feat(backend): watches fire on new, edited and deleted messages and on coming online, legacy keyword included

backend/src/utils/watch_rules.py

@@ -0,0 +1,145 @@
+import re
+from collections.abc import Iterable, Iterator
+from importlib import import_module
+from typing import Any, Literal, Self, get_args
+
+from pydantic import BaseModel, Field, ValidationError, model_validator
+
+MAX_PATTERN = 200
+MAX_TEXT = 4096
+MAX_NAME = 120
+
+WatchKind = Literal[
+    "keyword", "new_message", "message_deleted", "message_edited", "peer_online"
+]
+WATCH_KINDS: tuple[str, ...] = get_args(WatchKind)
+
+sre_parser = import_module("re._parser")
+_REPEATS = {sre_parser.MAX_REPEAT, sre_parser.MIN_REPEAT, sre_parser.POSSESSIVE_REPEAT}
+
+
+class WatchParamsError(ValueError):
+    pass
+
+
+def _subpatterns(value: object) -> Iterator[Any]:
+    if isinstance(value, sre_parser.SubPattern):
+        yield value
+    elif isinstance(value, tuple | list):
+        for item in value:
+            yield from _subpatterns(item)
+
+
+def _repeat_depth(items: Iterable[tuple[Any, Any]]) -> int:
+    depth = 0
+    for op, value in items:
+        if op in _REPEATS:
+            low, high, body = value
+            depth = max(depth, _repeat_depth(body) + (high > max(low, 1)))
+        else:
+            inner = (_repeat_depth(sub) for sub in _subpatterns(value))
+            depth = max(depth, *inner, 0)
+    return depth
+
+
+def compile_pattern(pattern: str) -> re.Pattern[str]:
+    try:
+        compiled = re.compile(pattern, re.IGNORECASE)
+    except re.error as exc:
+        msg = f"invalid regex: {exc}"
+        raise ValueError(msg) from exc
+    if _repeat_depth(sre_parser.parse(pattern)) > 1:
+        msg = "regex repeats inside a repeat, like (a+)+, which can hang"
+        raise ValueError(msg)
+    return compiled
+
+
+class KeywordParams(BaseModel):
+    pattern: str = Field(min_length=1, max_length=MAX_PATTERN)
+    regex: bool = False
+    chat_id: int | None = None
+    sender_id: int | None = None
+
+    @model_validator(mode="before")
+    @classmethod
+    def _legacy_keyword(cls, data: object) -> object:
+        if isinstance(data, dict) and "pattern" not in data and "keyword" in data:
+            return {**data, "pattern": data["keyword"]}
+        return data
+
+    @model_validator(mode="after")
+    def _valid_pattern(self) -> Self:
+        if not self.pattern.strip():
+            msg = "pattern is blank"
+            raise ValueError(msg)
+        if self.regex:
+            compile_pattern(self.pattern)
+        return self
+
+
+class NewMessageParams(BaseModel):
+    chat_id: int | None = None
+    sender_id: int | None = None
+
+    @model_validator(mode="after")
+    def _scoped(self) -> Self:
+        if self.chat_id is None and self.sender_id is None:
+            msg = "set chat_id, sender_id or both"
+            raise ValueError(msg)
+        return self
+
+
+class DeletedParams(BaseModel):
+    chat_id: int | None = None
+
+
+class EditedParams(BaseModel):
+    chat_id: int | None = None
+
+
+class PeerOnlineParams(BaseModel):
+    peer_id: int
+
+
+WatchParams = (
+    KeywordParams | NewMessageParams | DeletedParams | EditedParams | PeerOnlineParams
+)
+
+PARAMS_BY_KIND: dict[str, type[WatchParams]] = {
+    "keyword": KeywordParams,
+    "new_message": NewMessageParams,
+    "message_deleted": DeletedParams,
+    "message_edited": EditedParams,
+    "peer_online": PeerOnlineParams,
+}
+
+
+def _describe(exc: ValidationError) -> str:
+    parts = []
+    for error in exc.errors():
+        where = ".".join(str(part) for part in error["loc"])
+        message = error["msg"].removeprefix("Value error, ")
+        parts.append(f"{where}: {message}" if where else message)
+    return "; ".join(parts)
+
+
+def parse_params(kind: str, params: dict[str, Any]) -> WatchParams:
+    model = PARAMS_BY_KIND.get(kind)
+    if model is None:
+        msg = f"unknown kind {kind!r}, expected one of {', '.join(WATCH_KINDS)}"
+        raise WatchParamsError(msg)
+    try:
+        return model.model_validate(params)
+    except ValidationError as exc:
+        raise WatchParamsError(_describe(exc)) from exc
+
+
+def normalize_params(kind: str, params: dict[str, Any]) -> dict[str, Any]:
+    return parse_params(kind, params).model_dump(exclude_none=True)
+
+
+def clean_name(name: str | None) -> str | None:
+    if name is None:
+        return None
+    name = name.strip()[:MAX_NAME]
+    return name or None