signal-bot/tests/test_unknown.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

178 lines
6.3 KiB
Python

# Copyright (c) 2026 Krzysztof Rudnicki. MIT licence.
"""What the bot says about a "!" word it does not know, and what it keeps."""
from __future__ import annotations
from typing import TYPE_CHECKING
import pytest
from signal_bot.models import parse_envelope
from signal_bot.registry import CommandRegistry
from signal_bot.unknown import (
SUGGEST_PRIMARY,
answer_unknown,
reply_for,
unknown_in,
)
from tests.conftest import SECOND_GROUP, WATCHED_GROUP, envelope
if TYPE_CHECKING:
from signal_bot.context import CommandContext
from signal_bot.store import Store
from signal_bot.transport_fake import FakeTransport
async def _noop(ctx: CommandContext) -> None:
"""A handler that does nothing."""
def loaded() -> CommandRegistry:
"""A registry with one global command and one scoped to a single group."""
registry = CommandRegistry()
registry.register("cat", ["!kot", "!cat"], _noop, "kot")
# The album's shape: scoped, and named after a person.
registry.register("pack:p1", ["!alfa"], _noop, "zdjęcia", groups=frozenset({WATCHED_GROUP}))
return registry
@pytest.mark.parametrize("text", [None, "", " ", "just chatting", "hej! super"])
def test_a_message_that_is_not_a_command_attempt_is_not_one(text: str | None) -> None:
"""Nothing without a leading "!" word reaches the rest of this module."""
assert unknown_in(text, loaded()) is None
@pytest.mark.parametrize("text", ["!", "!!", "!?", "!...", "!😀", "!x"])
def test_punctuation_shaped_like_a_command_is_ignored(text: str) -> None:
"""No reply and no row: this is ordinary typing, not a mistyped command."""
assert unknown_in(text, loaded()) is None
def test_a_real_trigger_with_words_after_it_stays_silent() -> None:
""" "!kot ładny" has always been silent and this feature does not change it."""
assert unknown_in("!kot ładny", loaded()) is None
def test_a_command_scoped_to_another_group_is_never_reported() -> None:
"""The album's privacy guarantee.
"!alfa" is a person's name. In a group with no album it is unclaimed
*there*, so a check scoped to the asking group would answer it and store
it -- announcing the album and putting the name in the database. The
check is global, so this is silence.
"""
assert unknown_in("!alfa", loaded()) is None
def test_an_unclaimed_word_is_reported_with_the_whole_message() -> None:
"""The word is folded; the body is kept exactly as it was typed."""
found = unknown_in("!CTA teraz proszę", loaded())
assert found is not None
assert found.word == "!cta"
assert found.body == "!CTA teraz proszę"
def test_a_styled_or_cyrillic_spelling_folds_before_anything_else() -> None:
"""The same folding the registry does, so both sides can only agree."""
found = unknown_in("!ᴄᴛᴀ", loaded())
assert found is not None
assert found.word == "!cta"
def test_a_digit_survives_folding_and_is_still_recorded() -> None:
"""fold() maps 0134578 onto letters but leaves 2, 6 and 9 alone.
A letters-only shape guard would silently drop "!rok2", which is exactly
the kind of request this feature exists to collect.
"""
found = unknown_in("!rok2", loaded())
assert found is not None
assert found.word == "!rok2"
def test_a_near_miss_is_answered_with_the_polish_spelling() -> None:
"""`triggers` are folded skeletons; only `primary` still reads correctly."""
registry = CommandRegistry()
registry.register("pope", ["!papież", "!pope"], _noop, "papież")
live = registry.triggers_in(WATCHED_GROUP)
assert "!papież" in reply_for("!papiesz", live)
assert "!papiez?" not in reply_for("!papiesz", live)
def test_a_word_with_no_near_miss_points_at_the_suggestion_command() -> None:
"""Otherwise there is nothing for the person to do with the refusal."""
text = reply_for("!nosuchcommand", loaded().triggers_in(WATCHED_GROUP))
assert SUGGEST_PRIMARY in text
assert "!nosuchcommand" in text
def test_a_suggestion_never_names_a_command_from_another_group() -> None:
""" "!alfaa" must not come back as "chodziło o !alfa?" where there is no album.
Asserted on the correction clause rather than on the bare name, because
the refusal quotes the word that was typed and "!alfa" is a substring of
"!alfaa" -- a naive check passes for the wrong reason.
"""
registry = loaded()
live = registry.triggers_in(SECOND_GROUP)
assert "!alfa" not in live
answer = reply_for("!alfaa", live)
assert "Chodziło" not in answer
assert SUGGEST_PRIMARY in answer
# ...and the album's own group is where the correction is allowed.
assert "Chodziło" in reply_for("!alfaa", registry.triggers_in(WATCHED_GROUP))
async def test_answering_records_the_token_and_the_body(
store: Store, transport: FakeTransport
) -> None:
"""One row, one reply, keyed on the envelope's own timestamp."""
message = parse_envelope(envelope(text="!cta", timestamp=4242))
assert message is not None
await answer_unknown(
message=message,
group_id=WATCHED_GROUP,
registry=loaded(),
store=store,
transport=transport,
)
assert len(transport.sent) == 1
assert "!kot" in (transport.sent[0].text or "")
(row,) = store.suggestions.unknown()
assert row.word == "!cta"
assert row.bodies == ("!cta",)
assert row.last_seen == 4242
async def test_answering_quotes_the_person_who_asked(
store: Store, transport: FakeTransport
) -> None:
"""A busy group has to be able to tell which "!cta" this is about."""
message = parse_envelope(envelope(text="!cta"))
assert message is not None
await answer_unknown(
message=message,
group_id=WATCHED_GROUP,
registry=loaded(),
store=store,
transport=transport,
)
assert transport.sent[0].quote is not None
async def test_a_message_that_is_not_an_attempt_writes_nothing(
store: Store, transport: FakeTransport
) -> None:
"""The early return, from the caller's side."""
message = parse_envelope(envelope(text="just chatting"))
assert message is not None
await answer_unknown(
message=message,
group_id=WATCHED_GROUP,
registry=loaded(),
store=store,
transport=transport,
)
assert transport.sent == []
assert store.suggestions.unknown() == []