signal-bot/signal_bot/suggestions.py
Krzysztof kuhy Rudnicki 6ec746fad2 feat: a word the bot does not know gets an answer, and gets kept
An unrecognised "!word" was dropped in silence, which made a typo
indistinguishable from the bot being down and threw away the best available
list of commands worth writing. Three things now happen at the one place in
dispatcher.py that used to be a bare `return`.

A near miss is corrected -- "!cta" asks whether you meant "!kot". The
distance is optimal string alignment rather than difflib's ratio: a
transposition is the typo people actually make, it costs one edit here, and
it shreds the matching blocks a ratio counts ("!cta"/"!cat" scores 0.75 but
"!cta"/"!kot" scores 0.50, so no cutoff separates them on principle).

A word with no near miss points at !zaproponuj, which files an idea from
inside the group. Either way the word and the message are recorded, and the
management page grows an Ideas tab that lists them.

Three things stay silent, each for its own reason. "!kot ladny" is a real
trigger with words after it and always has been silent. "!!" is punctuation.
And a command claimed in some *other* group is neither answered nor recorded
-- the album's triggers are real people's names, and pack_scoping.yaml says
it out loud: a refusal would itself announce that the album exists. The
existence check is therefore global, which is what pack_add already asks.

Measured rather than guessed, against the shipped aliases.json:

- a minimum candidate length, because "!ja" is one edit from ~40 of the 676
  two-letter words somebody could type; 69 of them got a suggestion before
  the rule and 18 after.
- the two-edit budget starts at six characters, not five. At five it put
  four pairs of *different* commands within reach of each other --
  !pies/!popes, !days/!cats, !dzis/!dni, !mnie/!dni -- and costs nothing to
  give up, because every real typo this was written for is one edit away.
  test_no_two_commands_are_within_reach_of_each_other fails if a future
  alias reopens the gap.

Both new tables are keyed on the envelope's own sent_at_ms, never the wall
clock: signal-cli replays undelivered frames on reconnect, and a clock-keyed
table collects a duplicate row per reconnect instead of ignoring one.

The branch is gated on should_count -- which excludes the bot's own envelope,
and that is the loop guard, since handle() checks is_own nowhere else -- and
wrapped in its own try/except, because it runs above the one that wraps a
handler and Bot.listen has no exception handling of its own.

Suggestions live in SQLite, not in a tracked file, and the export goes to the
state directory. check_personal_data learns the names it refuses from the
live database, so it catches a name somebody has typed and cannot catch a
misspelling of one -- and an unrecognised-command corpus is made of
misspellings. A group-writable file in a public tree is both a leak and a
wedge. The export therefore carries
deliberate !zaproponuj notes in full and unrecognised words as counts only;
bodies stay on the loopback-only page.

Verified on the live bot: restarted, both tables created with every existing
row count unchanged, !cta and !zaproponuj !pogoda delivered through
/_selftest, both rows read back out of the live database, and the journal
names neither the word nor the body.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01416TegJ3sa8xL5qGGgrygP
2026-08-31 16:48:26 +02:00

127 lines
4.6 KiB
Python

# Copyright (c) 2026 Krzysztof Rudnicki. MIT licence.
"""What the group wished the bot could do, kept where git cannot reach it.
Two tables and one deliberate asymmetry between them. ``suggestions`` holds
notes somebody wrote *on purpose*, with !zaproponuj, knowing they were
writing to whoever maintains the bot. ``unknown_commands`` holds every "!"
word the registry did not recognise, body and all, which is a far less
considered kind of text -- it is full of half-typed names, and a misspelled
name is one the personal-data gate cannot know to refuse.
That asymmetry is why this lives in SQLite rather than in a tracked file.
The database is gitignored; the export the management page offers carries
the deliberate notes in full and the unrecognised words as counts only.
Shares the Store's connection, the same way stats.events.EventStore does.
"""
from __future__ import annotations
from dataclasses import dataclass
from typing import TYPE_CHECKING
if TYPE_CHECKING:
import sqlite3
@dataclass(frozen=True, slots=True)
class Suggestion:
"""One idea somebody filed with !zaproponuj."""
group_id: str
command: str
body: str
sent_at_ms: int
@dataclass(frozen=True, slots=True)
class UnknownTally:
"""One unrecognised word, how often it was tried, and what was typed."""
group_id: str
word: str
hits: int
last_seen: int
bodies: tuple[str, ...]
class SuggestionStore:
"""Reads and writes the wish list. Shares the Store's connection."""
def __init__(self, conn: sqlite3.Connection) -> None:
"""Wrap an open connection whose schema has already been applied."""
self._conn = conn
def record_unknown(self, group_id: str, word: str, body: str, sent_at_ms: int) -> None:
"""Note one unrecognised "!" word, ignoring a redelivery of it.
``OR IGNORE`` against the envelope's own timestamp, not the wall
clock. signal-cli replays undelivered frames when the bot
reconnects; keyed on a clock reading every replay looks new, and
without OR IGNORE a genuine collision raises IntegrityError out of
the dispatcher, where there is nothing above it to catch one.
"""
self._conn.execute(
"INSERT OR IGNORE INTO unknown_commands"
" (group_id, word, body, sent_at_ms) VALUES (?, ?, ?, ?)",
(group_id, word, body, sent_at_ms),
)
self._conn.commit()
def record_suggestion(self, group_id: str, command: str, body: str, sent_at_ms: int) -> None:
"""File one idea against the command name it is asking for."""
self._conn.execute(
"INSERT OR IGNORE INTO suggestions"
" (group_id, command, body, sent_at_ms) VALUES (?, ?, ?, ?)",
(group_id, command, body, sent_at_ms),
)
self._conn.commit()
def filed(self) -> list[Suggestion]:
"""Every filed suggestion, newest first."""
rows = self._conn.execute(
"SELECT group_id, command, body, sent_at_ms FROM suggestions"
" ORDER BY sent_at_ms DESC, command ASC"
)
return [
Suggestion(
group_id=str(row["group_id"]),
command=str(row["command"]),
body=str(row["body"]),
sent_at_ms=int(row["sent_at_ms"]),
)
for row in rows
]
def unknown(self) -> list[UnknownTally]:
"""Every unrecognised word, most-tried first, with what was typed.
Two passes rather than one grouped query with ``group_concat``: a
body may itself contain the separator, and a tally that silently
splits somebody's sentence in half is worse than a second SELECT.
"""
tallies = self._conn.execute(
"SELECT group_id, word, COUNT(*) AS hits, MAX(sent_at_ms) AS last_seen"
" FROM unknown_commands GROUP BY group_id, word"
" ORDER BY hits DESC, last_seen DESC, word ASC"
).fetchall()
return [
UnknownTally(
group_id=str(row["group_id"]),
word=str(row["word"]),
hits=int(row["hits"]),
last_seen=int(row["last_seen"]),
bodies=self._bodies(str(row["group_id"]), str(row["word"])),
)
for row in tallies
]
def _bodies(self, group_id: str, word: str) -> tuple[str, ...]:
"""The distinct messages that produced one unrecognised word."""
rows = self._conn.execute(
"SELECT DISTINCT body FROM unknown_commands"
" WHERE group_id = ? AND word = ? ORDER BY body ASC",
(group_id, word),
)
return tuple(str(row["body"]) for row in rows)