mirror of
https://github.com/kuhyx/signal-bot.git
synced 2026-11-10 19:27:20 +01:00
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
135 lines
5.3 KiB
Python
135 lines
5.3 KiB
Python
# Copyright (c) 2026 Krzysztof Rudnicki. MIT licence.
|
|
"""What to do with a "!" word the registry does not recognise.
|
|
|
|
Until now the answer was "nothing at all", which conflated three unrelated
|
|
situations the dispatcher could not tell apart: a word nobody registered, a
|
|
word registered to some *other* group, and a real trigger with junk after it.
|
|
The first two deserve an answer. The third is "!kot ładny", which has always
|
|
been silent and stays silent -- a bot that pipes up every time somebody adds
|
|
a word to a working command is a bot people mute.
|
|
|
|
Silence is decided by the *global* question -- does anything, in any group,
|
|
claim this word -- and the near-miss search then reads only the triggers live
|
|
in the asking group. Both halves are needed. The first keeps the album's
|
|
triggers, which are real people's names, from being echoed or recorded in a
|
|
group that has no album; the second keeps them out of a suggestion.
|
|
"""
|
|
|
|
from __future__ import annotations
|
|
|
|
import logging
|
|
import re
|
|
from dataclasses import dataclass
|
|
from typing import TYPE_CHECKING, Final
|
|
|
|
from signal_bot.folding import fold
|
|
from signal_bot.nearest import nearest
|
|
from signal_bot.quoting import quote_of
|
|
|
|
if TYPE_CHECKING:
|
|
from collections.abc import Mapping
|
|
|
|
from signal_bot.models import IncomingMessage
|
|
from signal_bot.registry import Command, CommandRegistry
|
|
from signal_bot.store import Store
|
|
from signal_bot.transport_base import Transport
|
|
|
|
logger = logging.getLogger(__name__)
|
|
|
|
# What may count as a mistyped command. Anything else -- a bare "!", "!!",
|
|
# "!?", an emoji, a pasted paragraph -- is ordinary punctuation and is
|
|
# ignored entirely: no reply and no row.
|
|
#
|
|
# Digits are in the class on purpose. fold() maps 0134578 onto letters but
|
|
# leaves 2, 6 and 9 alone, so "!rok2" really does arrive with a digit in it,
|
|
# and a letters-only guard would swallow a perfectly plausible typo.
|
|
WORD_SHAPE: Final = re.compile(r"^![a-z0-9_]{2,32}$")
|
|
|
|
# The spelling every one of these messages points people at. It lives here
|
|
# rather than in suggest_command because that module needs WORD_SHAPE from
|
|
# this one, and two modules importing each other is not a seam. A test
|
|
# pins it against aliases.json so the two cannot drift.
|
|
SUGGEST_PRIMARY: Final = "!zaproponuj"
|
|
|
|
DID_YOU_MEAN: Final = "Nie znam {word}. Chodziło o {primary}?"
|
|
NO_MATCH: Final = "Nie znam {word}. Napisz {hint} {word} <co ma robić>, a zapiszę pomysł."
|
|
|
|
|
|
@dataclass(frozen=True, slots=True)
|
|
class Unknown:
|
|
"""A "!" word nothing claims here, and the message that carried it."""
|
|
|
|
word: str
|
|
body: str
|
|
|
|
|
|
def unknown_in(text: str | None, registry: CommandRegistry) -> Unknown | None:
|
|
"""Read an unrecognised command out of a message, if there is one.
|
|
|
|
Args:
|
|
text: The raw message body.
|
|
registry: The command registry, asked the *global* question.
|
|
|
|
Returns:
|
|
The unrecognised word and the message it came in, or None when the
|
|
message is not a command attempt at all.
|
|
"""
|
|
if not text:
|
|
return None
|
|
words = text.split()
|
|
if not words:
|
|
return None
|
|
word = fold(words[0])
|
|
if WORD_SHAPE.match(word) is None:
|
|
return None
|
|
# Claimed anywhere at all, not merely here -- ``resolve`` with no group
|
|
# is the same question pack_add asks before minting a label. This one
|
|
# check covers two very different cases:
|
|
#
|
|
# "!kot ładny" a real trigger with words after it. resolve_with_args
|
|
# already decided that means nothing for a command
|
|
# taking no arguments, and it is not this file's to
|
|
# revisit.
|
|
# "!alfa" an album command, in a group that has no album. The
|
|
# album's triggers are real people's names, and
|
|
# pack_scoping says it out loud: "a refusal would itself
|
|
# announce that the album exists". Answering here would
|
|
# announce it, and recording the row would put the name
|
|
# in the database and on the management page.
|
|
if registry.resolve(word) is not None:
|
|
return None
|
|
return Unknown(word=word, body=text)
|
|
|
|
|
|
def reply_for(word: str, live: Mapping[str, Command]) -> str:
|
|
"""What to say about one unrecognised word.
|
|
|
|
The suggestion prints ``Command.primary`` -- the same spelling !pomoc
|
|
shows -- rather than the folded trigger it matched, which is a skeleton
|
|
with the diacritics stripped off and reads as a second typo.
|
|
"""
|
|
hit = nearest(word, live)
|
|
if hit is None:
|
|
return NO_MATCH.format(word=word, hint=SUGGEST_PRIMARY)
|
|
return DID_YOU_MEAN.format(word=word, primary=live[hit].primary)
|
|
|
|
|
|
async def answer_unknown(
|
|
*,
|
|
message: IncomingMessage,
|
|
group_id: str,
|
|
registry: CommandRegistry,
|
|
store: Store,
|
|
transport: Transport,
|
|
) -> None:
|
|
"""Remember one unrecognised command and answer it, or do neither."""
|
|
unknown = unknown_in(message.text, registry)
|
|
if unknown is None:
|
|
return
|
|
live = registry.triggers_in(group_id)
|
|
# No word and no body in the log line. A misspelled name is still a
|
|
# name, log lines reach CI output, and the database is the record.
|
|
logger.debug("unrecognised command in %s", group_id)
|
|
store.suggestions.record_unknown(group_id, unknown.word, unknown.body, message.timestamp)
|
|
await transport.send_text(reply_for(unknown.word, live), group_id, quote_of(message))
|