iris_x_hermes/gateway-plugin/purge.py

"""Complete (hard) deletion of iris messages from the hermes session store.

The iris plugin mints its own message ids (``m_<hex>``) that are **not**
persisted in the hermes session DB (``state.db``), so a delete request cannot
join on an id. Instead a message is matched to its ``messages`` row by
(session, role, content, timestamp proximity) and that row is deleted.

Deleting a ``messages`` row also drops it from the FTS5 search index via the
``messages_fts_*_delete`` triggers, so **no search trace survives** and the
message is not recoverable.

Channel / thread deletion wipes the whole lane: every session (and its
messages) for the chat (a channel) or the specific thread.

The session DB is opened read-write with a busy timeout. It is normally held
by the gateway core in WAL mode, which allows one writer at a time, so a brief
write from a second connection is safe.
"""

import contextlib
import logging
import sqlite3
from pathlib import Path

logger = logging.getLogger(__name__)


def _connect(db_path: Path) -> sqlite3.Connection:
    conn = sqlite3.connect(str(db_path), timeout=10.0)
    conn.row_factory = sqlite3.Row
    return conn


def _flat_session_id(conn: sqlite3.Connection, chat_id: str) -> str | None:
    """The flat-lane session (thread_id NULL / empty) for *chat_id*."""
    row = conn.execute(
        "SELECT id FROM sessions WHERE chat_id = ? "
        "AND (thread_id IS NULL OR thread_id = '') LIMIT 1",
        (chat_id,),
    ).fetchone()
    return row["id"] if row else None


def _thread_session_id(conn: sqlite3.Connection, chat_id: str, thread_id: str) -> str | None:
    row = conn.execute(
        "SELECT id FROM sessions WHERE chat_id = ? AND thread_id = ? LIMIT 1",
        (chat_id, thread_id),
    ).fetchone()
    return row["id"] if row else None


def delete_lane(db_path: Path, chat_id: str, thread_id: str | None = None) -> int:
    """Hard-delete a whole lane from the session store.

    * ``thread_id is None`` -> a **channel**: every session under *chat_id*
      (the flat lane plus all of its threads) and all of their messages.
    * ``thread_id`` set -> a **thread**: the single session for
      (chat_id, thread_id) and its messages.

    Returns the number of message rows removed (0 when the lane is absent or
    the DB is missing). Never raises: any DB error yields 0 (the caller still
    deletes the directory entry + outbox frames, so the lane is gone from the
    app's point of view even if the session-store wipe fails).
    """
    db_path = Path(db_path)
    if not db_path.exists():
        return 0
    conn = _connect(db_path)
    try:
        if thread_id:
            rows = conn.execute(
                "SELECT id FROM sessions WHERE chat_id = ? AND thread_id = ?",
                (chat_id, thread_id),
            ).fetchall()
        else:
            rows = conn.execute("SELECT id FROM sessions WHERE chat_id = ?", (chat_id,)).fetchall()
        ids = [r["id"] for r in rows]
        if not ids:
            return 0
        # Delete per session with fully parameterized queries (a channel has
        # only a handful of sessions: the flat lane plus its threads). Messages
        # first (foreign_keys is not enforced on this DB), then the session
        # rows. The FTS delete triggers fire on the message deletes.
        n_msgs = 0
        for sid in ids:
            cur = conn.execute("DELETE FROM messages WHERE session_id = ?", (sid,))
            n_msgs += max(0, cur.rowcount)
            conn.execute("DELETE FROM sessions WHERE id = ?", (sid,))
        conn.commit()
        return n_msgs
    except sqlite3.Error as e:
        logger.warning("iris purge: delete_lane failed: %s", e)
        return 0
    finally:
        with contextlib.suppress(Exception):
            conn.close()


def delete_message(
    db_path: Path,
    chat_id: str,
    thread_id: str | None,
    role: str,
    content: str,
    ts_ms: int | None,
) -> int:
    """Hard-delete a single message from the session store.

    Matches the ``messages`` row by (session, role, content) and, when several
    rows share that content, the one whose timestamp is closest to *ts_ms*.
    The content must match exactly (after stripping) -- a mismatch deletes
    nothing rather than the wrong message. Returns 1 when a row was removed,
    0 otherwise. Never raises.
    """
    db_path = Path(db_path)
    if not db_path.exists() or not content or not role:
        return 0
    conn = _connect(db_path)
    try:
        sid = (
            _thread_session_id(conn, chat_id, thread_id)
            if thread_id
            else _flat_session_id(conn, chat_id)
        )
        if sid is None:
            return 0
        ts_s = (ts_ms or 0) / 1000.0
        want = content.strip()
        rows = conn.execute(
            "SELECT id, timestamp, content FROM messages WHERE session_id = ? AND role = ?",
            (sid, role),
        ).fetchall()
        target: int | None = None
        best_dt: float | None = None
        for r in rows:
            if (r["content"] or "").strip() != want:
                continue
            try:
                dt = abs(float(r["timestamp"]) - ts_s)
            except (TypeError, ValueError):
                dt = 0.0
            if best_dt is None or dt < best_dt:
                best_dt = dt
                target = r["id"]
        if target is None:
            return 0
        conn.execute("DELETE FROM messages WHERE id = ?", (target,))
        conn.commit()
        return 1
    except sqlite3.Error as e:
        logger.warning("iris purge: delete_message failed: %s", e)
        return 0
    finally:
        with contextlib.suppress(Exception):
            conn.close()