signal-bot/scripts/signal-bot.service
Krzysztof kuhy Rudnicki aefffb2fd9
Some checks are pending
CI / check (push) Waiting to run
feat: add a photo to the album from the chat, tagged, stripped and verified
Send a photo captioned "!add_photo alfa wakacje" in one of the album's own
groups and it lands in the pack tagged with both, retrievable through !alfa
and the new !losowe immediately -- no restart, and no tagger session on the PC.
A label matching nobody becomes a custom tag, and !losowe is the only door to
one, which is what keeps a live swap a *data* swap: it can never register a
trigger or widen a command's groups.

No typed label travels. Labels become pseudonym keys in a new registry
arg_resolver, at the dispatcher, before the line that logs the command name --
so no handler, journal entry, command_uses row or reply body ever holds a word
somebody typed. The reply is counts and pseudonyms: "dodano · p1: 4 · t1: 1".

Stripping is exiftool with a *positive* check on the output, never an exit
code. ffmpeg was tried first and fails silently on photos: given a JPEG with a
real APP1 Exif segment it returned a byte-identical file, GPS intact, exit 0.
The surviving-tag allowlist is per tag *name* rather than per group, because
measurement showed a group-level "must be empty" rule rejects every format
except JPEG -- a valid PNG always reports its IHDR, an MP4 some fifty QuickTime
tags that are the container itself. Codec configuration (CompressorName,
VendorID, the HEVC parameter set) is on the list by decision: CompressorName
lives in the avc1 sample description and exiftool cannot write it at all, so
the alternative was refusing every video and every iPhone HEIC. Location and
identity tags are still refused, and the e2e run re-reads the stored files to
prove a QuickTime geotag did not survive.

.webm and .svg stay on the format allowlist and are stored *unstripped*:
exiftool 13.55 can write neither, and its SVG reader does not even surface
<metadata>. That is a deliberate, documented hole, not an oversight.

Two things this had to defuse along the way:
- personal_data_sources now reads custom.json, so "!add_photo kot <photo>"
  would have outlawed the word "kot" and refused every future commit, fixable
  only by hand-editing an album the committer cannot see. Minting now refuses
  a word that is already a command, and the gate subtracts the repo's own
  alias vocabulary from custom labels only -- curated people.json names are
  never excused.
- signal_bot/external.py is now the single subprocess site in the repository,
  so ruff's S603 needs exactly one scoped per-file-ignore instead of one per
  module that shells out. A test asserts nothing else imports subprocess.

A failed add changes nothing: media and the three manifests are written, the
album is reloaded, and only a load that succeeds is published; a load that
fails rolls all four back and keeps serving the pack that was already working,
without raising -- Restart=always would turn a ConfigError into a loop.

977 tests, 100% branch coverage, no omits and no suppressions. scripts/e2e.sh
gains 8 scenarios driven against the real exiftool on media generated at run
time, and asserts the invariants a reply cannot: that a refused upload is not
stored, that a message with no attachment leaves custom.json byte-identical,
and that signal-cli's own copy of every ingested attachment is gone.

Closes TODO-add-photo.md.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01JqqE22jp94B1ypwxppdr1o
2026-08-31 11:20:00 +02:00

57 lines
2.5 KiB
Desktop File

# Installed to ~/.config/systemd/user/ by scripts/install_service.sh, which
# substitutes @REPO_ROOT@. Edit this template, never the installed copy.
[Unit]
Description=Signal group bot
Documentation=file://@REPO_ROOT@/README.md
# The bot's only outbound dependency is signal-cli-rest-api in Docker. This
# ordering is worth having and is NOT a fix: dockerd being up says nothing
# about whether the container inside it has finished starting, and signal-cli
# then spends further minutes loading the account. systemd cannot express
# "after that container is ready", so the bot retries the connection itself
# (signal_bot/startup.py) and serves an unhealthy /healthz meanwhile.
After=network-online.target docker.service
Wants=network-online.target
# The start limit now only has to catch answers, not outages. A config error
# -- a missing PHONE_NUMBER, a taken port, a container in the wrong mode --
# exits almost instantly and would otherwise spin forever, so five of those
# stop the unit in `failed`, which shows up in `systemctl --user --failed`.
#
# The arithmetic: 5 starts x RestartSec=30 is 120s, comfortably inside the
# 300s window, so a genuine config error still trips the limit. What used to
# trip it was 5 x 5s = 25s of a container being slow to boot; on 2026-08-30
# that left the group with no bot for eight hours.
StartLimitIntervalSec=300
StartLimitBurst=5
[Service]
Type=exec
WorkingDirectory=@REPO_ROOT@
# run.sh execs the interpreter, so this PID is the bot itself.
ExecStart=@REPO_ROOT@/run.sh
# `always`, not `on-failure`: the bot never exits on its own except when
# startup hits something no retry can fix, and that path asks uvicorn to
# shut down cleanly -- so exit 0 must still count as a restart.
Restart=always
RestartSec=30
# The bot writes its SQLite tally database inside the repo, and papal
# portraits into the cache directory. CacheDirectory= is load-bearing:
# ProtectHome=read-only makes the rest of $HOME immutable, and systemd
# exempts only the directory it created, exporting it as $CACHE_DIRECTORY.
CacheDirectory=signal-bot
NoNewPrivileges=true
PrivateTmp=true
ProtectSystem=strict
ProtectHome=read-only
ReadWritePaths=@REPO_ROOT@
# The photo album. ProtectHome=read-only covers it too, and !add_photo
# writes there -- media, tags.json, custom.json and adds.json. The pack
# does not move; this is the one path that has to become writable.
ReadWritePaths=%h/.local/share/signal-bot
ProtectKernelTunables=true
ProtectControlGroups=true
RestrictSUIDSGID=true
[Install]
WantedBy=default.target