public."Tag"
(
id integer NOT NULL GENERATED ALWAYS AS IDENTITY,
name text NOT NULL,
display text,
fg_color integer,
bg_color integer,
PRIMARY KEY (id)
);
ALTER TABLE IF EXISTS public."Email"
ADD CONSTRAINT email_avatar_fkey FOREIGN KEY (avatar_id)
REFERENCES public."Avatar" (id) MATCH SIMPLE
ON UPDATE NO ACTION
ON DELETE NO ACTION
NOT VALID;
ALTER TABLE IF EXISTS public."EmailDisplayName"
ADD CONSTRAINT email_id_fk FOREIGN KEY (email_id)
REFERENCES public."Email" (id) MATCH SIMPLE
ON UPDATE NO ACTION
ON DELETE NO ACTION
NOT VALID;
ALTER TABLE IF EXISTS public."Message"
ADD CONSTRAINT message_to_fkey FOREIGN KEY ("to")
REFERENCES public."Email" (id) MATCH SIMPLE
ON UPDATE NO ACTION
ON DELETE NO ACTION
NOT VALID;
ALTER TABLE IF EXISTS public."Message"
ADD CONSTRAINT message_cc_fkey FOREIGN KEY (cc)
REFERENCES public."Email" (id) MATCH SIMPLE
ON UPDATE NO ACTION
ON DELETE NO ACTION
NOT VALID;
ALTER TABLE IF EXISTS public."Message"
ADD CONSTRAINT message_from_fkey FOREIGN KEY ("from")
REFERENCES public."Email" (id) MATCH SIMPLE
ON UPDATE NO ACTION
ON DELETE NO ACTION
NOT VALID;
ALTER TABLE IF EXISTS public."Message"
ADD CONSTRAINT message_header_fkey FOREIGN KEY (header_id)
REFERENCES public."Header" (id) MATCH SIMPLE
ON UPDATE NO ACTION
ON DELETE NO ACTION
NOT VALID;
ALTER TABLE IF EXISTS public."Message"
ADD CONSTRAINT message_file_fkey FOREIGN KEY (file_id)
REFERENCES public."File" (id) MATCH SIMPLE
ON UPDATE NO ACTION
ON DELETE NO ACTION
NOT VALID;
ALTER TABLE IF EXISTS public."Message"
ADD CONSTRAINT message_body_id_fkey FOREIGN KEY (body_id)
REFERENCES public."Body" (id) MATCH SIMPLE
ON UPDATE NO ACTION
ON DELETE NO ACTION
NOT VALID;
ALTER TABLE IF EXISTS public."Message"
ADD CONSTRAINT message_thread_fkey FOREIGN KEY (thread_id)
REFERENCES public."Thread" (id) MATCH SIMPLE
ON UPDATE NO ACTION
ON DELETE NO ACTION
NOT VALID;
ALTER TABLE IF EXISTS public."Message"
ADD CONSTRAINT message_tag_fkey FOREIGN KEY (tag_id)
REFERENCES public."Tag" (id) MATCH SIMPLE
ON UPDATE NO ACTION
ON DELETE NO ACTION
NOT VALID;
END;

View File

@@ -1,3 +0,0 @@
-- Add down migration script here
ALTER TABLE
post DROP CONSTRAINT post_link_key;

View File

@@ -1,28 +0,0 @@
WITH dupes AS (
SELECT
uid,
link,
Row_number() over(
PARTITION by link
ORDER BY
link
) AS RowNumber
FROM
post
)
DELETE FROM
post
WHERE
uid IN (
SELECT
uid
FROM
dupes
WHERE
RowNumber > 1
);
ALTER TABLE
post
ADD
UNIQUE (link);

View File

@@ -1,7 +0,0 @@
ALTER TABLE
post
ALTER COLUMN
link DROP NOT NULL;
ALTER TABLE
post DROP CONSTRAINT link;

View File

@@ -1,17 +0,0 @@
DELETE FROM
post
WHERE
link IS NULL
OR link = '';
ALTER TABLE
post
ALTER COLUMN
link
SET
NOT NULL;
ALTER TABLE
post
ADD
CONSTRAINT link CHECK (link <> '');

View File

@@ -1,3 +0,0 @@
DROP TABLE IF EXISTS email_address;
DROP TABLE IF EXISTS photo;
DROP TABLE IF EXISTS google_person;

View File

@@ -1,19 +0,0 @@
-- Add up migration script here
CREATE TABLE IF NOT EXISTS google_person (
id SERIAL PRIMARY KEY,
resource_name TEXT NOT NULL UNIQUE,
display_name TEXT NOT NULL
);
CREATE TABLE IF NOT EXISTS email_photo (
id SERIAL PRIMARY KEY,
google_person_id INTEGER REFERENCES google_person (id) UNIQUE,
url TEXT NOT NULL
);
CREATE TABLE IF NOT EXISTS email_address (
id SERIAL PRIMARY KEY,
address TEXT NOT NULL UNIQUE,
email_photo_id INTEGER REFERENCES email_photo (id),
google_person_id INTEGER REFERENCES google_person (id)
);

View File

@@ -1,5 +0,0 @@
-- Add down migration script here
DROP INDEX post_summary_idx;
CREATE INDEX post_summary_idx ON post USING gin (
to_tsvector('english', summary)
);

View File

@@ -1,11 +0,0 @@
-- Something like this around summary in the idx w/ tsvector
DROP INDEX post_summary_idx;
CREATE INDEX post_summary_idx ON post USING gin (to_tsvector(
'english',
regexp_replace(
regexp_replace(summary, '<[^>]+>', ' ', 'g'),
'\s+',
' ',
'g'
)
));

View File

@@ -1,2 +0,0 @@
-- Add down migration script here
DROP INDEX nzb_posts_created_at_idx;

View File

@@ -1,2 +0,0 @@
-- Add up migration script here
CREATE INDEX nzb_posts_created_at_idx ON nzb_posts USING btree (created_at);

View File

@@ -1,15 +0,0 @@
-- Add down migration script here
BEGIN;
DROP INDEX IF EXISTS post_search_summary_idx;
ALTER TABLE post DROP search_summary;
-- CREATE INDEX post_summary_idx ON post USING gin (to_tsvector(
-- 'english',
-- regexp_replace(
-- regexp_replace(summary, '<[^>]+>', ' ', 'g'),
-- '\s+',
-- ' ',
-- 'g'
-- )
-- ));
COMMIT;

View File

@@ -1,14 +0,0 @@
-- Add up migration script here
BEGIN;
DROP INDEX IF EXISTS post_summary_idx;
ALTER TABLE post ADD search_summary TEXT;
CREATE INDEX post_search_summary_idx ON post USING gin (
to_tsvector('english', search_summary)
);
UPDATE post SET search_summary = regexp_replace(
regexp_replace(summary, '<[^>]+>', ' ', 'g'),
'\s+',
' ',
'g'
);
COMMIT;

View File

@@ -1,20 +0,0 @@
-- Bad examples:
-- https://nzbfinder.ws/getnzb/d2c3e5a08abadd985dccc6a574122892030b6a9a.nzb&i=95972&r=b55082d289937c050dedc203c9653850
-- https://nzbfinder.ws/getnzb?id=45add174-7da4-4445-bf2b-a67dbbfc07fe.nzb&r=b55082d289937c050dedc203c9653850
-- https://nzbfinder.ws/api/v1/getnzb?id=82486020-c192-4fa0-a7e7-798d7d72e973.nzb&r=b55082d289937c050dedc203c9653850
UPDATE nzb_posts
SET link =
regexp_replace(
regexp_replace(
regexp_replace(
link,
'https://nzbfinder.ws/getnzb/',
'https://nzbfinder.ws/api/v1/getnzb?id='
),
'https://nzbfinder.ws/getnzb',
'https://nzbfinder.ws/api/v1/getnzb'
),
'&r=',
'&apikey='
)
;

View File

@@ -1,3 +0,0 @@
DROP TABLE IF NOT EXISTS email_rule;
-- Add down migration script here

View File

@@ -1,5 +0,0 @@
CREATE TABLE IF NOT EXISTS email_rule (
id integer NOT NULL GENERATED ALWAYS AS IDENTITY,
sort_order integer NOT NULL,
rule jsonb NOT NULL
);

View File

@@ -1,2 +0,0 @@
-- Add down migration script here
ALTER TABLE feed DROP COLUMN IF EXISTS disabled;

View File

@@ -1,2 +0,0 @@
-- Add up migration script here
ALTER TABLE feed ADD disabled boolean;

View File

@@ -1,14 +0,0 @@
SELECT
site,
title,
summary,
link,
date,
is_read,
uid,
p.id id
FROM
post AS p
JOIN feed AS f ON p.site = f.slug -- necessary to weed out nzb posts
ORDER BY
date DESC;

View File

@@ -1,6 +0,0 @@
SELECT
uid
FROM
post AS p
JOIN feed AS f ON p.site = f.slug -- necessary to weed out nzb posts
;

View File

@@ -1,17 +0,0 @@
SELECT COUNT(*) AS count
FROM
post
WHERE
(
$1::text IS NULL
OR site = $1
)
AND (
NOT $2
OR NOT is_read
)
AND (
$3::text IS NULL
OR TO_TSVECTOR('english', search_summary)
@@ WEBSEARCH_TO_TSQUERY('english', $3)
)

View File

@@ -1 +0,0 @@
SELECT rule as "rule: Json<Rule>" FROM email_rule ORDER BY sort_order

View File

@@ -1,13 +0,0 @@
SELECT
p.id,
link,
clean_summary
FROM
post AS p
INNER JOIN feed AS f ON p.site = f.slug -- necessary to weed out nzb posts
WHERE
search_summary IS NULL
-- TODO remove AND link ~ '^<'
ORDER BY
ROW_NUMBER() OVER (PARTITION BY site ORDER BY date DESC)
LIMIT 100;

View File

@@ -1 +0,0 @@
SELECT url FROM email_photo ep JOIN email_address ea ON ep.id = ea.email_photo_id WHERE address = $1

View File

@@ -1,14 +0,0 @@
SELECT
site AS "site!",
title AS "title!",
summary AS "summary!",
link AS "link!",
date AS "date!",
is_read AS "is_read!",
uid AS "uid!",
p.id id
FROM
post p
JOIN feed f ON p.site = f.slug
WHERE
uid = ANY ($1);

View File

@@ -1,6 +0,0 @@
UPDATE
post
SET
is_read = $1
WHERE
uid = $2

View File

@@ -1,21 +0,0 @@
SELECT
site,
name,
count (
NOT is_read
OR NULL
) unread
FROM
post AS p
JOIN feed AS f ON p.site = f.slug --
-- TODO: figure this out to make the query faster when only looking for unread
--WHERE
-- (
-- NOT $1
-- OR NOT is_read
-- )
GROUP BY
1,
2
ORDER BY
site

View File

@@ -1,15 +0,0 @@
SELECT
date,
is_read,
link,
site,
summary,
clean_summary,
title,
name,
homepage
FROM
post AS p
INNER JOIN feed AS f ON p.site = f.slug
WHERE
uid = $1

View File

@@ -1,14 +0,0 @@
SELECT
site,
date,
is_read,
title,
uid,
name
FROM
post p
JOIN feed f ON p.site = f.slug
WHERE
uid = ANY ($1)
ORDER BY
date DESC;

View File

@@ -1,25 +0,0 @@
SELECT
site,
date,
is_read,
title,
uid,
name
FROM
post p
JOIN feed f ON p.site = f.slug
WHERE
($1::text IS NULL OR site = $1)
AND (
NOT $2
OR NOT is_read
)
AND (
$5 :: text IS NULL
OR to_tsvector('english', search_summary) @@ websearch_to_tsquery('english', $5)
)
ORDER BY
date DESC,
title OFFSET $3
LIMIT
$4

View File

@@ -1,13 +0,0 @@
select t.id, tt.tokid, tt.alias, length(t.token), t.token from (
select id, (ts_parse('default',
-- regexp_replace(
-- regexp_replace(summary, '<[^>]+>', ' ', 'g'),
-- '\s+',
-- ' ',
-- 'g'
-- )
summary
)).* from post) t
inner join ts_token_type('default') tt
on t.tokid = tt.tokid
where length(token) >= 2*1024;

View File

@@ -1,16 +0,0 @@
use std::fs;
use letterbox_server::sanitize_html;
fn main() -> anyhow::Result<()> {
let mut args = std::env::args().skip(1);
let src = args.next().expect("source not specified");
let dst = args.next().expect("destination not specified");
println!("Sanitizing {src} into {dst}");
let bytes = fs::read(src)?;
let html = String::from_utf8_lossy(&bytes);
let html = sanitize_html(&html, "", &None)?;
fs::write(dst, html)?;
Ok(())
}

View File

@@ -1,21 +0,0 @@
use std::fs;
use url::Url;
fn main() -> anyhow::Result<()> {
println!("PWD: {}", std::env::current_dir()?.display());
let _url = "https://slashdot.org/story/25/01/24/1813201/walgreens-replaced-fridge-doors-with-smart-screens-its-now-a-200-million-fiasco?utm_source=rss1.0mainlinkanon&utm_medium=feed";
let _url = "https://hackaday.com/2025/01/24/hackaday-podcast-episode-305-caustic-clocks-practice-bones-and-brick-layers/";
let _url = "https://theonion.com/monster-devastated-to-see-film-depicting-things-he-told-guillermo-del-toro-in-confidence/";
let _url = "https://trofi.github.io/posts/330-another-nix-language-nondeterminism-example.html";
let _url = "https://blog.cloudflare.com/ddos-threat-report-for-2024-q4/";
let url = "https://trofi.github.io/posts/330-another-nix-language-nondeterminism-example.html";
let body = reqwest::blocking::get(url)?.text()?;
let output = "/tmp/h2md/output.html";
let inliner = css_inline::CSSInliner::options()
.base_url(Url::parse(url).ok())
.build();
let inlined = inliner.inline(&body)?;
fs::write(output, inlined)?;
Ok(())
}

View File

@@ -1,343 +0,0 @@
// Rocket generates a lot of warnings for handlers
// TODO: figure out why
#![allow(unreachable_patterns)]
use std::{error::Error, net::SocketAddr, sync::Arc, time::Duration};
use async_graphql::{extensions, http::GraphiQLSource, Schema};
use async_graphql_axum::{GraphQL, GraphQLSubscription};
//allows to extract the IP of connecting user
use axum::extract::connect_info::ConnectInfo;
use axum::{
extract::{self, ws::WebSocketUpgrade, Query, State},
http::{header, StatusCode},
response::{self, IntoResponse, Response},
routing::{any, get, post},
Router,
};
use cacher::FilesystemCacher;
use clap::Parser;
use letterbox_notmuch::Notmuch;
#[cfg(feature = "tantivy")]
use letterbox_server::tantivy::TantivyConnection;
use letterbox_server::{
graphql::{compute_catchup_ids, Attachment, MutationRoot, QueryRoot, SubscriptionRoot},
nm::{attachment_bytes, cid_attachment_bytes, label_unprocessed},
ws::ConnectionTracker,
};
use letterbox_shared::WebsocketMessage;
use serde::Deserialize;
use sqlx::postgres::PgPool;
use tokio::{net::TcpListener, sync::Mutex};
use tower_http::trace::{DefaultMakeSpan, TraceLayer};
use tracing::{error, info};
// Make our own error that wraps `ServerError`.
struct AppError(letterbox_server::ServerError);
// Tell axum how to convert `AppError` into a response.
impl IntoResponse for AppError {
fn into_response(self) -> Response {
(
StatusCode::INTERNAL_SERVER_ERROR,
format!("Something went wrong: {}", self.0),
)
.into_response()
}
}
// This enables using `?` on functions that return `Result<_, letterbox_server::Error>` to turn them into
// `Result<_, AppError>`. That way you don't need to do that manually.
impl<E> From<E> for AppError
where
E: Into<letterbox_server::ServerError>,
{
fn from(err: E) -> Self {
Self(err.into())
}
}
fn inline_attachment_response(attachment: Attachment) -> impl IntoResponse {
info!("attachment filename {:?}", attachment.filename);
let mut hdr_map = headers::HeaderMap::new();
if let Some(filename) = attachment.filename {
hdr_map.insert(
header::CONTENT_DISPOSITION,
format!(r#"inline; filename="{}""#, filename)
.parse()
.unwrap(),
);
}
if let Some(ct) = attachment.content_type {
hdr_map.insert(header::CONTENT_TYPE, ct.parse().unwrap());
}
info!("hdr_map {hdr_map:?}");
(hdr_map, attachment.bytes).into_response()
}
fn download_attachment_response(attachment: Attachment) -> impl IntoResponse {
info!("attachment filename {:?}", attachment.filename);
let mut hdr_map = headers::HeaderMap::new();
if let Some(filename) = attachment.filename {
hdr_map.insert(
header::CONTENT_DISPOSITION,
format!(r#"attachment; filename="{}""#, filename)
.parse()
.unwrap(),
);
}
if let Some(ct) = attachment.content_type {
hdr_map.insert(header::CONTENT_TYPE, ct.parse().unwrap());
}
info!("hdr_map {hdr_map:?}");
(hdr_map, attachment.bytes).into_response()
}
#[axum_macros::debug_handler]
async fn view_attachment(
State(AppState { nm, .. }): State<AppState>,
extract::Path((id, idx, _)): extract::Path<(String, String, String)>,
) -> Result<impl IntoResponse, AppError> {
let mid = if id.starts_with("id:") {
id.to_string()
} else {
format!("id:{}", id)
};
info!("view attachment {mid} {idx}");
let idx: Vec<_> = idx
.split('.')
.map(|s| s.parse().expect("not a usize"))
.collect();
let attachment = attachment_bytes(&nm, &mid, &idx)?;
Ok(inline_attachment_response(attachment))
}
async fn download_attachment(
State(AppState { nm, .. }): State<AppState>,
extract::Path((id, idx, _)): extract::Path<(String, String, String)>,
) -> Result<impl IntoResponse, AppError> {
let mid = if id.starts_with("id:") {
id.to_string()
} else {
format!("id:{}", id)
};
info!("download attachment {mid} {idx}");
let idx: Vec<_> = idx
.split('.')
.map(|s| s.parse().expect("not a usize"))
.collect();
let attachment = attachment_bytes(&nm, &mid, &idx)?;
Ok(download_attachment_response(attachment))
}
async fn view_cid(
State(AppState { nm, .. }): State<AppState>,
extract::Path((id, cid)): extract::Path<(String, String)>,
) -> Result<impl IntoResponse, AppError> {
let mid = if id.starts_with("id:") {
id.to_string()
} else {
format!("id:{}", id)
};
info!("view cid attachment {mid} {cid}");
let attachment = cid_attachment_bytes(&nm, &mid, &cid)?;
Ok(inline_attachment_response(attachment))
}
// TODO make this work with gitea message ids like `wathiede/letterbox/pulls/91@git.z.xinu.tv`
async fn view_original(
State(AppState { nm, .. }): State<AppState>,
extract::Path(id): extract::Path<String>,
) -> Result<impl IntoResponse, AppError> {
info!("view_original {id}");
let bytes = nm.show_original(&id)?;
let s = String::from_utf8_lossy(&bytes).to_string();
Ok(s.into_response())
}
async fn graphiql() -> impl IntoResponse {
response::Html(
GraphiQLSource::build()
.endpoint("/api/graphql/")
.subscription_endpoint("/api/graphql/ws")
.finish(),
)
}
async fn start_ws(
ws: WebSocketUpgrade,
ConnectInfo(addr): ConnectInfo<SocketAddr>,
State(AppState {
connection_tracker, ..
}): State<AppState>,
) -> impl IntoResponse {
info!("intiating websocket connection for {addr}");
ws.on_upgrade(async move |socket| connection_tracker.lock().await.add_peer(socket, addr).await)
}
#[derive(Debug, Deserialize)]
struct NotificationParams {
delay_ms: Option<u64>,
num_unprocessed: Option<usize>,
}
async fn send_refresh_websocket_handler(
State(AppState {
nm,
pool,
connection_tracker,
..
}): State<AppState>,
params: Query<NotificationParams>,
) -> impl IntoResponse {
info!("send_refresh_websocket_handler params {params:?}");
if let Some(delay_ms) = params.delay_ms {
let delay = Duration::from_millis(delay_ms);
info!("sleeping {delay:?}");
tokio::time::sleep(delay).await;
}
let limit = match params.num_unprocessed {
Some(0) => None,
Some(limit) => Some(limit),
None => Some(10),
};
let mut ids = None;
match label_unprocessed(&nm, &pool, false, limit, "tag:unprocessed").await {
Ok(i) => ids = Some(i),
Err(err) => error!("Failed to label_unprocessed: {err:?}"),
};
connection_tracker
.lock()
.await
.send_message_all(WebsocketMessage::RefreshMessages)
.await;
if let Some(ids) = ids {
format!("{ids:?}")
} else {
"refresh triggered".to_string()
}
}
async fn watch_new(
nm: Notmuch,
pool: PgPool,
conn_tracker: Arc<Mutex<ConnectionTracker>>,
poll_time: Duration,
) -> Result<(), async_graphql::Error> {
async fn watch_new_iteration(
nm: &Notmuch,
pool: &PgPool,
conn_tracker: Arc<Mutex<ConnectionTracker>>,
old_ids: &[String],
) -> Result<Vec<String>, async_graphql::Error> {
let ids = compute_catchup_ids(&nm, &pool, "is:unread").await?;
info!("old_ids: {} ids: {}", old_ids.len(), ids.len());
if old_ids != ids {
label_unprocessed(&nm, &pool, false, Some(100), "tag:unprocessed").await?;
conn_tracker
.lock()
.await
.send_message_all(WebsocketMessage::RefreshMessages)
.await
}
Ok(ids)
}
let mut old_ids = Vec::new();
loop {
old_ids = match watch_new_iteration(&nm, &pool, conn_tracker.clone(), &old_ids).await {
Ok(old_ids) => old_ids,
Err(err) => {
error!("watch_new_iteration failed: {err:?}");
continue;
}
};
tokio::time::sleep(poll_time).await;
}
}
#[derive(Clone)]
struct AppState {
nm: Notmuch,
pool: PgPool,
connection_tracker: Arc<Mutex<ConnectionTracker>>,
}
#[derive(Parser)]
#[command(version, about, long_about = None)]
struct Cli {
#[arg(short, long, default_value = "0.0.0.0:9345")]
addr: SocketAddr,
newsreader_database_url: String,
newsreader_tantivy_db_path: String,
slurp_cache_path: String,
}
#[tokio::main]
async fn main() -> Result<(), Box<dyn Error>> {
let cli = Cli::parse();
let _guard = xtracing::init(env!("CARGO_BIN_NAME"))?;
build_info::build_info!(fn bi);
info!("Build Info: {}", letterbox_shared::build_version(bi));
if !std::fs::exists(&cli.slurp_cache_path)? {
info!("Creating slurp cache @ '{}'", &cli.slurp_cache_path);
std::fs::create_dir_all(&cli.slurp_cache_path)?;
}
let pool = PgPool::connect(&cli.newsreader_database_url).await?;
let nm = Notmuch::default();
sqlx::migrate!("./migrations").run(&pool).await?;
#[cfg(feature = "tantivy")]
let tantivy_conn = TantivyConnection::new(&cli.newsreader_tantivy_db_path)?;
let cacher = FilesystemCacher::new(&cli.slurp_cache_path)?;
let schema = Schema::build(QueryRoot, MutationRoot, SubscriptionRoot)
.data(nm.clone())
.data(cacher)
.data(pool.clone());
let schema = schema.extension(extensions::Logger).finish();
let connection_tracker = Arc::new(Mutex::new(ConnectionTracker::default()));
let ct = Arc::clone(&connection_tracker);
let poll_time = Duration::from_secs(60);
let _h = tokio::spawn(watch_new(nm.clone(), pool.clone(), ct, poll_time));
let api_routes = Router::new()
.route(
"/download/attachment/{id}/{idx}/{*rest}",
get(download_attachment),
)
.route("/view/attachment/{id}/{idx}/{*rest}", get(view_attachment))
.route("/original/{id}", get(view_original))
.route("/cid/{id}/{cid}", get(view_cid))
.route("/ws", any(start_ws))
.route_service("/graphql/ws", GraphQLSubscription::new(schema.clone()))
.route(
"/graphql/",
get(graphiql).post_service(GraphQL::new(schema.clone())),
);
let notification_routes = Router::new()
.route("/mail", post(send_refresh_websocket_handler))
.route("/news", post(send_refresh_websocket_handler));
let app = Router::new()
.nest("/api", api_routes)
.nest("/notification", notification_routes)
.with_state(AppState {
nm,
pool,
connection_tracker,
})
.layer(
TraceLayer::new_for_http()
.make_span_with(DefaultMakeSpan::default().include_headers(true)),
);
let listener = TcpListener::bind(cli.addr).await.unwrap();
tracing::info!("listening on {}", listener.local_addr().unwrap());
axum::serve(
listener,
app.into_make_service_with_connect_info::<SocketAddr>(),
)
.await
.unwrap();
Ok(())
}

172
server/src/bin/server.rs Normal file
View File

@@ -0,0 +1,172 @@
#[macro_use]
extern crate rocket;
use std::{error::Error, io::Cursor, str::FromStr};
use glog::Flags;
use notmuch::{Notmuch, NotmuchError, ThreadSet};
use rocket::{
http::{ContentType, Header},
request::Request,
response::{Debug, Responder},
serde::json::Json,
Response, State,
};
use rocket_cors::{AllowedHeaders, AllowedOrigins};
use server::{error::ServerError, nm::threadset_to_messages};
use shared::Message;
#[get("/")]
fn hello() -> &'static str {
"Hello, world!"
}
#[get("/refresh")]
async fn refresh(nm: &State<Notmuch>) -> Result<Json<String>, Debug<NotmuchError>> {
Ok(Json(String::from_utf8_lossy(&nm.new()?).to_string()))
}
#[get("/search")]
async fn search_all(
nm: &State<Notmuch>,
) -> Result<Json<shared::SearchResult>, Debug<NotmuchError>> {
search(nm, "*", None, None).await
}
#[get("/search/<query>?<page>&<results_per_page>")]
async fn search(
nm: &State<Notmuch>,
query: &str,
page: Option<usize>,
results_per_page: Option<usize>,
) -> Result<Json<shared::SearchResult>, Debug<NotmuchError>> {
let page = page.unwrap_or(0);
let results_per_page = results_per_page.unwrap_or(10);
let query = urlencoding::decode(query).map_err(NotmuchError::from)?;
info!(" search '{query}'");
let res = shared::SearchResult {
summary: nm.search(&query, page * results_per_page, results_per_page)?,
query: query.to_string(),
page,
results_per_page,
total: nm.count(&query)?,
};
Ok(Json(res))
}
#[get("/show/<query>/pretty")]
async fn show_pretty(
nm: &State<Notmuch>,
query: &str,
) -> Result<Json<Vec<Message>>, Debug<ServerError>> {
let query = urlencoding::decode(query).map_err(|e| ServerError::from(NotmuchError::from(e)))?;
let res = threadset_to_messages(nm.show(&query).map_err(ServerError::from)?)?;
Ok(Json(res))
}
#[get("/show/<query>")]
async fn show(nm: &State<Notmuch>, query: &str) -> Result<Json<ThreadSet>, Debug<NotmuchError>> {
let query = urlencoding::decode(query).map_err(NotmuchError::from)?;
let res = nm.show(&query)?;
Ok(Json(res))
}
struct PartResponder {
bytes: Vec<u8>,
filename: Option<String>,
}
impl<'r, 'o: 'r> Responder<'r, 'o> for PartResponder {
fn respond_to(self, _: &'r Request<'_>) -> rocket::response::Result<'o> {
let mut resp = Response::build();
if let Some(filename) = self.filename {
info!("filename {:?}", filename);
resp.header(Header::new(
"Content-Disposition",
format!(r#"attachment; filename="{}""#, filename),
))
.header(ContentType::Binary);
}
resp.sized_body(self.bytes.len(), Cursor::new(self.bytes))
.ok()
}
}
#[get("/original/<id>/part/<part>")]
async fn original_part(
nm: &State<Notmuch>,
id: &str,
part: usize,
) -> Result<PartResponder, Debug<NotmuchError>> {
let mid = if id.starts_with("id:") {
id.to_string()
} else {
format!("id:{}", id)
};
let meta = nm.show_part(&mid, part)?;
let res = nm.show_original_part(&mid, part)?;
Ok(PartResponder {
bytes: res,
filename: meta.filename,
})
}
#[get("/original/<id>")]
async fn original(
nm: &State<Notmuch>,
id: &str,
) -> Result<(ContentType, Vec<u8>), Debug<NotmuchError>> {
let mid = if id.starts_with("id:") {
id.to_string()
} else {
format!("id:{}", id)
};
let res = nm.show_original(&mid)?;
Ok((ContentType::Plain, res))
}
#[rocket::main]
async fn main() -> Result<(), Box<dyn Error>> {
glog::new()
.init(Flags {
colorlogtostderr: true,
//alsologtostderr: true, // use logtostderr to only write to stderr and not to files
logtostderr: true,
..Default::default()
})
.unwrap();
let allowed_origins = AllowedOrigins::all();
let cors = rocket_cors::CorsOptions {
allowed_origins,
allowed_methods: vec!["Get"]
.into_iter()
.map(|s| FromStr::from_str(s).unwrap())
.collect(),
allowed_headers: AllowedHeaders::some(&["Authorization", "Accept"]),
allow_credentials: true,
..Default::default()
}
.to_cors()?;
let _ = rocket::build()
.mount(
"/",
routes![
original_part,
original,
hello,
refresh,
search_all,
search,
show_pretty,
show
],
)
.attach(cors)
.manage(Notmuch::default())
//.manage(Notmuch::with_config("../notmuch/testdata/notmuch.config"))
.launch()
.await?;
Ok(())
}

View File

@@ -1,39 +0,0 @@
use std::error::Error;
use clap::Parser;
use letterbox_notmuch::Notmuch;
use letterbox_server::nm::label_unprocessed;
use sqlx::postgres::PgPool;
use tracing::info;
#[derive(Parser)]
#[command(version, about, long_about = None)]
struct Cli {
#[arg(short, long)]
newsreader_database_url: String,
#[arg(short, long, default_value = "10")]
/// Set to 0 to process all matches
messages_to_process: usize,
#[arg(short, long, default_value = "false")]
execute: bool,
/// Process messages matching this notmuch query
#[arg(short, long, default_value = "tag:unprocessed")]
query: String,
}
#[tokio::main]
async fn main() -> Result<(), Box<dyn Error>> {
let cli = Cli::parse();
let _guard = xtracing::init(env!("CARGO_BIN_NAME"))?;
build_info::build_info!(fn bi);
info!("Build Info: {}", letterbox_shared::build_version(bi));
let pool = PgPool::connect(&cli.newsreader_database_url).await?;
let nm = Notmuch::default();
let limit = if cli.messages_to_process > 0 {
Some(cli.messages_to_process)
} else {
None
};
label_unprocessed(&nm, &pool, !cli.execute, limit, &cli.query).await?;
Ok(())
}

File diff suppressed because it is too large Load Diff

View File

@@ -1,7 +0,0 @@
use serde::Deserialize;
#[derive(Deserialize)]
pub struct Config {
pub newsreader_database_url: String,
pub newsreader_tantivy_db_path: String,
pub slurp_cache_path: String,
}

File diff suppressed because it is too large Load Diff

View File

@@ -1,48 +1,9 @@
use std::{convert::Infallible, str::Utf8Error, string::FromUtf8Error};
use mailparse::MailParseError;
#[cfg(feature = "tantivy")]
use tantivy::{query::QueryParserError, TantivyError};
use thiserror::Error;
use crate::TransformError;
#[derive(Error, Debug)]
pub enum ServerError {
#[error("notmuch: {0}")]
NotmuchError(#[from] letterbox_notmuch::NotmuchError),
#[error("notmuch")]
NotmuchError(#[from] notmuch::NotmuchError),
#[error("flatten")]
FlattenError,
#[error("mail parse error: {0}")]
MailParseError(#[from] MailParseError),
#[error("IO error: {0}")]
IoError(#[from] std::io::Error),
#[error("attachement not found")]
PartNotFound,
#[error("sqlx error: {0}")]
SQLXError(#[from] sqlx::Error),
#[error("html transform error: {0}")]
TransformError(#[from] TransformError),
#[error("UTF8 error: {0}")]
Utf8Error(#[from] Utf8Error),
#[error("FromUTF8 error: {0}")]
FromUtf8Error(#[from] FromUtf8Error),
#[error("error: {0}")]
StringError(String),
#[error("invalid url: {0}")]
UrlParseError(#[from] url::ParseError),
#[cfg(feature = "tantivy")]
#[error("tantivy error: {0}")]
TantivyError(#[from] TantivyError),
#[cfg(feature = "tantivy")]
#[error("tantivy query parse error: {0}")]
QueryParseError(#[from] QueryParserError),
#[error("impossible: {0}")]
InfaillibleError(#[from] Infallible),
#[error("askama error: {0}")]
AskamaError(#[from] askama::Error),
#[error("xml error: {0}")]
XmlError(#[from] quick_xml::Error),
#[error("xml encoding error: {0}")]
XmlEncodingError(#[from] quick_xml::encoding::EncodingError),
}

View File

@@ -1,708 +0,0 @@
use std::{fmt, str::FromStr};
use async_graphql::{
connection::{self, Connection, Edge, OpaqueCursor},
futures_util::Stream,
Context, Enum, Error, FieldResult, InputObject, Object, Schema, SimpleObject, Subscription,
Union,
};
use cacher::FilesystemCacher;
use futures::stream;
use letterbox_notmuch::Notmuch;
use serde::{Deserialize, Serialize};
use sqlx::postgres::PgPool;
use tokio::join;
use tracing::{info, instrument};
#[cfg(feature = "tantivy")]
use crate::tantivy::TantivyConnection;
use crate::{newsreader, nm, nm::label_unprocessed, Query};
/// # Number of seconds since the Epoch
pub type UnixTime = isize;
/// # Thread ID, sans "thread:"
pub type ThreadId = String;
#[derive(Debug, Enum, Copy, Clone, Eq, PartialEq)]
pub enum Corpus {
Notmuch,
Newsreader,
Tantivy,
}
impl FromStr for Corpus {
type Err = String;
fn from_str(s: &str) -> Result<Self, Self::Err> {
Ok(match s {
"notmuch" => Corpus::Notmuch,
"newsreader" => Corpus::Newsreader,
"tantivy" => Corpus::Tantivy,
s => return Err(format!("unknown corpus: '{s}'")),
})
}
}
// TODO: add is_read field and remove all use of 'tag:unread'
#[derive(Debug, SimpleObject)]
pub struct ThreadSummary {
pub thread: ThreadId,
pub timestamp: UnixTime,
/// user-friendly timestamp
pub date_relative: String,
/// number of matched messages
pub matched: isize,
/// total messages in thread
pub total: isize,
/// comma-separated names with | between matched and unmatched
pub authors: String,
pub subject: String,
pub tags: Vec<String>,
pub corpus: Corpus,
}
#[derive(Debug, Union)]
pub enum Thread {
Email(EmailThread),
News(NewsPost),
}
#[derive(Debug, SimpleObject)]
pub struct NewsPost {
pub thread_id: String,
pub is_read: bool,
pub slug: String,
pub site: String,
pub title: String,
pub body: String,
pub url: String,
pub timestamp: i64,
}
#[derive(Debug, SimpleObject)]
pub struct EmailThread {
pub thread_id: String,
pub subject: String,
pub messages: Vec<Message>,
}
#[derive(Debug, SimpleObject)]
pub struct Message {
// Message-ID for message, prepend `id:<id>` to search in notmuch
pub id: String,
// First From header found in email
pub from: Option<Email>,
// All To headers found in email
pub to: Vec<Email>,
// All CC headers found in email
pub cc: Vec<Email>,
// X-Original-To header found in email
pub x_original_to: Option<Email>,
// Delivered-To header found in email
pub delivered_to: Option<Email>,
// First Subject header found in email
pub subject: Option<String>,
// Parsed Date header, if found and valid
pub timestamp: Option<i64>,
// Headers
pub headers: Vec<Header>,
// The body contents
pub body: Body,
// On disk location of message
pub path: String,
pub attachments: Vec<Attachment>,
pub tags: Vec<String>,
}
// Content-Type: image/jpeg; name="PXL_20231125_204826860.jpg"
// Content-Disposition: attachment; filename="PXL_20231125_204826860.jpg"
// Content-Transfer-Encoding: base64
// Content-ID: <f_lponoluo1>
// X-Attachment-Id: f_lponoluo1
#[derive(Default, Debug, SimpleObject)]
pub struct Attachment {
pub id: String,
pub idx: String,
pub filename: Option<String>,
pub size: usize,
pub content_type: Option<String>,
pub content_id: Option<String>,
pub disposition: DispositionType,
pub bytes: Vec<u8>,
}
#[derive(Debug, Clone, Eq, PartialEq)]
pub struct Disposition {
pub r#type: DispositionType,
pub filename: Option<String>,
pub size: Option<usize>,
}
#[derive(Debug, Enum, Copy, Clone, Eq, PartialEq)]
pub enum DispositionType {
Inline,
Attachment,
}
impl From<mailparse::DispositionType> for DispositionType {
fn from(value: mailparse::DispositionType) -> Self {
match value {
mailparse::DispositionType::Inline => DispositionType::Inline,
mailparse::DispositionType::Attachment => DispositionType::Attachment,
dt => panic!("unhandled DispositionType {dt:?}"),
}
}
}
impl Default for DispositionType {
fn default() -> Self {
DispositionType::Attachment
}
}
#[derive(Debug, SimpleObject)]
pub struct Header {
pub key: String,
pub value: String,
}
#[derive(Debug)]
pub struct UnhandledContentType {
pub text: String,
pub content_tree: String,
}
#[Object]
impl UnhandledContentType {
async fn contents(&self) -> &str {
&self.text
}
async fn content_tree(&self) -> &str {
&self.content_tree
}
}
#[derive(Debug)]
pub struct PlainText {
pub text: String,
pub content_tree: String,
}
#[Object]
impl PlainText {
async fn contents(&self) -> &str {
&self.text
}
async fn content_tree(&self) -> &str {
&self.content_tree
}
}
#[derive(Debug)]
pub struct Html {
pub html: String,
pub content_tree: String,
}
#[Object]
impl Html {
async fn contents(&self) -> &str {
&self.html
}
async fn content_tree(&self) -> &str {
&self.content_tree
}
async fn headers(&self) -> Vec<Header> {
Vec::new()
}
}
#[derive(Debug, Union)]
pub enum Body {
UnhandledContentType(UnhandledContentType),
PlainText(PlainText),
Html(Html),
}
impl Body {
pub fn html(html: String) -> Body {
Body::Html(Html {
html,
content_tree: "".to_string(),
})
}
pub fn text(text: String) -> Body {
Body::PlainText(PlainText {
text,
content_tree: "".to_string(),
})
}
pub fn to_html(&self) -> Option<String> {
match self {
Body::Html(h) => Some(h.html.clone()),
Body::PlainText(p) => Some(format!("<pre>{}</pre>", html_escape::encode_text(&p.text))),
Body::UnhandledContentType(u) => Some(format!("<pre>{}</pre>", html_escape::encode_text(&u.text))),
}
}
pub fn to_html_content_tree(&self) -> Option<String> {
match self {
Body::Html(h) => Some(h.content_tree.clone()),
Body::PlainText(p) => Some(p.content_tree.clone()),
Body::UnhandledContentType(u) => Some(u.content_tree.clone()),
}
}
}
#[derive(Debug, SimpleObject)]
pub struct Email {
pub name: Option<String>,
pub addr: Option<String>,
pub photo_url: Option<String>,
}
impl fmt::Display for Email {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> Result<(), std::fmt::Error> {
match (&self.name, &self.addr) {
(Some(name), Some(addr)) => write!(f, "{name} <{addr}>")?,
(Some(name), None) => write!(f, "{name}")?,
(None, Some(addr)) => write!(f, "{addr}")?,
(None, None) => write!(f, "<UNKNOWN>")?,
}
Ok(())
}
}
#[derive(SimpleObject)]
pub struct Tag {
pub name: String,
pub fg_color: String,
pub bg_color: String,
pub unread: usize,
}
#[derive(Serialize, Deserialize, Debug, InputObject)]
struct SearchCursor {
newsreader_offset: i32,
notmuch_offset: i32,
#[cfg(feature = "tantivy")]
tantivy_offset: i32,
}
fn request_id() -> String {
let now = std::time::SystemTime::now();
let nanos = now
.duration_since(std::time::SystemTime::UNIX_EPOCH)
.unwrap_or_default()
.as_nanos();
format!("{nanos:x}")
}
pub struct QueryRoot;
#[Object]
impl QueryRoot {
async fn version<'ctx>(&self, _ctx: &Context<'ctx>) -> Result<String, Error> {
build_info::build_info!(fn bi);
Ok(letterbox_shared::build_version(bi))
}
#[instrument(skip_all, fields(query=query, rid=request_id()))]
async fn count<'ctx>(&self, ctx: &Context<'ctx>, query: String) -> Result<usize, Error> {
let nm = ctx.data_unchecked::<Notmuch>();
let pool = ctx.data_unchecked::<PgPool>();
#[cfg(feature = "tantivy")]
let tantivy = ctx.data_unchecked::<TantivyConnection>();
let newsreader_query: Query = query.parse()?;
let newsreader_count = newsreader::count(pool, &newsreader_query).await?;
let notmuch_count = nm::count(nm, &newsreader_query).await?;
#[cfg(feature = "tantivy")]
let tantivy_count = tantivy.count(&newsreader_query).await?;
#[cfg(not(feature = "tantivy"))]
let tantivy_count = 0;
let total = newsreader_count + notmuch_count + tantivy_count;
info!("count {newsreader_query:?} newsreader count {newsreader_count} notmuch count {notmuch_count} tantivy count {tantivy_count} total {total}");
Ok(total)
}
#[instrument(skip_all, fields(query=query, rid=request_id()))]
async fn catchup<'ctx>(
&self,
ctx: &Context<'ctx>,
query: String,
) -> Result<Vec<String>, Error> {
let nm = ctx.data_unchecked::<Notmuch>();
let pool = ctx.data_unchecked::<PgPool>();
compute_catchup_ids(nm, pool, &query).await
}
// TODO: this function doesn't get parallelism, possibly because notmuch is sync and blocks,
// rewrite that with tokio::process:Command
#[instrument(skip_all, fields(query=query, rid=request_id()))]
async fn search<'ctx>(
&self,
ctx: &Context<'ctx>,
after: Option<String>,
before: Option<String>,
first: Option<i32>,
last: Option<i32>,
query: String,
) -> Result<Connection<OpaqueCursor<SearchCursor>, ThreadSummary>, Error> {
info!("search({after:?} {before:?} {first:?} {last:?} {query:?})",);
let nm = ctx.data_unchecked::<Notmuch>();
let pool = ctx.data_unchecked::<PgPool>();
#[cfg(feature = "tantivy")]
let tantivy = ctx.data_unchecked::<TantivyConnection>();
Ok(connection::query(
after,
before,
first,
last,
|after: Option<OpaqueCursor<SearchCursor>>,
before: Option<OpaqueCursor<SearchCursor>>,
first: Option<usize>,
last: Option<usize>| async move {
info!(
"search(after {:?} before {:?} first {first:?} last {last:?} query: {query:?})",
after.as_ref().map(|v| &v.0),
before.as_ref().map(|v| &v.0)
);
let newsreader_after = after.as_ref().map(|sc| sc.newsreader_offset);
let notmuch_after = after.as_ref().map(|sc| sc.notmuch_offset);
#[cfg(feature = "tantivy")]
let tantivy_after = after.as_ref().map(|sc| sc.tantivy_offset);
let newsreader_before = before.as_ref().map(|sc| sc.newsreader_offset);
let notmuch_before = before.as_ref().map(|sc| sc.notmuch_offset);
#[cfg(feature = "tantivy")]
let tantivy_before = before.as_ref().map(|sc| sc.tantivy_offset);
let first = first.map(|v| v as i32);
let last = last.map(|v| v as i32);
let query: Query = query.parse()?;
info!("newsreader_query {query:?}");
let newsreader_fut = newsreader_search(
pool,
newsreader_after,
newsreader_before,
first,
last,
&query,
);
let notmuch_fut =
notmuch_search(nm, notmuch_after, notmuch_before, first, last, &query);
#[cfg(feature = "tantivy")]
let tantivy_fut = tantivy_search(
tantivy,
pool,
tantivy_after,
tantivy_before,
first,
last,
&query,
);
#[cfg(not(feature = "tantivy"))]
let tantivy_fut =
async { Ok::<Vec<ThreadSummaryCursor>, async_graphql::Error>(Vec::new()) };
let (newsreader_results, notmuch_results, tantivy_results) =
join!(newsreader_fut, notmuch_fut, tantivy_fut);
let newsreader_results = newsreader_results?;
let notmuch_results = notmuch_results?;
let tantivy_results = tantivy_results?;
info!(
"newsreader_results ({}) notmuch_results ({}) tantivy_results ({})",
newsreader_results.len(),
notmuch_results.len(),
tantivy_results.len()
);
let mut results: Vec<_> = newsreader_results
.into_iter()
.chain(notmuch_results)
.chain(tantivy_results)
.collect();
// The leading '-' is to reverse sort
results.sort_by_key(|item| match item {
ThreadSummaryCursor::Newsreader(_, ts) => -ts.timestamp,
ThreadSummaryCursor::Notmuch(_, ts) => -ts.timestamp,
#[cfg(feature = "tantivy")]
ThreadSummaryCursor::Tantivy(_, ts) => -ts.timestamp,
});
let mut has_next_page = before.is_some();
if let Some(first) = first {
let first = first as usize;
if results.len() > first {
has_next_page = true;
results.truncate(first);
}
}
let mut has_previous_page = after.is_some();
if let Some(last) = last {
let last = last as usize;
if results.len() > last {
has_previous_page = true;
results.truncate(last);
}
}
let mut connection = Connection::new(has_previous_page, has_next_page);
// Set starting offset as the value from cursor to preserve state if no results from a corpus survived the truncation
let mut newsreader_offset =
after.as_ref().map(|sc| sc.newsreader_offset).unwrap_or(0);
let mut notmuch_offset = after.as_ref().map(|sc| sc.notmuch_offset).unwrap_or(0);
#[cfg(feature = "tantivy")]
let tantivy_offset = after.as_ref().map(|sc| sc.tantivy_offset).unwrap_or(0);
info!(
"newsreader_offset ({}) notmuch_offset ({})",
newsreader_offset, notmuch_offset,
);
connection.edges.extend(results.into_iter().map(|item| {
let thread_summary;
match item {
ThreadSummaryCursor::Newsreader(offset, ts) => {
thread_summary = ts;
newsreader_offset = offset;
}
ThreadSummaryCursor::Notmuch(offset, ts) => {
thread_summary = ts;
notmuch_offset = offset;
}
#[cfg(feature = "tantivy")]
ThreadSummaryCursor::Tantivy(offset, ts) => {
thread_summary = ts;
tantivy_offset = offset;
}
}
let cur = OpaqueCursor(SearchCursor {
newsreader_offset,
notmuch_offset,
#[cfg(feature = "tantivy")]
tantivy_offset,
});
Edge::new(cur, thread_summary)
}));
Ok::<_, async_graphql::Error>(connection)
},
)
.await?)
}
#[instrument(skip_all, fields(rid=request_id()))]
async fn tags<'ctx>(&self, ctx: &Context<'ctx>) -> FieldResult<Vec<Tag>> {
let nm = ctx.data_unchecked::<Notmuch>();
let pool = ctx.data_unchecked::<PgPool>();
let needs_unread = ctx.look_ahead().field("unread").exists();
let mut tags = newsreader::tags(pool, needs_unread).await?;
tags.append(&mut nm::tags(nm, needs_unread)?);
Ok(tags)
}
#[instrument(skip_all, fields(thread_id=thread_id, rid=request_id()))]
async fn thread<'ctx>(&self, ctx: &Context<'ctx>, thread_id: String) -> Result<Thread, Error> {
let nm = ctx.data_unchecked::<Notmuch>();
let cacher = ctx.data_unchecked::<FilesystemCacher>();
let pool = ctx.data_unchecked::<PgPool>();
let debug_content_tree = ctx
.look_ahead()
.field("messages")
.field("body")
.field("contentTree")
.exists();
if newsreader::is_newsreader_thread(&thread_id) {
Ok(newsreader::thread(cacher, pool, thread_id).await?)
} else {
Ok(nm::thread(nm, pool, thread_id, debug_content_tree).await?)
}
}
}
#[derive(Debug)]
enum ThreadSummaryCursor {
Newsreader(i32, ThreadSummary),
Notmuch(i32, ThreadSummary),
#[cfg(feature = "tantivy")]
Tantivy(i32, ThreadSummary),
}
async fn newsreader_search(
pool: &PgPool,
after: Option<i32>,
before: Option<i32>,
first: Option<i32>,
last: Option<i32>,
query: &Query,
) -> Result<Vec<ThreadSummaryCursor>, async_graphql::Error> {
Ok(newsreader::search(pool, after, before, first, last, &query)
.await?
.into_iter()
.map(|(cur, ts)| ThreadSummaryCursor::Newsreader(cur, ts))
.collect())
}
async fn notmuch_search(
nm: &Notmuch,
after: Option<i32>,
before: Option<i32>,
first: Option<i32>,
last: Option<i32>,
query: &Query,
) -> Result<Vec<ThreadSummaryCursor>, async_graphql::Error> {
Ok(nm::search(nm, after, before, first, last, &query)
.await?
.into_iter()
.map(|(cur, ts)| ThreadSummaryCursor::Notmuch(cur, ts))
.collect())
}
#[cfg(feature = "tantivy")]
async fn tantivy_search(
tantivy: &TantivyConnection,
pool: &PgPool,
after: Option<i32>,
before: Option<i32>,
first: Option<i32>,
last: Option<i32>,
query: &Query,
) -> Result<Vec<ThreadSummaryCursor>, async_graphql::Error> {
Ok(tantivy
.search(pool, after, before, first, last, &query)
.await?
.into_iter()
.map(|(cur, ts)| ThreadSummaryCursor::Tantivy(cur, ts))
.collect())
}
pub struct MutationRoot;
#[Object]
impl MutationRoot {
#[instrument(skip_all, fields(query=query, unread=unread, rid=request_id()))]
async fn set_read_status<'ctx>(
&self,
ctx: &Context<'ctx>,
query: String,
unread: bool,
) -> Result<bool, Error> {
let nm = ctx.data_unchecked::<Notmuch>();
let pool = ctx.data_unchecked::<PgPool>();
#[cfg(feature = "tantivy")]
let tantivy = ctx.data_unchecked::<TantivyConnection>();
let query: Query = query.parse()?;
newsreader::set_read_status(pool, &query, unread).await?;
#[cfg(feature = "tantivy")]
tantivy.reindex_thread(pool, &query).await?;
nm::set_read_status(nm, &query, unread).await?;
Ok(true)
}
#[instrument(skip_all, fields(query=query, tag=tag, rid=request_id()))]
async fn tag_add<'ctx>(
&self,
ctx: &Context<'ctx>,
query: String,
tag: String,
) -> Result<bool, Error> {
let nm = ctx.data_unchecked::<Notmuch>();
info!("tag_add({tag}, {query})");
nm.tag_add(&tag, &query)?;
Ok(true)
}
#[instrument(skip_all, fields(query=query, tag=tag, rid=request_id()))]
async fn tag_remove<'ctx>(
&self,
ctx: &Context<'ctx>,
query: String,
tag: String,
) -> Result<bool, Error> {
let nm = ctx.data_unchecked::<Notmuch>();
info!("tag_remove({tag}, {query})");
nm.tag_remove(&tag, &query)?;
Ok(true)
}
/// Drop and recreate tantivy index. Warning this is slow
#[cfg(feature = "tantivy")]
async fn drop_and_load_index<'ctx>(&self, ctx: &Context<'ctx>) -> Result<bool, Error> {
let tantivy = ctx.data_unchecked::<TantivyConnection>();
let pool = ctx.data_unchecked::<PgPool>();
tantivy.drop_and_load_index()?;
tantivy.reindex_all(pool).await?;
Ok(true)
}
#[instrument(skip_all, fields(rid=request_id()))]
async fn refresh<'ctx>(&self, ctx: &Context<'ctx>) -> Result<bool, Error> {
let nm = ctx.data_unchecked::<Notmuch>();
let cacher = ctx.data_unchecked::<FilesystemCacher>();
let pool = ctx.data_unchecked::<PgPool>();
info!("{}", String::from_utf8_lossy(&nm.new()?));
newsreader::refresh(pool, cacher).await?;
// Process email labels
label_unprocessed(&nm, &pool, false, Some(10), "tag:unprocessed").await?;
#[cfg(feature = "tantivy")]
{
let tantivy = ctx.data_unchecked::<TantivyConnection>();
// TODO: parallelize
tantivy.refresh(pool).await?;
}
Ok(true)
}
}
pub struct SubscriptionRoot;
#[Subscription]
impl SubscriptionRoot {
async fn values(&self, _ctx: &Context<'_>) -> Result<impl Stream<Item = usize>, Error> {
Ok(stream::iter(0..10))
}
}
pub type GraphqlSchema = Schema<QueryRoot, MutationRoot, SubscriptionRoot>;
#[instrument(skip_all, fields(query=query))]
pub async fn compute_catchup_ids(
nm: &Notmuch,
pool: &PgPool,
query: &str,
) -> Result<Vec<String>, Error> {
let query: Query = query.parse()?;
// TODO: implement optimized versions of fetching just IDs
let newsreader_fut = newsreader_search(pool, None, None, None, None, &query);
let notmuch_fut = notmuch_search(nm, None, None, None, None, &query);
let (newsreader_results, notmuch_results) = join!(newsreader_fut, notmuch_fut);
let newsreader_results = newsreader_results?;
let notmuch_results = notmuch_results?;
info!(
"newsreader_results ({}) notmuch_results ({})",
newsreader_results.len(),
notmuch_results.len(),
);
let mut results: Vec<_> = newsreader_results
.into_iter()
.chain(notmuch_results)
.collect();
// The leading '-' is to reverse sort
results.sort_by_key(|item| match item {
ThreadSummaryCursor::Newsreader(_, ts) => -ts.timestamp,
ThreadSummaryCursor::Notmuch(_, ts) => -ts.timestamp,
});
let ids = results
.into_iter()
.map(|r| match r {
ThreadSummaryCursor::Newsreader(_, ts) => ts.thread,
ThreadSummaryCursor::Notmuch(_, ts) => ts.thread,
})
.collect();
Ok(ids)
}

View File

@@ -1,963 +1,2 @@
pub mod config;
pub mod email_extract;
pub mod error;
pub mod graphql;
pub mod newsreader;
pub mod nm;
pub mod ws;
#[cfg(feature = "tantivy")]
pub mod tantivy;
use std::{
collections::{HashMap, HashSet},
convert::Infallible,
fmt,
str::FromStr,
sync::Arc,
};
use async_trait::async_trait;
use cacher::{Cacher, FilesystemCacher};
use css_inline::{CSSInliner, InlineError, InlineOptions};
pub use error::ServerError;
use linkify::{LinkFinder, LinkKind};
use lol_html::{
element, errors::RewritingError, html_content::ContentType, rewrite_str, text,
RewriteStrSettings,
};
use maplit::{hashmap, hashset};
use regex::Regex;
use reqwest::StatusCode;
use scraper::{Html, Selector};
use sqlx::types::time::PrimitiveDateTime;
use thiserror::Error;
use tracing::{debug, error, info, warn};
use url::Url;
use crate::{
graphql::{Corpus, ThreadSummary},
newsreader::is_newsreader_thread,
nm::is_notmuch_thread_or_id,
};
const NEWSREADER_TAG_PREFIX: &'static str = "News/";
const NEWSREADER_THREAD_PREFIX: &'static str = "news:";
// TODO: figure out how to use Cow
#[async_trait]
trait Transformer: Send + Sync {
fn should_run(&self, _addr: &Option<Url>, _html: &str) -> bool {
true
}
// TODO: should html be something like `html_escape` uses:
// <S: ?Sized + AsRef<str>>(text: &S) -> Cow<str>
async fn transform(&self, addr: &Option<Url>, html: &str) -> Result<String, TransformError>;
}
// TODO: how would we make this more generic to allow good implementations of Transformer outside
// of this module?
#[derive(Error, Debug)]
pub enum TransformError {
#[error("lol-html rewrite error: {0}")]
RewritingError(#[from] RewritingError),
#[error("css inline error: {0}")]
InlineError(#[from] InlineError),
#[error("failed to fetch url error: {0}")]
ReqwestError(#[from] reqwest::Error),
#[error("failed to parse HTML: {0}")]
HtmlParsingError(String),
#[error("got a retryable error code {0} for {1}")]
RetryableHttpStatusError(StatusCode, String),
}
struct SanitizeHtml<'a> {
cid_prefix: &'a str,
base_url: &'a Option<Url>,
}
#[async_trait]
impl<'a> Transformer for SanitizeHtml<'a> {
async fn transform(&self, _: &Option<Url>, html: &str) -> Result<String, TransformError> {
Ok(sanitize_html(html, self.cid_prefix, self.base_url)?)
}
}
struct EscapeHtml;
#[async_trait]
impl Transformer for EscapeHtml {
fn should_run(&self, _: &Option<Url>, html: &str) -> bool {
html.contains("&")
}
async fn transform(&self, _: &Option<Url>, html: &str) -> Result<String, TransformError> {
Ok(html_escape::decode_html_entities(html).to_string())
}
}
struct StripHtml;
#[async_trait]
impl Transformer for StripHtml {
fn should_run(&self, link: &Option<Url>, html: &str) -> bool {
debug!("StripHtml should_run {link:?} {}", html.contains("<"));
// Lame test
html.contains("<")
}
async fn transform(&self, link: &Option<Url>, html: &str) -> Result<String, TransformError> {
debug!("StripHtml {link:?}");
let mut text = String::new();
let element_content_handlers = vec![
element!("style", |el| {
el.remove();
Ok(())
}),
element!("script", |el| {
el.remove();
Ok(())
}),
];
let html = rewrite_str(
html,
RewriteStrSettings {
element_content_handlers,
..RewriteStrSettings::default()
},
)?;
let element_content_handlers = vec![text!("*", |t| {
text += t.as_str();
Ok(())
})];
let _ = rewrite_str(
&html,
RewriteStrSettings {
element_content_handlers,
..RewriteStrSettings::default()
},
)?;
let re = Regex::new(r"\s+").expect("failed to parse regex");
let text = re.replace_all(&text, " ").to_string();
Ok(text)
}
}
struct InlineStyle;
#[async_trait]
impl Transformer for InlineStyle {
async fn transform(&self, _: &Option<Url>, html: &str) -> Result<String, TransformError> {
let css = concat!(
"/* chrome-default.css */\n",
include_str!("chrome-default.css"),
//"\n/* mvp.css */\n",
//include_str!("mvp.css"),
//"\n/* Xinu Specific overrides */\n",
//include_str!("custom.css"),
);
let inline_opts = InlineOptions {
inline_style_tags: true,
keep_style_tags: false,
keep_link_tags: true,
base_url: None,
load_remote_stylesheets: true,
extra_css: Some(css.into()),
preallocate_node_capacity: 32,
..InlineOptions::default()
};
//info!("HTML:\n{html}");
Ok(match CSSInliner::new(inline_opts).inline(&html) {
Ok(inlined_html) => inlined_html,
Err(err) => {
error!("failed to inline CSS: {err}");
html.to_string()
}
})
}
}
/// Process images will extract any alt or title tags on images and place them as labels below said
/// image. It also handles data-src and data-cfsrc attributes
struct FrameImages;
#[async_trait]
impl Transformer for FrameImages {
async fn transform(&self, _: &Option<Url>, html: &str) -> Result<String, TransformError> {
Ok(rewrite_str(
html,
RewriteStrSettings {
element_content_handlers: vec![
element!("img[data-src]", |el| {
let src = el
.get_attribute("data-src")
.unwrap_or("https://placehold.co/600x400".to_string());
el.set_attribute("src", &src)?;
Ok(())
}),
element!("img[data-cfsrc]", |el| {
let src = el
.get_attribute("data-cfsrc")
.unwrap_or("https://placehold.co/600x400".to_string());
el.set_attribute("src", &src)?;
Ok(())
}),
element!("img[alt], img[title]", |el| {
let src = el
.get_attribute("src")
.unwrap_or("https://placehold.co/600x400".to_string());
let alt = el.get_attribute("alt");
let title = el.get_attribute("title");
let mut frags =
vec!["<figure>".to_string(), format!(r#"<img src="{src}">"#)];
alt.map(|t| {
if !t.is_empty() {
frags.push(format!("<figcaption>Alt: {t}</figcaption>"))
}
});
title.map(|t| {
if !t.is_empty() {
frags.push(format!("<figcaption>Title: {t}</figcaption>"))
}
});
frags.push("</figure>".to_string());
el.replace(&frags.join("\n"), ContentType::Html);
Ok(())
}),
],
..RewriteStrSettings::default()
},
)?)
}
}
struct AddOutlink;
#[async_trait]
impl Transformer for AddOutlink {
fn should_run(&self, link: &Option<Url>, html: &str) -> bool {
if let Some(link) = link {
link.scheme().starts_with("http") && !html.contains(link.as_str())
} else {
false
}
}
async fn transform(&self, link: &Option<Url>, html: &str) -> Result<String, TransformError> {
if let Some(link) = link {
Ok(format!(
r#"
{html}
<div><a href="{}">View on site</a></div>
"#,
link
))
} else {
Ok(html.to_string())
}
}
}
struct SlurpContents<'c> {
cacher: &'c FilesystemCacher,
inline_css: bool,
site_selectors: HashMap<String, Vec<Selector>>,
}
impl<'c> SlurpContents<'c> {
fn get_selectors(&self, link: &Url) -> Option<&[Selector]> {
for (host, selector) in self.site_selectors.iter() {
if link.host_str().map(|h| h.contains(host)).unwrap_or(false) {
return Some(&selector);
}
}
None
}
}
#[async_trait]
impl<'c> Transformer for SlurpContents<'c> {
fn should_run(&self, link: &Option<Url>, html: &str) -> bool {
debug!("SlurpContents should_run {link:?}");
let mut will_slurp = false;
if let Some(link) = link {
will_slurp = self.get_selectors(link).is_some();
}
if !will_slurp && self.inline_css {
return InlineStyle {}.should_run(link, html);
}
will_slurp
}
async fn transform(&self, link: &Option<Url>, html: &str) -> Result<String, TransformError> {
debug!("SlurpContents {link:?}");
let retryable_status: HashSet<StatusCode> = vec![
StatusCode::UNAUTHORIZED,
StatusCode::FORBIDDEN,
StatusCode::REQUEST_TIMEOUT,
StatusCode::TOO_MANY_REQUESTS,
]
.into_iter()
.collect();
if let Some(test_link) = link {
// If SlurpContents is configured for inline CSS, but no
// configuration found for this site, use the local InlineStyle
// transform.
if self.inline_css && self.get_selectors(test_link).is_none() {
debug!("local inline CSS for {link:?}");
return InlineStyle {}.transform(link, html).await;
}
}
let Some(link) = link else {
return Ok(html.to_string());
};
let Some(selectors) = self.get_selectors(&link) else {
return Ok(html.to_string());
};
let cacher = self.cacher;
let body = if let Some(body) = cacher.get(link.as_str()) {
String::from_utf8_lossy(&body).to_string()
} else {
let resp = reqwest::get(link.as_str()).await?;
let status = resp.status();
if status.is_server_error() {
error!("status error for {link}: {status}");
return Ok(html.to_string());
}
if retryable_status.contains(&status) {
error!("retryable error for {link}: {status}");
return Ok(html.to_string());
}
if !status.is_success() {
error!("unsuccessful for {link}: {status}");
return Ok(html.to_string());
}
let body = resp.text().await?;
cacher.set(link.as_str(), body.as_bytes());
body
};
let body = Arc::new(body);
let base_url = Some(link.clone());
let body = if self.inline_css {
debug!("inlining CSS for {link}");
let inner_body = Arc::clone(&body);
let res = tokio::task::spawn_blocking(move || {
let css = concat!(
"/* chrome-default.css */\n",
include_str!("chrome-default.css"),
"\n/* vars.css */\n",
include_str!("../static/vars.css"),
//"\n/* Xinu Specific overrides */\n",
//include_str!("custom.css"),
);
let res = CSSInliner::options()
.base_url(base_url)
.extra_css(Some(std::borrow::Cow::Borrowed(css)))
.build()
.inline(&inner_body);
match res {
Ok(inlined_html) => inlined_html,
Err(err) => {
error!("failed to inline remote CSS: {err}");
Arc::into_inner(inner_body).expect("failed to take body out of Arc")
}
}
})
.await;
match res {
Ok(inlined_html) => inlined_html,
Err(err) => {
error!("failed to spawn inline remote CSS: {err}");
Arc::into_inner(body).expect("failed to take body out of Arc")
}
}
} else {
debug!("using body as-is for {link:?}");
Arc::into_inner(body).expect("failed to take body out of Arc")
};
let doc = Html::parse_document(&body);
let mut results = Vec::new();
for selector in selectors {
for frag in doc.select(&selector) {
results.push(frag.html())
// TODO: figure out how to warn if there were no hits
//warn!("couldn't find '{:?}' in {}", selector, link);
}
}
Ok(results.join("<br>"))
}
}
pub fn linkify_html(text: &str) -> String {
let mut finder = LinkFinder::new();
let finder = finder.url_must_have_scheme(false).kinds(&[LinkKind::Url]);
let mut parts = Vec::new();
for span in finder.spans(text) {
// TODO(wathiede): use Cow<str>?
match span.kind() {
// Text as-is
None => parts.push(span.as_str().to_string()),
// Wrap in anchor tag
Some(LinkKind::Url) => {
let text = span.as_str();
let schema = if text.starts_with("http") {
""
} else {
"http://"
};
let a = format!(r#"<a href="{schema}{0}">{0}</a>"#, text);
parts.push(a);
}
_ => todo!("unhandled kind: {:?}", span.kind().unwrap()),
}
}
parts.join("")
}
// html contains the content to be cleaned, and cid_prefix is used to resolve mixed part image
// referrences
pub fn sanitize_html(
html: &str,
cid_prefix: &str,
base_url: &Option<Url>,
) -> Result<String, TransformError> {
let inline_opts = InlineOptions {
inline_style_tags: true,
keep_style_tags: true,
keep_link_tags: false,
base_url: None,
load_remote_stylesheets: false,
extra_css: None,
preallocate_node_capacity: 32,
..InlineOptions::default()
};
let html = match CSSInliner::new(inline_opts).inline(&html) {
Ok(inlined_html) => inlined_html,
Err(err) => {
error!("failed to inline CSS: {err}");
html.to_string()
}
};
let mut element_content_handlers = vec![
// Remove width and height attributes on elements
element!("[width],[height]", |el| {
el.remove_attribute("width");
el.remove_attribute("height");
Ok(())
}),
// Remove width and height values from inline styles
element!("[style]", |el| {
let style = el.get_attribute("style").unwrap();
let style = style
.split(";")
.filter(|s| {
let Some((k, _)) = s.split_once(':') else {
return true;
};
match k {
"width" | "max-width" | "min-width" | "height" | "max-height"
| "min-height" => false,
_ => true,
}
})
.collect::<Vec<_>>()
.join(";");
if let Err(e) = el.set_attribute("style", &style) {
error!("Failed to set style attribute: {e}");
}
Ok(())
}),
// Open links in new tab
element!("a[href]", |el| {
el.set_attribute("target", "_blank").unwrap();
Ok(())
}),
// Replace mixed part CID images with URL
element!("img[src]", |el| {
let src = el
.get_attribute("src")
.expect("src was required")
.replace("cid:", cid_prefix);
el.set_attribute("src", &src)?;
Ok(())
}),
// Only secure image URLs
element!("img[src]", |el| {
let src = el
.get_attribute("src")
.expect("src was required")
.replace("http:", "https:");
el.set_attribute("src", &src)?;
Ok(())
}),
// Add https to href with //<domain name>
element!("link[href]", |el| {
info!("found link[href] {el:?}");
let mut href = el.get_attribute("href").expect("href was required");
if href.starts_with("//") {
warn!("adding https to {href}");
href.insert_str(0, "https:");
}
el.set_attribute("href", &href)?;
Ok(())
}),
// Add https to src with //<domain name>
element!("style[src]", |el| {
let mut src = el.get_attribute("src").expect("src was required");
if src.starts_with("//") {
src.insert_str(0, "https:");
}
el.set_attribute("src", &src)?;
Ok(())
}),
];
if let Some(base_url) = base_url {
element_content_handlers.extend(vec![
// Make links with relative URLs absolute
element!("a[href]", |el| {
if let Some(Ok(href)) = el.get_attribute("href").map(|href| base_url.join(&href)) {
el.set_attribute("href", &href.as_str()).unwrap();
}
Ok(())
}),
// Make images with relative srcs absolute
element!("img[src]", |el| {
if let Some(Ok(src)) = el.get_attribute("src").map(|src| base_url.join(&src)) {
el.set_attribute("src", &src.as_str()).unwrap();
}
Ok(())
}),
]);
}
let html = rewrite_str(
&html,
RewriteStrSettings {
element_content_handlers,
..RewriteStrSettings::default()
},
)?;
// Default's don't allow style, but we want to preserve that.
// TODO: remove 'class' if rendering mails moves to a two phase process where abstract message
// types are collected, santized, and then grouped together as one big HTML doc
let attributes = hashset![
"align", "bgcolor", "class", "color", "height", "lang", "title", "width", "style",
];
let tags = hashset![
"a",
"abbr",
"acronym",
"area",
"article",
"aside",
"b",
"bdi",
"bdo",
"blockquote",
"br",
"caption",
"center",
"cite",
"code",
"col",
"colgroup",
"data",
"dd",
"del",
"details",
"dfn",
"div",
"dl",
"dt",
"em",
"figcaption",
"figure",
"footer",
"h1",
"h2",
"h3",
"h4",
"h5",
"h6",
"header",
"hgroup",
"hr",
"i",
"iframe", // wathiede
"img",
"ins",
"kbd",
"kbd",
"li",
"map",
"mark",
"nav",
"noscript", // wathiede
"ol",
"p",
"pre",
"q",
"rp",
"rt",
"rtc",
"ruby",
"s",
"samp",
"small",
"span",
"strike",
"strong",
"sub",
"summary",
"sup",
"table",
"tbody",
"td",
"th",
"thead",
"time",
"title", // wathiede
"tr",
"tt",
"u",
"ul",
"var",
"wbr",
];
let tag_attributes = hashmap![
"a" => hashset![
"href", "hreflang", "target",
],
"bdo" => hashset![
"dir"
],
"blockquote" => hashset![
"cite"
],
"col" => hashset![
"align", "char", "charoff", "span"
],
"colgroup" => hashset![
"align", "char", "charoff", "span"
],
"del" => hashset![
"cite", "datetime"
],
"hr" => hashset![
"align", "size", "width"
],
"iframe" => hashset![
"src", "allow", "allowfullscreen"
],
"img" => hashset![
"align", "alt", "height", "src", "width"
],
"ins" => hashset![
"cite", "datetime"
],
"ol" => hashset![
"start"
],
"q" => hashset![
"cite"
],
"table" => hashset![
"align", "border", "cellpadding", "cellspacing", "char", "charoff", "summary",
],
"tbody" => hashset![
"align", "char", "charoff"
],
"td" => hashset![
"align", "char", "charoff", "colspan", "headers", "rowspan"
],
"tfoot" => hashset![
"align", "char", "charoff"
],
"th" => hashset![
"align", "char", "charoff", "colspan", "headers", "rowspan", "scope"
],
"thead" => hashset![
"align", "char", "charoff"
],
"tr" => hashset![
"align", "char", "charoff"
],
];
let html = ammonia::Builder::default()
.tags(tags)
.tag_attributes(tag_attributes)
.generic_attributes(attributes)
.clean(&html)
.to_string();
Ok(html)
}
fn compute_offset_limit(
after: Option<i32>,
before: Option<i32>,
first: Option<i32>,
last: Option<i32>,
) -> (i32, i32) {
let default_page_size = 10000;
match (after, before, first, last) {
// Reasonable defaults
(None, None, None, None) => (0, default_page_size),
(None, None, Some(first), None) => (0, first),
(Some(after), None, None, None) => (after + 1, default_page_size),
(Some(after), None, Some(first), None) => (after + 1, first),
(None, Some(before), None, None) => (0.max(before - default_page_size), default_page_size),
(None, Some(before), None, Some(last)) => (0.max(before - last), last),
(None, None, None, Some(_)) => {
panic!("specifying last and no before doesn't make sense")
}
(None, None, Some(_), Some(_)) => {
panic!("specifying first and last doesn't make sense")
}
(None, Some(_), Some(_), _) => {
panic!("specifying before and first doesn't make sense")
}
(Some(_), Some(_), _, _) => {
panic!("specifying after and before doesn't make sense")
}
(Some(_), None, None, Some(_)) => {
panic!("specifying after and last doesn't make sense")
}
(Some(_), None, Some(_), Some(_)) => {
panic!("specifying after, first and last doesn't make sense")
}
}
}
#[derive(Debug, Default)]
pub struct Query {
pub unread_only: bool,
pub tags: Vec<String>,
pub uids: Vec<String>,
pub remainder: Vec<String>,
pub is_notmuch: bool,
pub is_newsreader: bool,
pub is_tantivy: bool,
pub corpus: Option<Corpus>,
}
impl fmt::Display for Query {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> Result<(), std::fmt::Error> {
if self.unread_only {
write!(f, "is:unread ")?;
}
for tag in &self.tags {
write!(f, "tag:{tag} ")?;
}
for uid in &self.uids {
write!(f, "id:{uid} ")?;
}
if self.is_notmuch {
write!(f, "is:mail ")?;
}
if self.is_newsreader {
write!(f, "is:newsreader ")?;
}
if self.is_newsreader {
write!(f, "is:news ")?;
}
match self.corpus {
Some(c) => write!(f, "corpus:{c:?}")?,
_ => (),
}
for rem in &self.remainder {
write!(f, "{rem} ")?;
}
Ok(())
}
}
impl Query {
// Converts the internal state of Query to something suitable for notmuch queries. Removes and
// letterbox specific '<key>:<value' tags
fn to_notmuch(&self) -> String {
let mut parts = Vec::new();
if !self.is_notmuch {
return String::new();
}
if self.unread_only {
parts.push("is:unread".to_string());
}
for tag in &self.tags {
parts.push(format!("tag:{tag}"));
}
for uid in &self.uids {
parts.push(uid.clone());
}
for r in &self.remainder {
// Rewrite "to:" to include ExtraTo:. ExtraTo: is configured in
// notmuch-config to index Delivered-To and X-Original-To headers.
if r.starts_with("to:") {
parts.push("(".to_string());
parts.push(r.to_string());
parts.push("OR".to_string());
parts.push(r.replace("to:", "ExtraTo:"));
parts.push(")".to_string());
} else {
parts.push(r.to_string());
}
}
parts.join(" ")
}
}
impl FromStr for Query {
type Err = Infallible;
fn from_str(s: &str) -> Result<Self, Self::Err> {
let mut unread_only = false;
let mut tags = Vec::new();
let mut uids = Vec::new();
let mut remainder = Vec::new();
let mut is_notmuch = false;
let mut is_newsreader = false;
let mut is_tantivy = false;
let mut corpus = None;
for word in s.split_whitespace() {
if word == "is:unread" {
unread_only = true
} else if word.starts_with("tag:") {
let t = &word["tag:".len()..];
// Per-address emails are faked as `tag:@<domain>/<username>`, rewrite to `to:` form
if t.starts_with('@') && t.contains('.') {
let t = match t.split_once('/') {
None => format!("to:{t}"),
Some((domain, user)) => format!("to:{user}{domain}"),
};
remainder.push(t);
} else {
tags.push(t.to_string());
};
/*
} else if word.starts_with("tag:") {
// Any tag that doesn't match site_prefix should explicitly set the site to something not in the
// database
site = Some(NON_EXISTENT_SITE_NAME.to_string());
*/
} else if word.starts_with("corpus:") {
let c = word["corpus:".len()..].to_string();
corpus = c.parse::<Corpus>().map(|c| Some(c)).unwrap_or_else(|e| {
warn!("Error parsing corpus '{c}': {e:?}");
None
});
} else if is_newsreader_thread(word) {
uids.push(word.to_string());
} else if is_notmuch_thread_or_id(word) {
uids.push(word.to_string());
} else if word == "is:mail" || word == "is:email" || word == "is:notmuch" {
is_notmuch = true;
} else if word == "is:news" {
is_newsreader = true;
} else if word == "is:newsreader" {
is_newsreader = true;
} else {
remainder.push(word.to_string());
}
}
// If we don't see any explicit filters for a corpus, flip them all on
if corpus.is_none() && !(is_notmuch || is_tantivy || is_newsreader) {
is_notmuch = true;
is_newsreader = true;
is_tantivy = true;
}
Ok(Query {
unread_only,
tags,
uids,
remainder,
is_notmuch,
is_newsreader,
is_tantivy,
corpus,
})
}
}
pub struct ThreadSummaryRecord {
pub site: Option<String>,
pub date: Option<PrimitiveDateTime>,
pub is_read: Option<bool>,
pub title: Option<String>,
pub uid: String,
pub name: Option<String>,
pub corpus: Corpus,
}
async fn thread_summary_from_row(r: ThreadSummaryRecord) -> ThreadSummary {
let site = r.site.unwrap_or("UNKOWN TAG".to_string());
let mut tags = vec![format!("{NEWSREADER_TAG_PREFIX}{site}")];
if !r.is_read.unwrap_or(true) {
tags.push("unread".to_string());
};
let mut title = r.title.unwrap_or("NO TITLE".to_string());
title = clean_title(&title).await.expect("failed to clean title");
ThreadSummary {
thread: format!("{NEWSREADER_THREAD_PREFIX}{}", r.uid),
timestamp: r
.date
.expect("post missing date")
.assume_utc()
.unix_timestamp() as isize,
date_relative: format!("{:?}", r.date),
//date_relative: "TODO date_relative".to_string(),
matched: 0,
total: 1,
authors: r.name.unwrap_or_else(|| site.clone()),
subject: title,
tags,
corpus: r.corpus,
}
}
async fn clean_title(title: &str) -> Result<String, ServerError> {
// Make title HTML so html parsers work
let mut title = format!("<html>{title}</html>");
let title_tranformers: Vec<Box<dyn Transformer>> =
vec![Box::new(EscapeHtml), Box::new(StripHtml)];
// Make title HTML so html parsers work
title = format!("<html>{title}</html>");
for t in title_tranformers.iter() {
if t.should_run(&None, &title) {
title = t.transform(&None, &title).await?;
}
}
Ok(title)
}
#[cfg(test)]
mod tests {
use super::{SanitizeHtml, Transformer};
#[tokio::test]
async fn strip_sizes() -> Result<(), Box<dyn std::error::Error>> {
let ss = SanitizeHtml {
cid_prefix: "",
base_url: &None,
};
let input = r#"<p width=16 height=16 style="color:blue;width:16px;height:16px;">This el has width and height attributes and inline styles</p>"#;
let want = r#"<p style="color:blue;">This el has width and height attributes and inline styles</p>"#;
let got = ss.transform(&None, input).await?;
assert_eq!(got, want);
Ok(())
}
}

View File

@@ -1,498 +0,0 @@
/* MVP.css v1.15 - https://github.com/andybrewer/mvp */
/* :root content stored in client side index.html */
html {
scroll-behavior: smooth;
}
@media (prefers-reduced-motion: reduce) {
html {
scroll-behavior: auto;
}
}
/* Layout */
article aside {
background: var(--color-secondary-accent);
border-left: 4px solid var(--color-secondary);
padding: 0.01rem 0.8rem;
}
body {
background: var(--color-bg);
color: var(--color-text);
font-family: var(--font-family);
line-height: var(--line-height);
margin: 0;
overflow-x: hidden;
padding: 0;
}
footer,
header,
main {
margin: 0 auto;
max-width: var(--width-content);
padding: 3rem 1rem;
}
hr {
background-color: var(--color-bg-secondary);
border: none;
height: 1px;
margin: 4rem 0;
width: 100%;
}
section {
display: flex;
flex-wrap: wrap;
justify-content: var(--justify-important);
}
section img,
article img {
max-width: 100%;
}
section pre {
overflow: auto;
}
section aside {
border: 1px solid var(--color-bg-secondary);
border-radius: var(--border-radius);
box-shadow: var(--box-shadow) var(--color-shadow);
margin: 1rem;
padding: 1.25rem;
width: var(--width-card);
}
section aside:hover {
box-shadow: var(--box-shadow) var(--color-bg-secondary);
}
[hidden] {
display: none;
}
/* Headers */
article header,
div header,
main header {
padding-top: 0;
}
header {
text-align: var(--justify-important);
}
header a b,
header a em,
header a i,
header a strong {
margin-left: 0.5rem;
margin-right: 0.5rem;
}
header nav img {
margin: 1rem 0;
}
section header {
padding-top: 0;
width: 100%;
}
/* Nav */
nav {
align-items: center;
display: flex;
font-weight: bold;
justify-content: space-between;
margin-bottom: 7rem;
}
nav ul {
list-style: none;
padding: 0;
}
nav ul li {
display: inline-block;
margin: 0 0.5rem;
position: relative;
text-align: left;
}
/* Nav Dropdown */
nav ul li:hover ul {
display: block;
}
nav ul li ul {
background: var(--color-bg);
border: 1px solid var(--color-bg-secondary);
border-radius: var(--border-radius);
box-shadow: var(--box-shadow) var(--color-shadow);
display: none;
height: auto;
left: -2px;
padding: .5rem 1rem;
position: absolute;
top: 1.7rem;
white-space: nowrap;
width: auto;
z-index: 1;
}
nav ul li ul::before {
/* fill gap above to make mousing over them easier */
content: "";
position: absolute;
left: 0;
right: 0;
top: -0.5rem;
height: 0.5rem;
}
nav ul li ul li,
nav ul li ul li a {
display: block;
}
/* Typography */
code,
samp {
background-color: var(--color-accent);
border-radius: var(--border-radius);
color: var(--color-text);
display: inline-block;
margin: 0 0.1rem;
padding: 0 0.5rem;
}
details {
margin: 1.3rem 0;
}
details summary {
font-weight: bold;
cursor: pointer;
}
h1,
h2,
h3,
h4,
h5,
h6 {
line-height: var(--line-height);
text-wrap: balance;
}
mark {
padding: 0.1rem;
}
ol li,
ul li {
padding: 0.2rem 0;
}
p {
margin: 0.75rem 0;
padding: 0;
width: 100%;
}
pre {
margin: 1rem 0;
max-width: var(--width-card-wide);
padding: 1rem 0;
}
pre code,
pre samp {
display: block;
max-width: var(--width-card-wide);
padding: 0.5rem 2rem;
white-space: pre-wrap;
}
small {
color: var(--color-text-secondary);
}
sup {
background-color: var(--color-secondary);
border-radius: var(--border-radius);
color: var(--color-bg);
font-size: xx-small;
font-weight: bold;
margin: 0.2rem;
padding: 0.2rem 0.3rem;
position: relative;
top: -2px;
}
/* Links */
a {
color: var(--color-link);
display: inline-block;
font-weight: bold;
text-decoration: underline;
}
a:hover {
filter: brightness(var(--hover-brightness));
}
a:active {
filter: brightness(var(--active-brightness));
}
a b,
a em,
a i,
a strong,
button,
input[type="submit"] {
border-radius: var(--border-radius);
display: inline-block;
font-size: medium;
font-weight: bold;
line-height: var(--line-height);
margin: 0.5rem 0;
padding: 1rem 2rem;
}
button,
input[type="submit"] {
font-family: var(--font-family);
}
button:hover,
input[type="submit"]:hover {
cursor: pointer;
filter: brightness(var(--hover-brightness));
}
button:active,
input[type="submit"]:active {
filter: brightness(var(--active-brightness));
}
a b,
a strong,
button,
input[type="submit"] {
background-color: var(--color-link);
border: 2px solid var(--color-link);
color: var(--color-bg);
}
a em,
a i {
border: 2px solid var(--color-link);
border-radius: var(--border-radius);
color: var(--color-link);
display: inline-block;
padding: 1rem 2rem;
}
article aside a {
color: var(--color-secondary);
}
/* Images */
figure {
margin: 0;
padding: 0;
}
figure img {
max-width: 100%;
}
figure figcaption {
color: var(--color-text-secondary);
}
/* Forms */
button:disabled,
input:disabled {
background: var(--color-bg-secondary);
border-color: var(--color-bg-secondary);
color: var(--color-text-secondary);
cursor: not-allowed;
}
button[disabled]:hover,
input[type="submit"][disabled]:hover {
filter: none;
}
form {
border: 1px solid var(--color-bg-secondary);
border-radius: var(--border-radius);
box-shadow: var(--box-shadow) var(--color-shadow);
display: block;
max-width: var(--width-card-wide);
min-width: var(--width-card);
padding: 1.5rem;
text-align: var(--justify-normal);
}
form header {
margin: 1.5rem 0;
padding: 1.5rem 0;
}
input,
label,
select,
textarea {
display: block;
font-size: inherit;
max-width: var(--width-card-wide);
}
input[type="checkbox"],
input[type="radio"] {
display: inline-block;
}
input[type="checkbox"]+label,
input[type="radio"]+label {
display: inline-block;
font-weight: normal;
position: relative;
top: 1px;
}
input[type="range"] {
padding: 0.4rem 0;
}
input,
select,
textarea {
border: 1px solid var(--color-bg-secondary);
border-radius: var(--border-radius);
margin-bottom: 1rem;
padding: 0.4rem 0.8rem;
}
input[type="text"],
input[type="password"] textarea {
width: calc(100% - 1.6rem);
}
input[readonly],
textarea[readonly] {
background-color: var(--color-bg-secondary);
}
label {
font-weight: bold;
margin-bottom: 0.2rem;
}
/* Popups */
dialog {
border: 1px solid var(--color-bg-secondary);
border-radius: var(--border-radius);
box-shadow: var(--box-shadow) var(--color-shadow);
position: fixed;
top: 50%;
left: 50%;
transform: translate(-50%, -50%);
width: 50%;
z-index: 999;
}
/* Tables */
table {
border: 1px solid var(--color-bg-secondary);
border-radius: var(--border-radius);
border-spacing: 0;
display: inline-block;
max-width: 100%;
overflow-x: auto;
padding: 0;
white-space: nowrap;
}
table td,
table th,
table tr {
padding: 0.4rem 0.8rem;
text-align: var(--justify-important);
}
table thead {
background-color: var(--color-table);
border-collapse: collapse;
border-radius: var(--border-radius);
color: var(--color-bg);
margin: 0;
padding: 0;
}
table thead tr:first-child th:first-child {
border-top-left-radius: var(--border-radius);
}
table thead tr:first-child th:last-child {
border-top-right-radius: var(--border-radius);
}
table thead th:first-child,
table tr td:first-child {
text-align: var(--justify-normal);
}
table tr:nth-child(even) {
background-color: var(--color-accent);
}
/* Quotes */
blockquote {
display: block;
font-size: x-large;
line-height: var(--line-height);
margin: 1rem auto;
max-width: var(--width-card-medium);
padding: 1.5rem 1rem;
text-align: var(--justify-important);
}
blockquote footer {
color: var(--color-text-secondary);
display: block;
font-size: small;
line-height: var(--line-height);
padding: 1.5rem 0;
}
/* Scrollbars */
* {
scrollbar-width: thin;
scrollbar-color: var(--color-scrollbar) transparent;
}
*::-webkit-scrollbar {
width: 5px;
height: 5px;
}
*::-webkit-scrollbar-track {
background: transparent;
}
*::-webkit-scrollbar-thumb {
background-color: var(--color-scrollbar);
border-radius: 10px;
}

View File

@@ -1,386 +0,0 @@
use std::collections::HashMap;
use cacher::FilesystemCacher;
use futures::{stream::FuturesUnordered, StreamExt};
use letterbox_shared::compute_color;
use maplit::hashmap;
use scraper::Selector;
use sqlx::postgres::PgPool;
use tracing::{error, info, instrument};
use url::Url;
use crate::{
clean_title, compute_offset_limit,
error::ServerError,
graphql::{Corpus, NewsPost, Tag, Thread, ThreadSummary},
thread_summary_from_row, AddOutlink, FrameImages, Query, SanitizeHtml, SlurpContents,
StripHtml, ThreadSummaryRecord, Transformer, NEWSREADER_TAG_PREFIX, NEWSREADER_THREAD_PREFIX,
};
pub fn is_newsreader_query(query: &Query) -> bool {
query.is_newsreader || query.corpus == Some(Corpus::Newsreader)
}
pub fn is_newsreader_thread(query: &str) -> bool {
query.starts_with(NEWSREADER_THREAD_PREFIX)
}
pub fn extract_thread_id(query: &str) -> &str {
if query.starts_with(NEWSREADER_THREAD_PREFIX) {
&query[NEWSREADER_THREAD_PREFIX.len()..]
} else {
query
}
}
pub fn extract_site(tag: &str) -> &str {
&tag[NEWSREADER_TAG_PREFIX.len()..]
}
pub fn make_news_tag(tag: &str) -> String {
format!("tag:{NEWSREADER_TAG_PREFIX}{tag}")
}
fn site_from_tags(tags: &[String]) -> Option<String> {
for t in tags {
if t.starts_with(NEWSREADER_TAG_PREFIX) {
return Some(extract_site(t).to_string());
}
}
None
}
#[instrument(name = "newsreader::count", skip_all, fields(query=%query))]
pub async fn count(pool: &PgPool, query: &Query) -> Result<usize, ServerError> {
if !is_newsreader_query(query) {
return Ok(0);
}
let site = site_from_tags(&query.tags);
if !query.tags.is_empty() && site.is_none() {
// Newsreader can only handle all sites read/unread queries, anything with a non-site tag
// isn't supported
return Ok(0);
}
let search_term = query.remainder.join(" ");
let search_term = search_term.trim();
let search_term = if search_term.is_empty() {
None
} else {
Some(search_term)
};
// TODO: add support for looking for search_term in title and site
let row = sqlx::query_file!("sql/count.sql", site, query.unread_only, search_term)
.fetch_one(pool)
.await?;
Ok(row.count.unwrap_or(0).try_into().unwrap_or(0))
}
#[instrument(name = "newsreader::search", skip_all, fields(query=%query))]
pub async fn search(
pool: &PgPool,
after: Option<i32>,
before: Option<i32>,
first: Option<i32>,
last: Option<i32>,
query: &Query,
) -> Result<Vec<(i32, ThreadSummary)>, async_graphql::Error> {
info!("search({after:?} {before:?} {first:?} {last:?} {query:?}");
if !is_newsreader_query(query) {
return Ok(Vec::new());
}
let site = site_from_tags(&query.tags);
if !query.tags.is_empty() && site.is_none() {
// Newsreader can only handle all sites read/unread queries, anything with a non-site tag
// isn't supported
return Ok(Vec::new());
}
let (offset, mut limit) = compute_offset_limit(after, before, first, last);
if before.is_none() {
// When searching forward, the +1 is to see if there are more pages of data available.
// Searching backwards implies there's more pages forward, because the value represented by
// `before` is on the next page.
limit = limit + 1;
}
info!(
"search offset {offset} limit {limit} site {site:?} unread_only {}",
query.unread_only
);
let search_term = query.remainder.join(" ");
let search_term = search_term.trim();
let search_term = if search_term.is_empty() {
None
} else {
Some(search_term)
};
// TODO: add support for looking for search_term in title and site
let rows = sqlx::query_file!(
"sql/threads.sql",
site,
query.unread_only,
offset as i64,
limit as i64,
search_term
)
.fetch_all(pool)
.await?;
let mut res = Vec::new();
for (i, r) in rows.into_iter().enumerate() {
res.push((
i as i32 + offset,
thread_summary_from_row(ThreadSummaryRecord {
site: r.site,
date: r.date,
is_read: r.is_read,
title: r.title,
uid: r.uid,
name: r.name,
corpus: Corpus::Newsreader,
})
.await,
));
}
Ok(res)
}
#[instrument(name = "newsreader::tags", skip_all, fields(needs_unread=%_needs_unread))]
pub async fn tags(pool: &PgPool, _needs_unread: bool) -> Result<Vec<Tag>, ServerError> {
// TODO: optimize query by using needs_unread
let tags = sqlx::query_file!("sql/tags.sql").fetch_all(pool).await?;
let tags = tags
.into_iter()
.map(|tag| {
let unread = tag.unread.unwrap_or(0).try_into().unwrap_or(0);
let name = format!(
"{NEWSREADER_TAG_PREFIX}{}",
tag.site.expect("tag must have site")
);
let hex = compute_color(&name);
Tag {
name,
fg_color: "white".to_string(),
bg_color: hex,
unread,
}
})
.collect();
Ok(tags)
}
#[instrument(name = "newsreader::thread", skip_all, fields(thread_id=%thread_id))]
pub async fn thread(
cacher: &FilesystemCacher,
pool: &PgPool,
thread_id: String,
) -> Result<Thread, ServerError> {
let id = thread_id
.strip_prefix(NEWSREADER_THREAD_PREFIX)
.expect("news thread doesn't start with '{NEWSREADER_THREAD_PREFIX}'")
.to_string();
let r = sqlx::query_file!("sql/thread.sql", id)
.fetch_one(pool)
.await?;
let slug = r.site.unwrap_or("no-slug".to_string());
let site = r.name.unwrap_or("NO SITE".to_string());
// TODO: remove the various places that have this as an Option
let link = Some(Url::parse(&r.link)?);
let mut body = r.clean_summary.unwrap_or("NO SUMMARY".to_string());
let body_transformers: Vec<Box<dyn Transformer>> = vec![
Box::new(SlurpContents {
cacher,
inline_css: true,
site_selectors: slurp_contents_selectors(),
}),
Box::new(FrameImages),
Box::new(AddOutlink),
// TODO: causes doubling of images in cloudflare blogs
//Box::new(EscapeHtml),
Box::new(SanitizeHtml {
cid_prefix: "",
base_url: &link,
}),
];
for t in body_transformers.iter() {
if t.should_run(&link, &body) {
body = t.transform(&link, &body).await?;
}
}
let title = clean_title(&r.title.unwrap_or("NO TITLE".to_string())).await?;
let is_read = r.is_read.unwrap_or(false);
let timestamp = r
.date
.expect("post missing date")
.assume_utc()
.unix_timestamp();
Ok(Thread::News(NewsPost {
thread_id,
is_read,
slug,
site,
title,
body,
url: link
.as_ref()
.map(|url| url.to_string())
.unwrap_or("NO URL".to_string()),
timestamp,
}))
}
#[instrument(name = "newsreader::set_read_status", skip_all, fields(query=%query,unread=%unread))]
pub async fn set_read_status<'ctx>(
pool: &PgPool,
query: &Query,
unread: bool,
) -> Result<bool, ServerError> {
// TODO: make single query when query.uids.len() > 1
let uids: Vec<_> = query
.uids
.iter()
.filter(|uid| is_newsreader_thread(uid))
.map(
|uid| extract_thread_id(uid), // TODO strip prefix
)
.collect();
for uid in uids {
sqlx::query_file!("sql/set_unread.sql", !unread, uid)
.execute(pool)
.await?;
}
Ok(true)
}
#[instrument(name = "newsreader::refresh", skip_all)]
pub async fn refresh<'ctx>(pool: &PgPool, cacher: &FilesystemCacher) -> Result<bool, ServerError> {
async fn update_search_summary(
pool: &PgPool,
cacher: &FilesystemCacher,
link: String,
body: String,
id: i32,
) -> Result<(), ServerError> {
let slurp_contents = SlurpContents {
cacher,
inline_css: true,
site_selectors: slurp_contents_selectors(),
};
let strip_html = StripHtml;
info!("adding {link} to search index");
let mut body = body;
if let Ok(link) = Url::parse(&link) {
let link = Some(link);
if slurp_contents.should_run(&link, &body) {
body = slurp_contents.transform(&link, &body).await?;
}
} else {
error!("failed to parse link: {}", link);
}
body = strip_html.transform(&None, &body).await?;
sqlx::query!(
"UPDATE post SET search_summary = $1 WHERE id = $2",
body,
id
)
.execute(pool)
.await?;
Ok(())
}
let mut unordered: FuturesUnordered<_> = sqlx::query_file!("sql/need-search-summary.sql",)
.fetch_all(pool)
.await?
.into_iter()
.filter_map(|r| {
let Some(body) = r.clean_summary else {
error!("clean_summary missing for {}", r.link);
return None;
};
let id = r.id;
Some(update_search_summary(pool, cacher, r.link, body, id))
})
.collect();
while let Some(res) = unordered.next().await {
//let res = res;
match res {
Ok(()) => {}
Err(err) => {
info!("failed refresh {err:?}");
// TODO:
//fd.error = Some(err);
}
};
}
Ok(true)
}
fn slurp_contents_selectors() -> HashMap<String, Vec<Selector>> {
hashmap![
"atmeta.com".to_string() => vec![
Selector::parse("div.entry-content").unwrap(),
],
"blog.prusa3d.com".to_string() => vec![
Selector::parse("article.content .post-block").unwrap(),
],
"blog.cloudflare.com".to_string() => vec![
Selector::parse(".author-lists .author-name-tooltip").unwrap(),
Selector::parse(".post-full-content").unwrap()
],
"blog.zsa.io".to_string() => vec![
Selector::parse("section.blog-article").unwrap(),
],
"engineering.fb.com".to_string() => vec![
Selector::parse("article").unwrap(),
],
"grafana.com".to_string() => vec![
Selector::parse(".blog-content").unwrap(),
],
"hackaday.com".to_string() => vec![
Selector::parse("div.entry-featured-image").unwrap(),
Selector::parse("div.entry-content").unwrap()
],
"ingowald.blog".to_string() => vec![
Selector::parse("article").unwrap(),
],
"jvns.ca".to_string() => vec![
Selector::parse("article").unwrap(),
],
"mitchellh.com".to_string() => vec![Selector::parse("div.w-full").unwrap()],
"natwelch.com".to_string() => vec![
Selector::parse("article div.prose").unwrap(),
],
"seiya.me".to_string() => vec![
Selector::parse("header + div").unwrap(),
],
"rustacean-station.org".to_string() => vec![
Selector::parse("article").unwrap(),
],
"slashdot.org".to_string() => vec![
Selector::parse("span.story-byline").unwrap(),
Selector::parse("div.p").unwrap(),
],
"theonion.com".to_string() => vec![
// Single image joke w/ title
Selector::parse("article > section > div > figure").unwrap(),
// Single cartoon
Selector::parse("article > div > div > figure").unwrap(),
// Image at top of article
Selector::parse("article > header > div > div > figure").unwrap(),
// Article body
Selector::parse("article .entry-content > *").unwrap(),
],
"trofi.github.io".to_string() => vec![
Selector::parse("#content").unwrap(),
],
"www.redox-os.org".to_string() => vec![
Selector::parse("div.content").unwrap(),
],
"www.smbc-comics.com".to_string() => vec![
Selector::parse("img#cc-comic").unwrap(),
Selector::parse("div#aftercomic img").unwrap(),
],
]
}

View File

@@ -1,813 +1,15 @@
use std::{
collections::{HashMap, HashSet},
fs::File,
io::{Cursor, Read},
};
use shared::Message;
use letterbox_notmuch::Notmuch;
use letterbox_shared::{compute_color, Rule};
use mailparse::{parse_mail, MailHeader, MailHeaderMap};
use memmap::MmapOptions;
use sqlx::{types::Json, PgPool};
use tracing::{error, info, info_span, instrument, warn};
use zip::ZipArchive;
use crate::{
compute_offset_limit,
email_extract::*,
error::ServerError,
graphql::{
Attachment, Body, Corpus, EmailThread, Header, Html, Message, PlainText, Tag, Thread,
ThreadSummary, UnhandledContentType,
},
linkify_html, InlineStyle, Query, SanitizeHtml, Transformer,
};
const APPLICATION_GZIP: &'static str = "application/gzip";
const APPLICATION_ZIP: &'static str = "application/zip";
const MULTIPART_REPORT: &'static str = "multipart/report";
const MAX_RAW_MESSAGE_SIZE: usize = 100_000;
fn is_notmuch_query(query: &Query) -> bool {
query.is_notmuch || query.corpus == Some(Corpus::Notmuch)
}
pub fn is_notmuch_thread_or_id(id: &str) -> bool {
id.starts_with("id:") || id.starts_with("thread:")
}
use crate::error;
// TODO(wathiede): decide good error type
pub fn threadset_to_messages(
thread_set: letterbox_notmuch::ThreadSet,
) -> Result<Vec<Message>, ServerError> {
thread_set: notmuch::ThreadSet,
) -> Result<Vec<Message>, error::ServerError> {
for t in thread_set.0 {
for _tn in t.0 {}
for tn in t.0 {
todo!()
}
}
Ok(Vec::new())
}
#[instrument(name="nm::count", skip_all, fields(query=%query))]
pub async fn count(nm: &Notmuch, query: &Query) -> Result<usize, ServerError> {
if !is_notmuch_query(query) {
return Ok(0);
}
let query = query.to_notmuch();
Ok(nm.count(&query)?)
}
#[instrument(name="nm::search", skip_all, fields(query=%query))]
pub async fn search(
nm: &Notmuch,
after: Option<i32>,
before: Option<i32>,
first: Option<i32>,
last: Option<i32>,
query: &Query,
) -> Result<Vec<(i32, ThreadSummary)>, async_graphql::Error> {
if !is_notmuch_query(query) {
return Ok(Vec::new());
}
let query = query.to_notmuch();
let (offset, mut limit) = compute_offset_limit(after, before, first, last);
if before.is_none() {
// When searching forward, the +1 is to see if there are more pages of data available.
// Searching backwards implies there's more pages forward, because the value represented by
// `before` is on the next page.
limit = limit + 1;
}
Ok(nm
.search(&query, offset as usize, limit as usize)?
.0
.into_iter()
.enumerate()
.map(|(i, ts)| {
(
offset + i as i32,
ThreadSummary {
thread: format!("thread:{}", ts.thread),
timestamp: ts.timestamp,
date_relative: ts.date_relative,
matched: ts.matched,
total: ts.total,
authors: ts.authors,
subject: ts.subject,
tags: ts.tags,
corpus: Corpus::Notmuch,
},
)
})
.collect())
}
#[instrument(name="nm::tags", skip_all, fields(needs_unread=needs_unread))]
pub fn tags(nm: &Notmuch, needs_unread: bool) -> Result<Vec<Tag>, ServerError> {
let unread_msg_cnt: HashMap<String, usize> = if needs_unread {
// 10000 is an arbitrary number, if there's more than 10k unread messages, we'll
// get an inaccurate count.
nm.search("is:unread", 0, 10000)?
.0
.iter()
.fold(HashMap::new(), |mut m, ts| {
ts.tags.iter().for_each(|t| {
m.entry(t.clone()).and_modify(|c| *c += 1).or_insert(1);
});
m
})
} else {
HashMap::new()
};
let tags: Vec<_> = nm
.tags()?
.into_iter()
.map(|tag| {
let hex = compute_color(&tag);
let unread = if needs_unread {
*unread_msg_cnt.get(&tag).unwrap_or(&0)
} else {
0
};
Tag {
name: tag,
fg_color: "white".to_string(),
bg_color: hex,
unread,
}
})
.chain(
nm.unread_recipients()?
.into_iter()
.filter_map(|(name, unread)| {
let Some(idx) = name.find('@') else {
return None;
};
let name = format!("{}/{}", &name[idx..], &name[..idx]);
let bg_color = compute_color(&name);
Some(Tag {
name,
fg_color: "white".to_string(),
bg_color,
unread,
})
}),
)
.collect();
Ok(tags)
}
#[instrument(name="nm::thread", skip_all, fields(thread_id=thread_id))]
pub async fn thread(
nm: &Notmuch,
pool: &PgPool,
thread_id: String,
debug_content_tree: bool,
) -> Result<Thread, ServerError> {
// TODO(wathiede): normalize all email addresses through an address book with preferred
// display names (that default to the most commonly seen name).
let mut messages = Vec::new();
for (path, id) in std::iter::zip(nm.files(&thread_id)?, nm.message_ids(&thread_id)?) {
let mut html_report_summary: Option<String> = None;
let tags = nm.tags_for_query(&format!("id:{}", id))?;
let file = File::open(&path)?;
let mmap = unsafe { MmapOptions::new().map(&file)? };
let m = parse_mail(&mmap)?;
let from = email_addresses(&path, &m, "from")?;
let mut from = match from.len() {
0 => None,
1 => from.into_iter().next(),
_ => {
warn!(
"Got {} from addresses in message, truncating: {:?}",
from.len(),
from
);
from.into_iter().next()
}
};
match from.as_mut() {
Some(from) => {
if let Some(addr) = from.addr.as_mut() {
let photo_url = photo_url_for_email_address(&pool, &addr).await?;
from.photo_url = photo_url;
}
}
_ => (),
}
let to = email_addresses(&path, &m, "to")?;
let cc = email_addresses(&path, &m, "cc")?;
let delivered_to = email_addresses(&path, &m, "delivered-to")?.pop();
let x_original_to = email_addresses(&path, &m, "x-original-to")?.pop();
let subject = m.headers.get_first_value("subject");
let timestamp = m
.headers
.get_first_value("date")
.and_then(|d| mailparse::dateparse(&d).ok());
let cid_prefix = letterbox_shared::urls::cid_prefix(None, &id);
let base_url = None;
let mut part_addr = Vec::new();
part_addr.push(id.to_string());
let body = match extract_body(&m, &mut part_addr)? {
Body::PlainText(PlainText { text, content_tree }) => {
let text = if text.len() > MAX_RAW_MESSAGE_SIZE {
format!(
"{}...\n\nMESSAGE WAS TRUNCATED @ {} bytes",
&text[..MAX_RAW_MESSAGE_SIZE],
MAX_RAW_MESSAGE_SIZE
)
} else {
text
};
Body::Html(Html {
html: {
let body_tranformers: Vec<Box<dyn Transformer>> = vec![
Box::new(InlineStyle),
Box::new(SanitizeHtml {
cid_prefix: &cid_prefix,
base_url: &base_url,
}),
];
let mut html = linkify_html(&text.trim_matches('\n'));
for t in body_tranformers.iter() {
if t.should_run(&None, &html) {
html = t.transform(&None, &html).await?;
}
}
format!(
r#"<p class="view-part-text-plain font-mono whitespace-pre-line">{}</p>"#,
// Trim newlines to prevent excessive white space at the beginning/end of
// presenation. Leave tabs and spaces incase plain text attempts to center a
// header on the first line.
html
)
},
content_tree: if debug_content_tree {
render_content_type_tree(&m)
} else {
content_tree
},
})
}
Body::Html(Html {
mut html,
content_tree,
}) => Body::Html(Html {
html: {
let body_tranformers: Vec<Box<dyn Transformer>> = vec![
// TODO: this breaks things like emails from calendar
//Box::new(InlineStyle),
Box::new(SanitizeHtml {
cid_prefix: &cid_prefix,
base_url: &base_url,
}),
];
for t in body_tranformers.iter() {
if t.should_run(&None, &html) {
html = t.transform(&None, &html).await?;
}
}
html
},
content_tree: if debug_content_tree {
render_content_type_tree(&m)
} else {
content_tree
},
}),
Body::UnhandledContentType(UnhandledContentType { content_tree, .. }) => {
let body_start = mmap
.windows(2)
.take(20_000)
.position(|w| w == b"\n\n")
.unwrap_or(0);
let body = mmap[body_start + 2..].to_vec();
Body::UnhandledContentType(UnhandledContentType {
text: String::from_utf8(body)?,
content_tree: if debug_content_tree {
render_content_type_tree(&m)
} else {
content_tree
},
})
}
};
let headers = m
.headers
.iter()
.map(|h| Header {
key: h.get_key(),
value: h.get_value(),
})
.collect();
// TODO(wathiede): parse message and fill out attachments
let attachments = extract_attachments(&m, &id)?;
let mut final_body = body;
let mut raw_report_content: Option<String> = None;
// Append TLS report if available
if m.ctype.mimetype.as_str() == MULTIPART_REPORT {
if let Ok(Body::Html(_html_body)) = extract_report(&m, &mut part_addr) {
// Extract raw JSON for pretty printing
if let Some(sp) = m
.subparts
.iter()
.find(|sp| sp.ctype.mimetype.as_str() == "application/tlsrpt+gzip")
{
if let Ok(gz_bytes) = sp.get_body_raw() {
let mut decoder = flate2::read::GzDecoder::new(&gz_bytes[..]);
let mut buffer = Vec::new();
if decoder.read_to_end(&mut buffer).is_ok() {
if let Ok(json_str) = String::from_utf8(buffer) {
raw_report_content = Some(json_str);
}
}
}
}
}
}
// Append DMARC report if available
if m.ctype.mimetype.as_str() == APPLICATION_ZIP {
if let Ok(Body::Html(html_body)) = extract_zip(&m) {
html_report_summary = Some(html_body.html);
// Extract raw XML for pretty printing
if let Ok(zip_bytes) = m.get_body_raw() {
if let Ok(mut archive) = ZipArchive::new(Cursor::new(&zip_bytes)) {
for i in 0..archive.len() {
if let Ok(mut file) = archive.by_index(i) {
let name = file.name().to_lowercase();
if is_dmarc_report_filename(&name) {
let mut xml = String::new();
use std::io::Read;
if file.read_to_string(&mut xml).is_ok() {
raw_report_content = Some(xml);
}
}
}
}
}
}
}
}
if m.ctype.mimetype.as_str() == APPLICATION_GZIP {
// Call extract_gzip to get the HTML summary and also to determine if it's a DMARC report
if let Ok((Body::Html(html_body), _)) = extract_gzip(&m) {
html_report_summary = Some(html_body.html);
// If extract_gzip successfully parsed a DMARC report, then extract the raw content
if let Ok(gz_bytes) = m.get_body_raw() {
let mut decoder = flate2::read::GzDecoder::new(&gz_bytes[..]);
let mut xml = String::new();
use std::io::Read;
if decoder.read_to_string(&mut xml).is_ok() {
raw_report_content = Some(xml);
}
}
}
}
let mut current_html = final_body.to_html().unwrap_or_default();
if let Some(html_summary) = html_report_summary {
current_html.push_str(&html_summary);
}
error!(
"mimetype {} raw_report_content.is_some() {}",
m.ctype.mimetype.as_str(),
raw_report_content.is_some()
);
if let Some(raw_content) = raw_report_content {
let pretty_printed_content = if m.ctype.mimetype.as_str() == MULTIPART_REPORT {
// Pretty print JSON
if let Ok(parsed_json) = serde_json::from_str::<serde_json::Value>(&raw_content) {
serde_json::to_string_pretty(&parsed_json).unwrap_or(raw_content)
} else {
raw_content
}
} else {
// DMARC reports are XML
// Pretty print XML
match pretty_print_xml_with_trimming(&raw_content) {
Ok(pretty_xml) => pretty_xml,
Err(e) => {
error!("Failed to pretty print XML: {:?}", e);
raw_content
}
}
};
current_html.push_str(&format!(
"\n<pre>{}</pre>",
html_escape::encode_text(&pretty_printed_content)
));
}
final_body = Body::Html(Html {
html: current_html,
content_tree: final_body.to_html_content_tree().unwrap_or_default(),
});
messages.push(Message {
id: format!("id:{}", id),
from,
to,
cc,
subject,
tags,
timestamp,
headers,
body: final_body,
path,
attachments,
delivered_to,
x_original_to,
});
}
messages.reverse();
// Find the first subject that's set. After reversing the vec, this should be the oldest
// message.
let subject: String = messages
.iter()
.skip_while(|m| m.subject.is_none())
.next()
.and_then(|m| m.subject.clone())
.unwrap_or("(NO SUBJECT)".to_string());
Ok(Thread::Email(EmailThread {
thread_id,
subject,
messages,
}))
}
pub fn cid_attachment_bytes(nm: &Notmuch, id: &str, cid: &str) -> Result<Attachment, ServerError> {
let files = nm.files(id)?;
let Some(path) = files.first() else {
warn!("failed to find files for message {}", id);
return Err(ServerError::PartNotFound);
};
let file = File::open(&path)?;
let mmap = unsafe { MmapOptions::new().map(&file)? };
let m = parse_mail(&mmap)?;
if let Some(attachment) = walk_attachments(&m, |sp, _cur_idx| {
info!("{} {:?}", cid, get_content_id(&sp.headers));
if let Some(h_cid) = get_content_id(&sp.headers) {
let h_cid = &h_cid[1..h_cid.len() - 1];
if h_cid == cid {
let attachment = extract_attachment(&sp, id, &[]).unwrap_or(Attachment {
..Attachment::default()
});
return Some(attachment);
}
}
None
}) {
return Ok(attachment);
}
Err(ServerError::PartNotFound)
}
pub fn attachment_bytes(nm: &Notmuch, id: &str, idx: &[usize]) -> Result<Attachment, ServerError> {
let files = nm.files(id)?;
let Some(path) = files.first() else {
warn!("failed to find files for message {}", id);
return Err(ServerError::PartNotFound);
};
let file = File::open(&path)?;
let mmap = unsafe { MmapOptions::new().map(&file)? };
let m = parse_mail(&mmap)?;
if let Some(attachment) = walk_attachments(&m, |sp, cur_idx| {
if cur_idx == idx {
let attachment = extract_attachment(&sp, id, idx).unwrap_or(Attachment {
..Attachment::default()
});
return Some(attachment);
}
None
}) {
return Ok(attachment);
}
Err(ServerError::PartNotFound)
}
#[instrument(name="nm::set_read_status", skip_all, fields(query=%query, unread=unread))]
pub async fn set_read_status<'ctx>(
nm: &Notmuch,
query: &Query,
unread: bool,
) -> Result<bool, ServerError> {
let uids: Vec<_> = query
.uids
.iter()
.filter(|uid| is_notmuch_thread_or_id(uid))
.collect();
info!("set_read_status({} {:?})", unread, uids);
for uid in uids {
if unread {
nm.tag_add("unread", uid)?;
} else {
nm.tag_remove("unread", uid)?;
}
}
Ok(true)
}
async fn photo_url_for_email_address(
pool: &PgPool,
addr: &str,
) -> Result<Option<String>, ServerError> {
let row =
sqlx::query_as::<_, (String,)>(include_str!("../sql/photo_url_for_email_address.sql"))
.bind(addr)
.fetch_optional(pool)
.await?;
Ok(row.map(|r| r.0))
}
/*
* grab email_rules table from sql
* For each message with `unprocessed` label
* parse the message
* pass headers for each message through a matcher using email rules
* for each match, add label to message
* if any matches were found, remove unprocessed
* TODO: how to handle inbox label
*/
#[instrument(name="nm::label_unprocessed", skip_all, fields(dryrun=dryrun, limit=?limit, query=%query))]
pub async fn label_unprocessed(
nm: &Notmuch,
pool: &PgPool,
dryrun: bool,
limit: Option<usize>,
query: &str,
) -> Result<Box<[String]>, ServerError> {
use futures::StreamExt;
let ids = nm.message_ids(query)?;
info!(
"Processing {:?} of {} messages with '{}'",
limit,
ids.len(),
query
);
let rules: Vec<_> =
sqlx::query_as::<_, (Json<Rule>,)>(include_str!("../sql/label_unprocessed.sql"))
.fetch(pool)
.map(|r| r.unwrap().0 .0)
.collect()
.await;
/*
use letterbox_shared::{Match, MatchType};
let rules = vec![Rule {
stop_on_match: false,
matches: vec![Match {
match_type: MatchType::From,
needle: "eftours".to_string(),
}],
tag: "EFTours".to_string(),
}];
*/
info!("Loaded {} rules", rules.len());
let limit = limit.unwrap_or(ids.len());
let limit = limit.min(ids.len());
let ids = &ids[..limit];
let mut add_mutations = HashMap::new();
let mut rm_mutations = HashMap::new();
for id in ids {
let id = format!("id:{}", id);
let files = nm.files(&id)?;
// Only process the first file path is multiple files have the same id
let Some(path) = files.iter().next() else {
error!("No files for message-ID {}", id);
let t = "Letterbox/Bad";
nm.tag_add(t, &id)?;
let t = "unprocessed";
nm.tag_remove(t, &id)?;
continue;
};
let file = File::open(&path)?;
info!("parsing {}", path);
let mmap = unsafe { MmapOptions::new().map(&file)? };
let m = match info_span!("parse_mail", path = path).in_scope(|| parse_mail(&mmap)) {
Ok(m) => m,
Err(err) => {
error!("Failed to parse {}: {}", path, err);
let t = "Letterbox/Bad";
nm.tag_add(t, &id)?;
let t = "unprocessed";
nm.tag_remove(t, &id)?;
continue;
}
};
let (matched_rule, add_tags) = find_tags(&rules, &m.headers);
if matched_rule {
if dryrun {
info!(
"\nAdd tags: {:?}\nTo: {} From: {} Subject: {}\n",
add_tags,
m.headers.get_first_value("to").expect("no from header"),
m.headers.get_first_value("from").expect("no from header"),
m.headers
.get_first_value("subject")
.expect("no subject header")
);
}
for t in &add_tags {
//nm.tag_add(t, &id)?;
add_mutations
.entry(t.to_string())
.or_insert_with(|| Vec::new())
.push(id.clone());
}
if add_tags.contains("spam") || add_tags.contains("Spam") {
//nm.tag_remove("unread", &id)?;
let t = "unread".to_string();
rm_mutations
.entry(t)
.or_insert_with(|| Vec::new())
.push(id.clone());
}
if !add_tags.contains("inbox") {
//nm.tag_remove("inbox", &id)?;
let t = "inbox".to_string();
rm_mutations
.entry(t)
.or_insert_with(|| Vec::new())
.push(id.clone());
}
//nm.tag_remove("unprocessed", &id)?;
} else {
if add_tags.is_empty() {
let t = "Grey".to_string();
add_mutations
.entry(t)
.or_insert_with(|| Vec::new())
.push(id.clone());
}
//nm.tag_remove("inbox", &id)?;
let t = "inbox".to_string();
rm_mutations
.entry(t)
.or_insert_with(|| Vec::new())
.push(id.clone());
}
let t = "unprocessed".to_string();
rm_mutations
.entry(t)
.or_insert_with(|| Vec::new())
.push(id.clone());
}
info!("Adding {} distinct labels", add_mutations.len());
for (tag, ids) in add_mutations.iter() {
info!(" {}: {}", tag, ids.len());
if !dryrun {
let ids: Vec<_> = ids.iter().map(|s| s.as_str()).collect();
info_span!("tags_add", tag = tag, count = ids.len())
.in_scope(|| nm.tags_add(tag, &ids))?;
}
}
info!("Removing {} distinct labels", rm_mutations.len());
for (tag, ids) in rm_mutations.iter() {
info!(" {}: {}", tag, ids.len());
if !dryrun {
let ids: Vec<_> = ids.iter().map(|s| s.as_str()).collect();
info_span!("tags_remove", tag = tag, count = ids.len())
.in_scope(|| nm.tags_remove(tag, &ids))?;
}
}
Ok(ids.into())
}
fn find_tags<'a, 'b>(rules: &'a [Rule], headers: &'b [MailHeader]) -> (bool, HashSet<&'a str>) {
let mut matched_rule = false;
let mut add_tags = HashSet::new();
for rule in rules {
for hdr in headers {
if rule.is_match(&hdr.get_key(), &hdr.get_value()) {
//info!("Matched {:?}", rule);
matched_rule = true;
add_tags.insert(rule.tag.as_str());
if rule.stop_on_match {
return (true, add_tags);
}
}
}
}
(matched_rule, add_tags)
}
#[cfg(test)]
mod tests {
use super::*;
const REPORT_V1: &str = r#"
{
"organization-name": "Google Inc.",
"date-range": {
"start-datetime": "2025-08-09T00:00:00Z",
"end-datetime": "2025-08-09T23:59:59Z"
},
"contact-info": "smtp-tls-reporting@google.com",
"report-id": "2025-08-09T00:00:00Z_xinu.tv",
"policies": [
{
"policy": {
"policy-type": "sts",
"policy-string": [
"version: STSv1",
"mode": "testing",
"mx": "mail.xinu.tv",
"max_age": "86400"
],
"policy-domain": "xinu.tv"
},
"summary": {
"total-successful-session-count": 20,
"total-failure-session-count": 0
}
}
]
}
"#;
const REPORT_V2: &str = r#"
{
"organization-name": "Google Inc.",
"date-range": {
"start-datetime": "2025-08-09T00:00:00Z",
"end-datetime": "2025-08-09T23:59:59Z"
},
"contact-info": "smtp-tls-reporting@google.com",
"report-id": "2025-08-09T00:00:00Z_xinu.tv",
"policies": [
{
"policy": {
"policy-type": "sts",
"policy-string": [
"version: STSv1",
"mode": "testing",
"mx": "mail.xinu.tv",
"max_age": "86400"
],
"policy-domain": "xinu.tv",
"mx-host": [
"mail.xinu.tv"
]
},
"summary": {
"total-successful-session-count": 3,
"total-failure-session-count": 0
}
}
]
}
"#;
const REPORT_V3: &str = r#"
{
"organization-name": "Google Inc.",
"date-range": {
"start-datetime": "2025-08-09T00:00:00Z",
"end-datetime": "2025-08-09T23:59:59Z"
},
"contact-info": "smtp-tls-reporting@google.com",
"report-id": "2025-08-09T00:00:00Z_xinu.tv",
"policies": [
{
"policy": {
"policy-type": "sts",
"policy-string": [
"version: STSv1",
"mode": "testing",
"mx": "mail.xinu.tv",
"max_age": "86400"
],
"policy-domain": "xinu.tv",
"mx-host": [
{
"hostname": "mail.xinu.tv",
"failure-count": 0,
"result-type": "success"
}
]
},
"summary": {
"total-successful-session-count": 3,
"total-failure-session-count": 0
}
}
]
}
"#;
#[test]
fn test_parse_tls_report_v1() {
let report: TlsRpt = serde_json::from_str(REPORT_V1).unwrap();
}
}

View File

@@ -1,353 +0,0 @@
use std::collections::HashSet;
use log::{debug, error, info, warn};
use sqlx::{postgres::PgPool, types::time::PrimitiveDateTime};
use tantivy::{
collector::{DocSetCollector, TopDocs},
doc, query,
query::{AllQuery, BooleanQuery, Occur, QueryParser, TermQuery},
schema::{Facet, IndexRecordOption, Value},
DocAddress, Index, IndexReader, Searcher, TantivyDocument, TantivyError, Term,
};
use tracing::{info_span, instrument, Instrument};
use crate::{
compute_offset_limit,
error::ServerError,
graphql::{Corpus, ThreadSummary},
newsreader::{extract_thread_id, is_newsreader_thread},
thread_summary_from_row, Query, ThreadSummaryRecord,
};
pub fn is_tantivy_query(query: &Query) -> bool {
query.is_tantivy || query.corpus == Some(Corpus::Tantivy)
}
pub struct TantivyConnection {
db_path: String,
index: Index,
reader: IndexReader,
}
fn get_index(db_path: &str) -> Result<Index, TantivyError> {
Ok(match Index::open_in_dir(db_path) {
Ok(idx) => idx,
Err(err) => {
warn!("Failed to open {db_path}: {err}");
create_news_db(db_path)?;
Index::open_in_dir(db_path)?
}
})
}
impl TantivyConnection {
pub fn new(tantivy_db_path: &str) -> Result<TantivyConnection, TantivyError> {
let index = get_index(tantivy_db_path)?;
let reader = index.reader()?;
Ok(TantivyConnection {
db_path: tantivy_db_path.to_string(),
index,
reader,
})
}
#[instrument(name = "tantivy::refresh", skip_all)]
pub async fn refresh(&self, pool: &PgPool) -> Result<(), ServerError> {
let start_time = std::time::Instant::now();
let p_uids: Vec<_> = sqlx::query_file!("sql/all-uids.sql")
.fetch_all(pool)
.instrument(info_span!("postgres query"))
.await?
.into_iter()
.map(|r| r.uid)
.collect();
info!(
"refresh from postgres got {} uids in {}",
p_uids.len(),
start_time.elapsed().as_secs_f32()
);
let t_span = info_span!("tantivy query");
let _enter = t_span.enter();
let start_time = std::time::Instant::now();
let (searcher, _query) = self.searcher_and_query(&Query::default())?;
let docs = searcher.search(&AllQuery, &DocSetCollector)?;
let uid = self.index.schema().get_field("uid")?;
let t_uids: Vec<_> = docs
.into_iter()
.map(|doc_address| {
searcher
.doc(doc_address)
.map(|doc: TantivyDocument| {
debug!("doc: {doc:#?}");
doc.get_first(uid)
.expect("uid")
.as_str()
.expect("as_str")
.to_string()
})
.expect("searcher.doc")
})
.collect();
drop(_enter);
info!(
"refresh tantivy got {} uids in {}",
t_uids.len(),
start_time.elapsed().as_secs_f32()
);
let t_set: HashSet<_> = t_uids.into_iter().collect();
let need: Vec<_> = p_uids
.into_iter()
.filter(|uid| !t_set.contains(uid.as_str()))
.collect();
if !need.is_empty() {
info!(
"need to reindex {} uids: {:?}...",
need.len(),
&need[..need.len().min(10)]
);
}
let batch_size = 1000;
let uids: Vec<_> = need[..need.len().min(batch_size)]
.into_iter()
.cloned()
.collect();
self.reindex_uids(pool, &uids).await
}
#[instrument(skip(self, pool))]
async fn reindex_uids(&self, pool: &PgPool, uids: &[String]) -> Result<(), ServerError> {
if uids.is_empty() {
return Ok(());
}
// TODO: add SlurpContents and convert HTML to text
let pool: &PgPool = pool;
let mut index_writer = self.index.writer(50_000_000)?;
let schema = self.index.schema();
let site = schema.get_field("site")?;
let title = schema.get_field("title")?;
let summary = schema.get_field("summary")?;
let link = schema.get_field("link")?;
let date = schema.get_field("date")?;
let is_read = schema.get_field("is_read")?;
let uid = schema.get_field("uid")?;
let id = schema.get_field("id")?;
let tag = schema.get_field("tag")?;
info!("reindexing {} posts", uids.len());
let rows = sqlx::query_file_as!(PostgresDoc, "sql/posts-from-uids.sql", uids)
.fetch_all(pool)
.await?;
if uids.len() != rows.len() {
error!(
"Had {} uids and only got {} rows: uids {uids:?}",
uids.len(),
rows.len()
);
}
for r in rows {
let id_term = Term::from_field_text(uid, &r.uid);
index_writer.delete_term(id_term);
let slug = r.site;
let tag_facet = Facet::from(&format!("/News/{slug}"));
index_writer.add_document(doc!(
site => slug.clone(),
title => r.title,
// TODO: clean and extract text from HTML
summary => r.summary,
link => r.link,
date => tantivy::DateTime::from_primitive(r.date),
is_read => r.is_read,
uid => r.uid,
id => r.id as u64,
tag => tag_facet,
))?;
}
info_span!("IndexWriter.commit").in_scope(|| index_writer.commit())?;
info_span!("IndexReader.reload").in_scope(|| self.reader.reload())?;
Ok(())
}
#[instrument(name = "tantivy::reindex_thread", skip_all, fields(query=%query))]
pub async fn reindex_thread(&self, pool: &PgPool, query: &Query) -> Result<(), ServerError> {
let uids: Vec<_> = query
.uids
.iter()
.filter(|uid| is_newsreader_thread(uid))
.map(|uid| extract_thread_id(uid).to_string())
.collect();
Ok(self.reindex_uids(pool, &uids).await?)
}
#[instrument(name = "tantivy::reindex_all", skip_all)]
pub async fn reindex_all(&self, pool: &PgPool) -> Result<(), ServerError> {
let rows = sqlx::query_file!("sql/all-posts.sql")
.fetch_all(pool)
.await?;
let uids: Vec<String> = rows.into_iter().map(|r| r.uid).collect();
self.reindex_uids(pool, &uids).await?;
Ok(())
}
fn searcher_and_query(
&self,
query: &Query,
) -> Result<(Searcher, Box<dyn query::Query>), ServerError> {
// TODO: only create one reader
// From https://tantivy-search.github.io/examples/basic_search.html
// "For a search server you will typically create one reader for the entire lifetime of
// your program, and acquire a new searcher for every single request."
//
// I think there's some challenge in making the reader work if we reindex, so reader my
// need to be stored indirectly, and be recreated on reindex
// I think creating a reader takes 200-300 ms.
let schema = self.index.schema();
let searcher = self.reader.searcher();
let title = schema.get_field("title")?;
let summary = schema.get_field("summary")?;
let query_parser = QueryParser::for_index(&self.index, vec![title, summary]);
// Tantivy uses '*' to match all docs, not empty string
let term = &query.remainder.join(" ");
let term = if term.is_empty() { "*" } else { term };
info!("query_parser('{term}')");
let tantivy_query = query_parser.parse_query(&term)?;
let tag = schema.get_field("tag")?;
let is_read = schema.get_field("is_read")?;
let mut terms = vec![(Occur::Must, tantivy_query)];
for t in &query.tags {
let facet = Facet::from(&format!("/{t}"));
let facet_term = Term::from_facet(tag, &facet);
let facet_term_query = Box::new(TermQuery::new(facet_term, IndexRecordOption::Basic));
terms.push((Occur::Must, facet_term_query));
}
if query.unread_only {
info!("searching for unread only");
let term = Term::from_field_bool(is_read, false);
terms.push((
Occur::Must,
Box::new(TermQuery::new(term, IndexRecordOption::Basic)),
));
}
let search_query = BooleanQuery::new(terms);
Ok((searcher, Box::new(search_query)))
}
#[instrument(name="tantivy::count", skip_all, fields(query=%query))]
pub async fn count(&self, query: &Query) -> Result<usize, ServerError> {
if !is_tantivy_query(query) {
return Ok(0);
}
info!("tantivy::count {query:?}");
use tantivy::collector::Count;
let (searcher, query) = self.searcher_and_query(&query)?;
Ok(searcher.search(&query, &Count)?)
}
#[instrument(name="tantivy::search", skip_all, fields(query=%query))]
pub async fn search(
&self,
pool: &PgPool,
after: Option<i32>,
before: Option<i32>,
first: Option<i32>,
last: Option<i32>,
query: &Query,
) -> Result<Vec<(i32, ThreadSummary)>, async_graphql::Error> {
if !is_tantivy_query(query) {
return Ok(Vec::new());
}
let (offset, mut limit) = compute_offset_limit(after, before, first, last);
if before.is_none() {
// When searching forward, the +1 is to see if there are more pages of data available.
// Searching backwards implies there's more pages forward, because the value represented by
// `before` is on the next page.
limit = limit + 1;
}
let (searcher, search_query) = self.searcher_and_query(&query)?;
info!("Tantivy::search(query '{query:?}', off {offset}, lim {limit}, search_query {search_query:?})");
let top_docs = searcher.search(
&search_query,
&TopDocs::with_limit(limit as usize)
.and_offset(offset as usize)
.order_by_u64_field("date", tantivy::index::Order::Desc),
)?;
info!("search found {} docs", top_docs.len());
let uid = self.index.schema().get_field("uid")?;
let uids = top_docs
.into_iter()
.map(|(_, doc_address): (u64, DocAddress)| {
searcher.doc(doc_address).map(|doc: TantivyDocument| {
debug!("doc: {doc:#?}");
doc.get_first(uid)
.expect("doc missing uid")
.as_str()
.expect("doc str missing")
.to_string()
})
})
.collect::<Result<Vec<String>, TantivyError>>()?;
//let uids = format!("'{}'", uids.join("','"));
info!("uids {uids:?}");
let rows = sqlx::query_file!("sql/threads-from-uid.sql", &uids as &[String])
.fetch_all(pool)
.await?;
let mut res = Vec::new();
info!("found {} hits joining w/ tantivy", rows.len());
for (i, r) in rows.into_iter().enumerate() {
res.push((
i as i32 + offset,
thread_summary_from_row(ThreadSummaryRecord {
site: r.site,
date: r.date,
is_read: r.is_read,
title: r.title,
uid: r.uid,
name: r.name,
corpus: Corpus::Tantivy,
})
.await,
));
}
Ok(res)
}
pub fn drop_and_load_index(&self) -> Result<(), TantivyError> {
create_news_db(&self.db_path)
}
}
fn create_news_db(tantivy_db_path: &str) -> Result<(), TantivyError> {
info!("create_news_db");
// Don't care if directory didn't exist
let _ = std::fs::remove_dir_all(tantivy_db_path);
std::fs::create_dir_all(tantivy_db_path)?;
use tantivy::schema::*;
let mut schema_builder = Schema::builder();
schema_builder.add_text_field("site", STRING | STORED);
schema_builder.add_text_field("title", TEXT | STORED);
schema_builder.add_text_field("summary", TEXT);
schema_builder.add_text_field("link", STRING | STORED);
schema_builder.add_date_field("date", FAST | INDEXED | STORED);
schema_builder.add_bool_field("is_read", FAST | INDEXED | STORED);
schema_builder.add_text_field("uid", STRING | STORED);
schema_builder.add_u64_field("id", FAST);
schema_builder.add_facet_field("tag", FacetOptions::default());
let schema = schema_builder.build();
Index::create_in_dir(tantivy_db_path, schema)?;
Ok(())
}
struct PostgresDoc {
site: String,
title: String,
summary: String,
link: String,
date: PrimitiveDateTime,
is_read: bool,
uid: String,
id: i32,
}

View File

@@ -1,7 +0,0 @@
use askama::Template;
#[derive(Template)]
#[template(path = "dmarc_report.html")]
pub struct DmarcReportTemplate<'a> {
pub feedback: &'a crate::nm::Feedback,
}

View File

@@ -1,35 +0,0 @@
use std::{collections::HashMap, net::SocketAddr};
use axum::extract::ws::{Message, WebSocket};
use letterbox_shared::WebsocketMessage;
use tracing::{info, warn};
#[derive(Default)]
pub struct ConnectionTracker {
peers: HashMap<SocketAddr, WebSocket>,
}
impl ConnectionTracker {
pub async fn add_peer(&mut self, socket: WebSocket, who: SocketAddr) {
warn!("adding {who:?} to connection tracker");
self.peers.insert(who, socket);
self.send_message_all(WebsocketMessage::RefreshMessages)
.await;
}
pub async fn send_message_all(&mut self, msg: WebsocketMessage) {
info!("send_message_all {msg}");
let m = serde_json::to_string(&msg).expect("failed to json encode WebsocketMessage");
let mut bad_peers = Vec::new();
for (who, socket) in &mut self.peers.iter_mut() {
if let Err(e) = socket.send(Message::Text(m.clone().into())).await {
warn!("{:?} is bad, scheduling for removal: {e}", who);
bad_peers.push(who.clone());
}
}
for b in bad_peers {
info!("removing bad peer {b:?}");
self.peers.remove(&b);
}
}
}

View File

@@ -1,59 +0,0 @@
<!DOCTYPE html>
<html>
<head>
<meta charset=utf-8 />
<meta name="viewport" content="user-scalable=no, initial-scale=1.0, minimum-scale=1.0, maximum-scale=1.0, minimal-ui">
<title>GraphQL Playground</title>
<link rel="stylesheet" href="//cdn.jsdelivr.net/npm/graphql-playground-react/build/static/css/index.css" />
<link rel="shortcut icon" href="//cdn.jsdelivr.net/npm/graphql-playground-react/build/favicon.png" />
<script src="//cdn.jsdelivr.net/npm/graphql-playground-react/build/static/js/middleware.js"></script>
</head>
<body>
<div id="root">
<style>
body {
background-color: rgb(23, 42, 58);
font-family: Open Sans, sans-serif;
height: 90vh;
}
#root {
height: 100%;
width: 100%;
display: flex;
align-items: center;
justify-content: center;
}
.loading {
font-size: 32px;
font-weight: 200;
color: rgba(255, 255, 255, .6);
margin-left: 20px;
}
img {
width: 78px;
height: 78px;
}
.title {
font-weight: 400;
}
</style>
<img src='//cdn.jsdelivr.net/npm/graphql-playground-react/build/logo.png' alt=''>
<div class="loading"> Loading
<span class="title">GraphQL Playground</span>
</div>
</div>
<script>window.addEventListener('load', function (event) {
GraphQLPlayground.init(document.getElementById('root'), {
// options as 'endpoint' belong here
endpoint: "/api/graphql",
})
})</script>
</body>
</html>

View File

@@ -1,42 +0,0 @@
:root {
--active-brightness: 0.85;
--border-radius: 5px;
--box-shadow: 2px 2px 10px;
--color-accent: #118bee15;
--color-bg: #fff;
--color-bg-secondary: #e9e9e9;
--color-link: #118bee;
--color-secondary: #920de9;
--color-secondary-accent: #920de90b;
--color-shadow: #f4f4f4;
--color-table: #118bee;
--color-text: #000;
--color-text-secondary: #999;
--color-scrollbar: #cacae8;
--font-family: -apple-system, BlinkMacSystemFont, "Segoe UI", Roboto, Oxygen-Sans, Ubuntu, Cantarell, "Helvetica Neue", sans-serif;
--hover-brightness: 1.2;
--justify-important: center;
--justify-normal: left;
--line-height: 1.5;
/*
--width-card: 285px;
--width-card-medium: 460px;
--width-card-wide: 800px;
*/
--width-content: 1080px;
}
@media (prefers-color-scheme: dark) {
:root[color-mode="user"] {
--color-accent: #0097fc4f;
--color-bg: #333;
--color-bg-secondary: #555;
--color-link: #0097fc;
--color-secondary: #e20de9;
--color-secondary-accent: #e20de94f;
--color-shadow: #bbbbbb20;
--color-table: #0097fc;
--color-text: #f7f7f7;
--color-text-secondary: #aaa;
}
}

View File

@@ -1,93 +0,0 @@
<!DOCTYPE html>
<html>
<head>
<title>DMARC Report</title>
</head>
<body>
{% if report.report_metadata.is_some() %}
{% let meta = report.report_metadata.as_ref().unwrap() %}
<b>Reporter:</b> {{ meta.org_name }}<br>
<b>Contact:</b> {{ meta.email }}<br>
<b>Report ID:</b> {{ meta.report_id }}<br>
{% if meta.date_range.is_some() %}
{% let dr = meta.date_range.as_ref().unwrap() %}
<b>Date range:</b>
{{ dr.begin }}
to
{{ dr.end }}
<br>
{% endif %}
{% endif %}
{% if report.policy_published.is_some() %}
{% let pol = report.policy_published.as_ref().unwrap() %}
<b>Policy Published:</b>
<ul>
<li>Domain: {{ pol.domain }}</li>
<li>ADKIM: {{ pol.adkim }}</li>
<li>ASPF: {{ pol.aspf }}</li>
<li>Policy: {{ pol.p }}</li>
<li>Subdomain Policy: {{ pol.sp }}</li>
<li>Percent: {{ pol.pct }}</li>
</ul>
{% endif %}
{% if report.record.is_some() %}
<b>Records:</b>
<table style="border-collapse:collapse;width:100%;font-size:0.95em;">
<thead>
<tr style="background:#f0f0f0;">
<th style="border:1px solid #bbb;padding:4px 8px;">Source IP</th>
<th style="border:1px solid #bbb;padding:4px 8px;">Count</th>
<th style="border:1px solid #bbb;padding:4px 8px;">Header From</th>
<th style="border:1px solid #bbb;padding:4px 8px;">Disposition</th>
<th style="border:1px solid #bbb;padding:4px 8px;">DKIM</th>
<th style="border:1px solid #bbb;padding:4px 8px;">SPF</th>
<th style="border:1px solid #bbb;padding:4px 8px;">Auth Results</th>
</tr>
</thead>
<tbody>
{% for rec in report.record.as_ref().unwrap() %}
<tr>
<td style="border:1px solid #bbb;padding:4px 8px;">{{ rec.source_ip }}</td>
<td style="border:1px solid #bbb;padding:4px 8px;">{{ rec.count }}</td>
<td style="border:1px solid #bbb;padding:4px 8px;">{{ rec.header_from }}</td>
<td style="border:1px solid #bbb;padding:4px 8px;">{{ rec.disposition }}</td>
<td style="border:1px solid #bbb;padding:4px 8px;">{{ rec.dkim }}</td>
<td style="border:1px solid #bbb;padding:4px 8px;">{{ rec.spf }}</td>
<td style="border:1px solid #bbb;padding:4px 8px;">
{% if rec.auth_results.is_some() %}
{% let auth = rec.auth_results.as_ref().unwrap() %}
{% for dkimres in auth.dkim %}
<span style="white-space:nowrap;">
DKIM: domain=<b>{{ dkimres.domain }}</b>
selector=<b>{{ dkimres.selector }}</b>
result=<b>{{ dkimres.result }}</b>
</span><br>
{% endfor %}
{% for spfres in auth.spf %}
<span style="white-space:nowrap;">
SPF: domain=<b>{{ spfres.domain }}</b>
scope=<b>{{ spfres.scope }}</b>
result=<b>{{ spfres.result }}</b>
</span><br>
{% endfor %}
{% for reason in rec.reason %}
<span style="white-space:nowrap;">Reason: {{ reason }}</span><br>
{% endfor %}
{% endif %}
</td>
</tr>
{% endfor %}
</tbody>
</table>
{% endif %}
{% if report.report_metadata.is_none() && report.policy_published.is_none() && report.record.is_none() %}
<p>No DMARC summary found.</p>
{% endif %}
</body>
</html>

View File

@@ -1,43 +0,0 @@
<!DOCTYPE html>
<html>
<head>
<title>TLS Report</title>
</head>
<body>
<h3>TLS Report Summary:</h3>
<p>Organization: {{ report.organization_name }}</p>
<p>Date Range: {{ report.date_range.start_datetime }} to {{ report.date_range.end_datetime }}</p>
<p>Contact: {{ report.contact_info }}</p>
<p>Report ID: {{ report.report_id }}</p>
<h4>Policies:</h4>
{% for policy in report.policies %}
<h5>Policy Domain: {{ policy.policy.policy_domain }}</h5>
<ul>
<li>Policy Type: {{ policy.policy.policy_type }}</li>
<li>Policy String: {{ policy.policy.policy_string | join(", ") }}</li>
<li>Successful Sessions: {{ policy.summary.total_successful_session_count }}</li>
<li>Failed Sessions: {{ policy.summary.total_failure_session_count }}</li>
</ul>
<ul>
{% for mx_host in policy.policy.mx_host %}
<li>Hostname: {{ mx_host.hostname }}, Failures: {{ mx_host.failure_count }}, Result: {{ mx_host.result_type }}</li>
{% endfor %}
</ul>
<ul>
{% for detail in policy.failure_details %}
<li>Result: {{ detail.result_type }}, Sending IP: {{ detail.sending_mta_ip }}, Failed Sessions: {{ detail.failed_session_count }}
{% if detail.failure_reason_code != "" %}
(Reason: {{ detail.failure_reason_code }})
{% endif %}
</li>
(Receiving IP: {{ detail.receiving_ip }})
(Receiving MX: {{ detail.receiving_mx_hostname }})
(Additional Info: {{ detail.additional_info }})
{% endfor %}
</ul>
{% endfor %}
</body>
</html>

View File

@@ -1,20 +1,10 @@
[package]
name = "letterbox-shared"
description = "Shared module for letterbox"
authors.workspace = true
edition.workspace = true
license.workspace = true
publish.workspace = true
repository.workspace = true
version.workspace = true
name = "shared"
version = "0.1.0"
edition = "2021"
# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
[dependencies]
build-info = "0.0.41"
letterbox-notmuch = { path = "../notmuch", version = "0.17.32", registry = "xinu" }
regex = "1.11.1"
serde = { version = "1.0.219", features = ["derive"] }
sqlx = "0.8.5"
strum_macros = "0.27.1"
tracing = "0.1.41"
notmuch = { path = "../notmuch" }
serde = { version = "1.0.147", features = ["derive"] }

View File

@@ -1,14 +1,5 @@
use std::{
convert::Infallible,
hash::{DefaultHasher, Hash, Hasher},
str::FromStr,
};
use build_info::{BuildInfo, VersionControl};
use letterbox_notmuch::SearchSummary;
use regex::{RegexBuilder, RegexSetBuilder};
use notmuch::SearchSummary;
use serde::{Deserialize, Serialize};
use tracing::debug;
#[derive(Serialize, Deserialize, Debug)]
pub struct SearchResult {
@@ -18,251 +9,5 @@ pub struct SearchResult {
pub results_per_page: usize,
pub total: usize,
}
#[derive(Serialize, Deserialize, Debug, strum_macros::Display)]
pub enum WebsocketMessage {
RefreshMessages,
}
pub mod urls {
pub const MOUNT_POINT: &'static str = "/api";
pub fn view_original(host: Option<&str>, id: &str) -> String {
if let Some(host) = host {
format!("//{host}/api/original/{id}")
} else {
format!("/api/original/{id}")
}
}
pub fn cid_prefix(host: Option<&str>, cid: &str) -> String {
if let Some(host) = host {
format!("//{host}/api/cid/{cid}/")
} else {
format!("/api/cid/{cid}/")
}
}
pub fn download_attachment(host: Option<&str>, id: &str, idx: &str, filename: &str) -> String {
if let Some(host) = host {
format!(
"//{host}/api/download/attachment/{}/{}/{}",
id, idx, filename
)
} else {
format!("/api/download/attachment/{}/{}/{}", id, idx, filename)
}
}
}
pub fn build_version(bi: fn() -> &'static BuildInfo) -> String {
fn commit(git: &Option<VersionControl>) -> String {
let Some(VersionControl::Git(git)) = git else {
return String::new();
};
let mut s = vec!["-".to_string(), git.commit_short_id.clone()];
if let Some(branch) = &git.branch {
s.push(format!(" ({branch})"));
}
s.join("")
}
let bi = bi();
format!("v{}{}", bi.crate_info.version, commit(&bi.version_control)).to_string()
}
pub fn compute_color(data: &str) -> String {
let mut hasher = DefaultHasher::new();
data.hash(&mut hasher);
format!("#{:06x}", hasher.finish() % (1 << 24))
}
#[derive(
Copy, Clone, Debug, Default, PartialEq, Eq, Hash, Ord, PartialOrd, Serialize, Deserialize,
)]
pub enum MatchType {
From,
Sender,
To,
Cc,
Subject,
ListId,
DeliveredTo,
XForwardedTo,
ReplyTo,
XOriginalTo,
XSpam,
Body,
#[default]
Unknown,
}
#[derive(Debug, Default, Serialize, Deserialize)]
pub struct Match {
pub match_type: MatchType,
pub needle: String,
}
#[derive(Debug, Default, Serialize, Deserialize)]
pub struct Rule {
pub stop_on_match: bool,
pub matches: Vec<Match>,
pub tag: String,
}
impl Rule {
pub fn is_match(&self, header_key: &str, header_value: &str) -> bool {
let pats: Vec<_> = self
.matches
.iter()
.filter_map(|m| match m.match_type {
MatchType::To => Some("^(to|cc|bcc|x-original-to)$"),
MatchType::From => Some("^from$"),
MatchType::Sender => Some("^sender$"),
MatchType::Subject => Some("^subject$"),
MatchType::ListId => Some("^list-id$"),
MatchType::XOriginalTo => Some("^x-original-to$"),
MatchType::ReplyTo => Some("^reply-to$"),
MatchType::XSpam => Some("^x-spam$"),
MatchType::Body => None,
c => panic!("TODO handle '{c:?}' match type"),
})
.collect();
let set = RegexSetBuilder::new(&pats)
.case_insensitive(true)
.build()
.expect("failed to compile regex for matches");
let matches: Vec<_> = set.matches(header_key).into_iter().collect();
if !matches.is_empty() {
//info!("matched key '{header_key}' '{header_value}'");
for m_idx in matches {
let needle = regex::escape(&self.matches[m_idx].needle);
let pat = RegexBuilder::new(&needle)
.case_insensitive(true)
.build()
.expect("failed to compile regex for needle");
if pat.is_match(header_value) {
debug!("{header_key} matched {header_value} against {needle}");
return true;
}
}
}
false
}
}
mod matches {
// From https://linux.die.net/man/5/procmailrc
// If the regular expression contains '^TO_' it will be substituted by '(^((Original-)?(Resent-)?(To|Cc|Bcc)|(X-Envelope |Apparently(-Resent)?)-To):(.*[^-a-zA-Z0-9_.])?)'
// If the regular expression contains '^TO' it will be substituted by '(^((Original-)?(Resent-)?(To|Cc|Bcc)|(X-Envelope |Apparently(-Resent)?)-To):(.*[^a-zA-Z])?)', which should catch all destination specifications containing a specific word.
pub const TO: &'static str = "TO";
pub const CC: &'static str = "Cc";
pub const TOCC: &'static str = "(TO|Cc)";
pub const FROM: &'static str = "From";
pub const SENDER: &'static str = "Sender";
pub const SUBJECT: &'static str = "Subject";
pub const DELIVERED_TO: &'static str = "Delivered-To";
pub const X_FORWARDED_TO: &'static str = "X-Forwarded-To";
pub const REPLY_TO: &'static str = "Reply-To";
pub const X_ORIGINAL_TO: &'static str = "X-Original-To";
pub const LIST_ID: &'static str = "List-ID";
pub const X_SPAM: &'static str = "X-Spam";
pub const X_SPAM_FLAG: &'static str = "X-Spam-Flag";
}
impl FromStr for Match {
type Err = Infallible;
fn from_str(s: &str) -> Result<Self, Self::Err> {
// Examples:
// "* 1^0 ^TOsonyrewards.com@xinu.tv"
// "* ^TOsonyrewards.com@xinu.tv"
let mut it = s.split_whitespace().skip(1);
let mut needle = it.next().unwrap();
if needle == "1^0" {
needle = it.next().unwrap();
}
let mut needle = vec![needle];
needle.extend(it);
let needle = needle.join(" ");
let first = needle.chars().nth(0).unwrap_or(' ');
use matches::*;
if first == '^' {
let needle = &needle[1..];
if needle.starts_with(TO) {
return Ok(Match {
match_type: MatchType::To,
needle: cleanup_match(TO, needle),
});
} else if needle.starts_with(FROM) {
return Ok(Match {
match_type: MatchType::From,
needle: cleanup_match(FROM, needle),
});
} else if needle.starts_with(CC) {
return Ok(Match {
match_type: MatchType::Cc,
needle: cleanup_match(CC, needle),
});
} else if needle.starts_with(TOCC) {
return Ok(Match {
match_type: MatchType::To,
needle: cleanup_match(TOCC, needle),
});
} else if needle.starts_with(SENDER) {
return Ok(Match {
match_type: MatchType::Sender,
needle: cleanup_match(SENDER, needle),
});
} else if needle.starts_with(SUBJECT) {
return Ok(Match {
match_type: MatchType::Subject,
needle: cleanup_match(SUBJECT, needle),
});
} else if needle.starts_with(X_ORIGINAL_TO) {
return Ok(Match {
match_type: MatchType::XOriginalTo,
needle: cleanup_match(X_ORIGINAL_TO, needle),
});
} else if needle.starts_with(LIST_ID) {
return Ok(Match {
match_type: MatchType::ListId,
needle: cleanup_match(LIST_ID, needle),
});
} else if needle.starts_with(REPLY_TO) {
return Ok(Match {
match_type: MatchType::ReplyTo,
needle: cleanup_match(REPLY_TO, needle),
});
} else if needle.starts_with(X_SPAM_FLAG) {
return Ok(Match {
match_type: MatchType::XSpam,
needle: '*'.to_string(),
});
} else if needle.starts_with(X_SPAM) {
return Ok(Match {
match_type: MatchType::XSpam,
needle: '*'.to_string(),
});
} else if needle.starts_with(DELIVERED_TO) {
return Ok(Match {
match_type: MatchType::DeliveredTo,
needle: cleanup_match(DELIVERED_TO, needle),
});
} else if needle.starts_with(X_FORWARDED_TO) {
return Ok(Match {
match_type: MatchType::XForwardedTo,
needle: cleanup_match(X_FORWARDED_TO, needle),
});
} else {
unreachable!("needle: '{needle}'")
}
} else {
return Ok(Match {
match_type: MatchType::Body,
needle: cleanup_match("", &needle),
});
}
}
}
fn unescape(s: &str) -> String {
s.replace('\\', "")
}
pub fn cleanup_match(prefix: &str, s: &str) -> String {
unescape(&s[prefix.len()..]).replace(".*", "")
}
#[derive(Serialize, Deserialize, Debug)]
pub struct Message {}

View File

@@ -1,58 +1,39 @@
[package]
name = "letterbox-web"
description = "Web frontend for letterbox"
authors.workspace = true
edition.workspace = true
license.workspace = true
publish.workspace = true
repository.workspace = true
version.workspace = true
version = "0.1.0"
name = "letterbox"
repository = "https://github.com/seed-rs/seed-quickstart"
authors = ["Bill Thiede <git@xinu.tv>"]
description = "App Description"
categories = ["category"]
license = "MIT"
readme = "./README.md"
edition = "2018"
[build-dependencies]
build-info-build = "0.0.41"
[lib]
crate-type = ["cdylib"]
[dev-dependencies]
wasm-bindgen-test = "0.3.50"
wasm-bindgen-test = "0.3.33"
[dependencies]
console_error_panic_hook = "0.1.7"
log = "0.4.27"
seed = { version = "0.10.0", features = ["routing"] }
#seed = "0.9.2"
console_log = { version = "0.1.4", registry = "xinu" }
serde = { version = "1.0.219", features = ["derive"] }
itertools = "0.14.0"
serde_json = { version = "1.0.140", features = ["unbounded_depth"] }
chrono = "0.4.40"
graphql_client = "0.14.0"
thiserror = "2.0.12"
gloo-net = { version = "0.6.0", features = ["json", "serde_json"] }
human_format = "1.1.0"
build-info = "0.0.41"
wasm-bindgen = "=0.2.100"
uuid = { version = "1.16.0", features = [
"js",
] } # direct dep to set js feature, prevents Rng issues
letterbox-shared = { path = "../shared/", version = "0.17.32", registry = "xinu" }
seed_hooks = { version = "0.4.1", registry = "xinu" }
strum_macros = "0.27.1"
gloo-console = "0.3.0"
[target.'cfg(target_arch = "wasm32")'.dependencies]
wasm-sockets = "1.0.0"
log = "0.4.17"
seed = "0.9.2"
console_log = {git = "http://git-private.h.xinu.tv/wathiede/console_log.git"}
serde = { version = "1.0.147", features = ["derive"] }
notmuch = {path = "../notmuch"}
shared = {path = "../shared"}
itertools = "0.10.5"
serde_json = { version = "1.0.93", features = ["unbounded_depth"] }
wasm-timer = "0.2.5"
css-inline = "0.8.5"
[package.metadata.wasm-pack.profile.release]
wasm-opt = ['-Os']
[dependencies.web-sys]
version = "0.3.77"
version = "0.3.58"
features = [
"Clipboard",
"DomRect",
"Element",
"History",
"MediaQueryList",
"Navigator",
"Performance",
"ScrollRestoration",
"Window",
"Window"
]

6
web/Makefile Normal file
View File

@@ -0,0 +1,6 @@
.PHONY: all
# Build in release mode and push to minio for serving.
all:
trunk build --release
mc mirror --overwrite --remove dist/ m/letterbox/

View File

@@ -1,5 +1,5 @@
[build]
release = false
release = true
[serve]
# The address to serve on.
@@ -7,21 +7,5 @@ address = "0.0.0.0"
port = 6758
[[proxy]]
ws = true
backend = "ws://localhost:9345/api/ws"
[[proxy]]
backend = "http://localhost:9345/api/"
[[proxy]]
backend = "http://localhost:9345/notification/"
[[hooks]]
stage = "pre_build"
command = "printf"
command_arguments = ["\\033c"]
#[[hooks]]
#stage = "pre_build"
#command = "cargo"
#command_arguments = [ "test" ]
backend = "http://localhost:9345/"
rewrite= "/api/"

View File

@@ -1,5 +0,0 @@
fn main() {
// Calling `build_info_build::build_script` collects all data and makes it available to `build_info::build_info!`
// and `build_info::format!` in the main program.
build_info_build::build_script();
}

View File

@@ -1,3 +0,0 @@
mutation AddTagMutation($query: String!, $tag: String!) {
tagAdd(query:$query, tag:$tag)
}

View File

@@ -1,3 +0,0 @@
query CatchupQuery($query: String!) {
catchup(query: $query)
}

View File

@@ -1,27 +0,0 @@
query FrontPageQuery($query: String!, $after: String $before: String, $first: Int, $last: Int) {
count(query: $query)
search(query: $query, after: $after, before: $before, first: $first, last: $last) {
pageInfo {
hasPreviousPage
hasNextPage
startCursor
endCursor
}
nodes {
thread
total
timestamp
subject
authors
tags
corpus
}
}
tags {
name
bgColor
fgColor
unread
}
version
}

View File

@@ -1,3 +0,0 @@
mutation MarkReadMutation($query: String!, $unread: Boolean!) {
setReadStatus(query:$query, unread:$unread)
}

View File

@@ -1,3 +0,0 @@
mutation RefreshMutation {
refresh
}

View File

@@ -1,3 +0,0 @@
mutation RemoveTagMutation($query: String!, $tag: String!) {
tagRemove(query:$query, tag:$tag)
}

File diff suppressed because it is too large Load Diff

View File

@@ -1,77 +0,0 @@
query ShowThreadQuery($threadId: String!) {
thread(threadId: $threadId) {
__typename ... on NewsPost{
threadId
isRead
slug
site
title
body
url
timestamp
# TODO: unread
}
__typename ... on EmailThread{
threadId,
subject
messages {
id
subject
tags
from {
name
addr
photoUrl
}
to {
name
addr
}
cc {
name
addr
}
xOriginalTo {
name
addr
}
deliveredTo {
name
addr
}
timestamp
body {
__typename
... on UnhandledContentType {
contents
contentTree
}
... on PlainText {
contents
contentTree
}
... on Html {
contents
contentTree
}
}
path
attachments {
id
idx
filename
contentType
contentId
size
}
}
}
}
tags {
name
bgColor
fgColor
unread
}
version
}

View File

@@ -1,4 +0,0 @@
DEV_HOST=localhost
DEV_PORT=9345
graphql-client introspect-schema http://${DEV_HOST:?}:${DEV_PORT:?}/api/graphql --output schema.json
git diff schema.json

View File

@@ -2,25 +2,100 @@
<html lang="en">
<head>
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1, shrink-to-fit=no">
<link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/6.7.2/css/all.min.css"
integrity="sha512-Evv84Mr4kqVGRNSgIGL/F/aIDqQb7xQ2vcrdIwxfjThSH8CSR7PBEakCr51Ck+w+/U6swU2Im1vVX0SVk9ABhg=="
crossorigin="anonymous" referrerpolicy="no-referrer" />
<link rel="icon" href="https://static.xinu.tv/favicon/letterbox.svg" />
<!-- tall thin font for user icon -->
<link rel="preconnect" href="https://fonts.googleapis.com">
<link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>
<link href="https://fonts.googleapis.com/css2?family=Poppins:wght@700&display=swap" rel="stylesheet">
<!-- <link data-trunk rel="css" href="static/site-specific.css" /> -->
<link data-trunk rel="css" href="static/vars.css" />
<link data-trunk rel="tailwind-css" href="./src/tailwind.css" />
<link data-trunk rel="css" href="static/overrides.css" />
<link data-trunk rel="css" href="static/email-specific.css" />
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1, shrink-to-fit=no">
<link rel="stylesheet", href="https://jenil.github.io/bulmaswatch/cyborg/bulmaswatch.min.css">
<link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/6.3.0/css/all.min.css" integrity="sha512-SzlrxWUlpfuzQ+pcUCosxcglQRNAq/DZjVsC0lE40xsADsfeQoEypE+enwcOiGjk/bSuGGKHEyjSoQ1zVisanQ==" crossorigin="anonymous" referrerpolicy="no-referrer" />
<style>
.message {
padding: 0.5em;*/
}
.body {
background: white;
color: black;
padding: 0.5em;
margin-left: -0.5em;
margin-right: -0.5em;
margin-top: 0.5em;
}
.error {
background-color: red;
}
.view-part-text-plain {
white-space: pre-line;
}
iframe {
height: 100%;
width: 100%;
}
.index {
table-layout: fixed;
width: 100%;
}
.index .from {
overflow: hidden;
text-overflow: ellipsis;
white-space: nowrap;
width: 15em;
}
.index .subject {
overflow: hidden;
text-overflow: ellipsis;
white-space: nowrap;
}
.index .date {
width: 8em;
white-space: nowrap;
}
.footer {
background-color: #eee;
color: #222;
position: fixed;
bottom: 0;
left: 0;
right: 0;
height: 3em;
padding: 1em;
}
.tag {
margin-right: 2px;
}
.debug ul {
padding-left: 2em;
}
.debug li {
}
.loading {
animation-name: spin;
animation-duration: 1000ms;
animation-iteration-count: infinite;
animation-timing-function: linear;
}
@keyframes spin {
from {
transform:rotate(0deg);
}
to {
transform:rotate(360deg);
}
}
@media (max-width: 768px) {
.section {
padding: 1.5em;
}
}
input, .input {
color: #000;
}
input::placeholder, .input::placeholder{
color: #555;
}
</style>
</head>
<body>
<section id="app"></section>
<section id="app"></section>
</body>
</html>

View File

@@ -1,14 +0,0 @@
pub mod urls {
use seed::Url;
pub fn search(query: &str, page: usize) -> Url {
let query = Url::encode_uri_component(query);
if page > 0 {
Url::new().set_hash_path(["s", &query, &format!("p{page}")])
} else {
Url::new().set_hash_path(["s", &query])
}
}
pub fn thread(tid: &str) -> Url {
Url::new().set_hash_path(["t", tid])
}
}

View File

@@ -1 +0,0 @@
pub const SEARCH_RESULTS_PER_PAGE: usize = 20;

Some files were not shown because too many files have changed in this diff Show More