cc-transcript 14.14.0__tar.gz → 14.16.0__tar.gz

{cc_transcript-14.14.0 → cc_transcript-14.16.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: cc-transcript
-Version: 14.14.0
+Version: 14.16.0
 Classifier: Development Status :: 3 - Alpha
 Classifier: Environment :: Console
 Classifier: Intended Audience :: Developers

{cc_transcript-14.14.0 → cc_transcript-14.16.0}/cc_transcript/decision_store.py

@@ -1,13 +1,23 @@
 from __future__ import annotations
 
+import errno
+import fcntl
 import hashlib
+import logging
 import sqlite3
+import time
+from collections.abc import Iterator
+from contextlib import contextmanager
 from pathlib import Path
 
+logger = logging.getLogger(__name__)
+
 SCHEMA_COMPONENT = "cc-review-decisions-v1"
 SCHEMA_VERSION = 1
 EXPECTED_DDL_FINGERPRINT = "6ae938038f3420cdd4a00189b678fb399d60bb7647d009acb0fa9cc4a653040f"
 EXPECTED_OBJECT_FINGERPRINT = "a993521f1ae402d85545d9cd841b58c7e9ba755babba32c7d59cc3a97ee17af9"
+_ARCHIVE_LOCK_TIMEOUT_SECONDS = 2.0
+_ARCHIVE_LOCK_RETRY_SECONDS = 0.01
 
 _SCHEMA_DDL_ACTIONS = frozenset(
     {
@@ -146,6 +156,66 @@ def _authorize_exact_schema(
     return sqlite3.SQLITE_OK
 
 
+@contextmanager
+def _archive_lock(db_path: Path) -> Iterator[None]:
+    lock_path = db_path.with_name(f"{db_path.name}.archive.lock")
+    with lock_path.open("a+b") as lock:
+        deadline = time.monotonic() + _ARCHIVE_LOCK_TIMEOUT_SECONDS
+        while True:
+            try:
+                fcntl.flock(lock, fcntl.LOCK_EX | fcntl.LOCK_NB)
+                break
+            except OSError as error:
+                if error.errno not in (errno.EACCES, errno.EAGAIN):
+                    raise
+                remaining = deadline - time.monotonic()
+                if remaining <= 0:
+                    raise TimeoutError(
+                        f"timed out acquiring decisions archive lock {lock_path} "
+                        f"after {_ARCHIVE_LOCK_TIMEOUT_SECONDS:g} seconds"
+                    ) from error
+                time.sleep(min(_ARCHIVE_LOCK_RETRY_SECONDS, remaining))
+        try:
+            yield
+        finally:
+            fcntl.flock(lock, fcntl.LOCK_UN)
+
+
+def _archive_incompatible_database(db_path: Path, found_version: int) -> None:
+    checkpoint = sqlite3.connect(db_path, autocommit=True)
+    try:
+        try:
+            checkpoint.execute("PRAGMA wal_checkpoint(TRUNCATE)")
+        except sqlite3.Error:
+            pass
+    finally:
+        checkpoint.close()
+
+    timestamp_ns = time.time_ns()
+    seconds = timestamp_ns // 1_000_000_000
+    microseconds = timestamp_ns // 1_000 % 1_000_000
+    archive_stem = (
+        f"{db_path.name}.v{found_version}."
+        f"{time.strftime('%Y%m%d-%H%M%S', time.localtime(seconds))}-{microseconds:06d}"
+    )
+    archive_path = db_path.with_name(f"{archive_stem}.bak")
+    collision = 0
+    while archive_path.exists():
+        collision += 1
+        archive_path = db_path.with_name(f"{archive_stem}.{collision}.bak")
+    db_path.rename(archive_path)
+    for suffix in ("-wal", "-shm"):
+        sidecar = db_path.with_name(f"{db_path.name}{suffix}")
+        sidecar.unlink(missing_ok=True)
+    logger.warning(
+        "Archived incompatible decisions database %s with schema version %s (expected %s) to %s",
+        db_path,
+        found_version,
+        SCHEMA_VERSION,
+        archive_path,
+    )
+
+
 def open_decisions_sqlite(path: Path | None) -> sqlite3.Connection:
     if ddl_fingerprint() != EXPECTED_DDL_FINGERPRINT:
         raise RuntimeError(
@@ -153,32 +223,40 @@ def open_decisions_sqlite(path: Path | None) -> sqlite3.Connection:
         )
     db_path = path or Path.home() / ".cc-transcript" / "decisions.db"
     db_path.parent.mkdir(parents=True, exist_ok=True)
-    conn = sqlite3.connect(db_path, autocommit=True)
-    conn.row_factory = sqlite3.Row
-    conn.execute("PRAGMA busy_timeout = 2000")
-    committed = False
     created = False
-    try:
-        conn.execute("BEGIN IMMEDIATE")
-        version = conn.execute("PRAGMA user_version").fetchone()[0]
-        app_objects = conn.execute(
-            "SELECT count(*) FROM sqlite_schema "
-            "WHERE type IN ('table', 'index', 'trigger', 'view') "
-            "AND lower(substr(name, 1, 7)) <> 'sqlite_'"
-        ).fetchone()[0]
-        created = version == 0 and app_objects == 0
-        create_schema(conn) if created else verify_schema(conn)
-        conn.execute("COMMIT")
-        committed = True
-    finally:
-        if not committed:
-            if conn.in_transaction:
-                conn.execute("ROLLBACK")
+    with _archive_lock(db_path):
+        while True:
+            conn = sqlite3.connect(db_path, autocommit=True)
+            committed = False
+            try:
+                conn.row_factory = sqlite3.Row
+                conn.execute("PRAGMA busy_timeout = 2000")
+                conn.execute("BEGIN IMMEDIATE")
+                version = conn.execute("PRAGMA user_version").fetchone()[0]
+                has_objects = conn.execute("SELECT EXISTS(SELECT 1 FROM sqlite_master)").fetchone()[0]
+                created = version == 0 and not has_objects
+                if version != SCHEMA_VERSION and not created:
+                    conn.execute("ROLLBACK")
+                else:
+                    create_schema(conn) if created else verify_schema(conn)
+                    conn.execute("COMMIT")
+                    committed = True
+            except BaseException:
+                if conn.in_transaction:
+                    conn.execute("ROLLBACK")
+                conn.close()
+                raise
+            if committed:
+                break
+            conn.close()
+            _archive_incompatible_database(db_path, version)
+        try:
+            if created:
+                db_path.chmod(0o600)
+            if (journal_mode := conn.execute("PRAGMA journal_mode = WAL").fetchone()[0]) != "wal":
+                raise RuntimeError(f"enable decisions WAL: mode {journal_mode!r}")
+            conn.set_authorizer(_authorize_exact_schema)
+        except BaseException:
             conn.close()
-    if created:
-        db_path.chmod(0o600)
-    if (journal_mode := conn.execute("PRAGMA journal_mode = WAL").fetchone()[0]) != "wal":
-        conn.close()
-        raise RuntimeError(f"enable decisions WAL: mode {journal_mode!r}")
-    conn.set_authorizer(_authorize_exact_schema)
+            raise
     return conn

{cc_transcript-14.14.0 → cc_transcript-14.16.0}/cc_transcript/filterspec.py

@@ -73,7 +73,8 @@ STRUCTURAL_GROUPS: tuple[tuple[str, str], ...] = (
 # Start-anchored so a mid-text mention is not a banner.
 # Dialect-hardened for Rust-regex parity — see cc-notes doc "rust-regex-dialect-parity".
 AGENT_INJECTION_GROUPS: tuple[tuple[str, str], ...] = (
-    ("xml_tags_extra", r"\A\s*<(?:teammate-message|scheduled-task)(?:[\s/>]|$)"),
+    ("xml_tags_extra", r"\A\s*<(?:teammate-message|scheduled-task|agent-message|cross-session-message)(?:[\s/>]|$)"),
+    ("relay_preamble", r"\A\s*Another Claude sess(?-i:[Ii])on sent a message:"),
     ("augment_agent", r"\A\s*# Augment Agent(?:\s|$)"),
     ("role_reminder", r"\A\s*\[Role Rem(?-i:[Ii])nder(?:[\s:\]]|$)"),
 )

{cc_transcript-14.14.0 → cc_transcript-14.16.0}/cc_transcript/query.py

@@ -11,6 +11,7 @@ predicates compose over progressively narrower windows.
 from __future__ import annotations
 
 import re
+import threading
 from collections.abc import Mapping
 from dataclasses import dataclass
 from fnmatch import fnmatch
@@ -46,6 +47,31 @@ if TYPE_CHECKING:
     from cc_transcript.models import TranscriptEvent
 
 
+DEEP_LIFT_BUDGET = 64 * 1024 * 1024
+
+DEEP_LIFTS: dict[LiftKey, tuple[LiftStamp, DeepSession]] = {}
+"""Lifted sidechain transcripts, keyed by tree position and stamped with the file's ``(size, mtime_ns)``.
+
+:meth:`Session.walk` re-reads every node of a tree in the same order, so evicting the
+least-recently-used entry thrashes on any tree larger than the cache — each walk drops
+exactly what the next one asks for first. Admission stops at :data:`DEEP_LIFT_BUDGET`
+source bytes instead: what is held stays held, and a tree that outgrows the budget
+reparses only its tail. A sidechain that grew replaces its own entry, budget or not.
+"""
+
+DEEP_LIFT_GUARD = threading.Lock()
+
+type LiftStamp = tuple[int, int]
+type LiftKey = tuple[Path, int, ToolUseId | None]
+
+
+def hold_lift(key: LiftKey, stamp: LiftStamp, deep: DeepSession) -> DeepSession:
+    with DEEP_LIFT_GUARD:
+        if key in DEEP_LIFTS or sum(size for (size, _), _ in DEEP_LIFTS.values()) < DEEP_LIFT_BUDGET:
+            DEEP_LIFTS[key] = (stamp, deep)
+    return deep
+
+
 def is_failure(use: ToolUse) -> bool:
     return use.result is not None and use.result.is_error
 
@@ -236,13 +262,18 @@ class ToolCallQuery:
         return bool(self.items)
 
 
-@dataclass(frozen=True, slots=True)
+@dataclass(frozen=True)
 class Session:
     """An immutable windowed view of a session's turns.
 
     Every slicing operation returns another :class:`Session`; turns at a
     window boundary are trimmed copies, so mid-turn slices stay event-precise.
 
+    Unslotted, unlike its neighbours: the window is immutable, so every
+    derivation over it — :attr:`tool_calls`, :meth:`commands`,
+    :meth:`command_lines` — memoizes per instance, and a predicate that
+    queries one window repeatedly pays each derivation once.
+
     Attributes:
         turns: The turns in the window.
         path: The transcript file the session was loaded from, when known —
@@ -301,7 +332,7 @@ class Session:
         """Every event in the window, in order."""
         return tuple(event for turn in self.turns for event in turn.events)
 
-    @property
+    @cached_property
     def tool_calls(self) -> ToolCallQuery:
         """The window's tool calls as a chainable query."""
         return ToolCallQuery(tuple(use for turn in self.turns for use in turn.tool_uses))
@@ -344,6 +375,10 @@ class Session:
         cycles terminate. An unreadable transcript is skipped but its children
         are still walked; a structurally malformed line raises, as
         :attr:`subagents` does.
+
+        Each reached transcript is parsed and lifted once per ``(size, mtime)``
+        stamp and memoized across walks, so a predicate that walks repeatedly —
+        or a resident process that re-walks per event — reparses only what grew.
         """
         return deep_sessions(self)
 
@@ -489,13 +524,21 @@ class Session:
 
     def commands(self) -> tuple[str, ...]:
         """The shell command strings of the window's Bash calls."""
-        return tuple(call.command for use in self.tool_calls.named("Bash") if isinstance(call := use.call, BashCall))
+        return self._commands
 
     def command_lines(self) -> tuple[CommandLine, ...]:
         """The window's Bash commands parsed into :class:`~cc_transcript.command.CommandLine` objects."""
+        return self._command_lines
+
+    @cached_property
+    def _commands(self) -> tuple[str, ...]:
+        return tuple(call.command for use in self.tool_calls.named("Bash") if isinstance(call := use.call, BashCall))
+
+    @cached_property
+    def _command_lines(self) -> tuple[CommandLine, ...]:
         from cc_transcript.command import parse_command_line
 
-        return tuple(parse_command_line(command) for command in self.commands())
+        return tuple(parse_command_line(command) for command in self._commands)
 
     def __len__(self) -> int:
         return sum(len(turn.events) for turn in self.turns)
@@ -659,11 +702,8 @@ def visit_transcript(path: Path, depth: int, spawned_by: ToolUseId | None, seen:
     yield from descend_sidechains(path, depth + 1, seen)
 
 
-def load_deep_session(path: Path, depth: int, spawned_by: ToolUseId | None) -> DeepSession | None:
-    try:
-        transcript = parse(path)
-    except OSError:
-        return None
+def lift_deep_session(path: Path, depth: int, spawned_by: ToolUseId | None) -> DeepSession:
+    transcript = parse(path)
     session_id = session_id_of(transcript.events) or SessionId(path.stem)
     return DeepSession(
         session=Session.from_activity(SessionActivity.from_events(session_id, transcript.events), path=path),
@@ -672,3 +712,18 @@ def load_deep_session(path: Path, depth: int, spawned_by: ToolUseId | None) -> D
         depth=depth,
         spawned_by=spawned_by,
     )
+
+
+def load_deep_session(path: Path, depth: int, spawned_by: ToolUseId | None) -> DeepSession | None:
+    try:
+        stat = path.stat()
+    except OSError:
+        return None
+    stamp = (stat.st_size, stat.st_mtime_ns)
+    key = (path, depth, spawned_by)
+    if (held := DEEP_LIFTS.get(key)) is not None and held[0] == stamp:
+        return held[1]
+    try:
+        return hold_lift(key, stamp, lift_deep_session(path, depth, spawned_by))
+    except OSError:
+        return None

{cc_transcript-14.14.0 → cc_transcript-14.16.0}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "cc-transcript"
-version = "14.14.0" # informational — releases stamp the version from the git tag
+version = "14.16.0" # informational — releases stamp the version from the git tag
 description = "Grep every Claude Code session you've ever run."
 readme = "README.md"
 license = "PolyForm-Noncommercial-1.0.0"

{cc_transcript-14.14.0 → cc_transcript-14.16.0}/rust/crates/core/src/filter.rs

@@ -11,7 +11,7 @@ use crate::value::{field, field_bool, field_str};
 /// bench reuses: `--signal` keeps conversational turns minus junk, `--no-junk`
 /// (`builders.NOISE_SPEC` in Python) drops structural noise. The one source of
 /// truth; `tests/test_show_specs.py` pins Python's `NOISE_SPEC` to `NOISE_SPEC_JSON`.
-pub const SIGNAL_SPEC_JSON: &str = r#"{"clauses":[{"predicate":{"kind":"KindIs","kinds":["assistant","user"]},"action":"drop","applies_to":[],"negate":true,"label":null},{"predicate":{"kind":"TextMatchesAny","groups":[["xml_tags","<(?:system[_-](?:instruction|reminder)|local-command-(?:stdout|stderr|caveat)|command-(?:name|message|args)|task-notification|persisted-output|output-file)\\b"],["local_command_caveat","Caveat: The messages below were generated by the user while running local commands\\."],["remaining_tasks_ack","REMAINING_TASKS_ACKNOWLEDGED"],["angle_token","<<[a-z][a-z0-9-]*>>"],["skill_base_dir","^Base directory for this skill:"],["hash_banner","(?:###\\s+[\\w][\\w \\-]{0,30}\\s+){3,}###"],["xml_tags_extra","\\A\\s*<(?:teammate-message|scheduled-task)(?:[\\s/>]|$)"],["augment_agent","\\A\\s*# Augment Agent(?:\\s|$)"],["role_reminder","\\A\\s*\\[Role Rem(?-i:[Ii])nder(?:[\\s:\\]]|$)"],["command_echo","\\A\\s*<bash-(?:input|stdout|stderr)\\b"]],"ignore_case":true},"action":"drop","applies_to":["user"],"negate":false,"label":null},{"predicate":{"kind":"ModelIs","models":["<synthetic>"]},"action":"drop","applies_to":["assistant"],"negate":false,"label":null},{"predicate":{"kind":"MetaFlag","flag":"is_sidechain"},"action":"drop","applies_to":[],"negate":false,"label":null},{"predicate":{"kind":"TextEmpty","consider_tool_use":false},"action":"drop","applies_to":["user"],"negate":false,"label":null},{"predicate":{"kind":"TextEmpty","consider_tool_use":true},"action":"drop","applies_to":["assistant"],"negate":false,"label":null}]}"#;
+pub const SIGNAL_SPEC_JSON: &str = r#"{"clauses":[{"predicate":{"kind":"KindIs","kinds":["assistant","user"]},"action":"drop","applies_to":[],"negate":true,"label":null},{"predicate":{"kind":"TextMatchesAny","groups":[["xml_tags","<(?:system[_-](?:instruction|reminder)|local-command-(?:stdout|stderr|caveat)|command-(?:name|message|args)|task-notification|persisted-output|output-file)\\b"],["local_command_caveat","Caveat: The messages below were generated by the user while running local commands\\."],["remaining_tasks_ack","REMAINING_TASKS_ACKNOWLEDGED"],["angle_token","<<[a-z][a-z0-9-]*>>"],["skill_base_dir","^Base directory for this skill:"],["hash_banner","(?:###\\s+[\\w][\\w \\-]{0,30}\\s+){3,}###"],["xml_tags_extra","\\A\\s*<(?:teammate-message|scheduled-task|agent-message|cross-session-message)(?:[\\s/>]|$)"],["relay_preamble","\\A\\s*Another Claude sess(?-i:[Ii])on sent a message:"],["augment_agent","\\A\\s*# Augment Agent(?:\\s|$)"],["role_reminder","\\A\\s*\\[Role Rem(?-i:[Ii])nder(?:[\\s:\\]]|$)"],["command_echo","\\A\\s*<bash-(?:input|stdout|stderr)\\b"]],"ignore_case":true},"action":"drop","applies_to":["user"],"negate":false,"label":null},{"predicate":{"kind":"ModelIs","models":["<synthetic>"]},"action":"drop","applies_to":["assistant"],"negate":false,"label":null},{"predicate":{"kind":"MetaFlag","flag":"is_sidechain"},"action":"drop","applies_to":[],"negate":false,"label":null},{"predicate":{"kind":"TextEmpty","consider_tool_use":false},"action":"drop","applies_to":["user"],"negate":false,"label":null},{"predicate":{"kind":"TextEmpty","consider_tool_use":true},"action":"drop","applies_to":["assistant"],"negate":false,"label":null}]}"#;
 pub const NOISE_SPEC_JSON: &str = r#"{"clauses":[{"predicate":{"kind":"TextMatchesAny","groups":[["xml_tags","<(?:system[_-](?:instruction|reminder)|local-command-(?:stdout|stderr|caveat)|command-(?:name|message|args)|task-notification|persisted-output|output-file)\\b"],["local_command_caveat","Caveat: The messages below were generated by the user while running local commands\\."],["remaining_tasks_ack","REMAINING_TASKS_ACKNOWLEDGED"],["angle_token","<<[a-z][a-z0-9-]*>>"],["skill_base_dir","^Base directory for this skill:"],["hash_banner","(?:###\\s+[\\w][\\w \\-]{0,30}\\s+){3,}###"]],"ignore_case":true},"action":"drop","applies_to":["user"],"negate":false,"label":null}]}"#;
 
 pub static SIGNAL_SPEC: Lazy<CompiledSpec> =

{cc_transcript-14.14.0 → cc_transcript-14.16.0}/rust/crates/core/src/literals/protocol.rs

@@ -10,7 +10,7 @@ pub const USER_SAID_TRAILER: &str = "Note: The user's next message";
 pub const ANSWERED_PREFIX: &str = "Your questions have been answered: ";
 pub const ANSWERED_TRAILER: &str = ". You can now continue with these answers in mind.";
 pub const INTERRUPT_MARKER_PATTERN: &str = "(?:^\\s*\\[Request (?-i:[Ii])nterrupted by user)";
-pub const AGENT_INJECTION_PATTERN: &str = "(?:\\A\\s*<(?:teammate-message|scheduled-task)(?:[\\s/>]|$))|(?:\\A\\s*# Augment Agent(?:\\s|$))|(?:\\A\\s*\\[Role Rem(?-i:[Ii])nder(?:[\\s:\\]]|$))";
+pub const AGENT_INJECTION_PATTERN: &str = "(?:\\A\\s*<(?:teammate-message|scheduled-task|agent-message|cross-session-message)(?:[\\s/>]|$))|(?:\\A\\s*Another Claude sess(?-i:[Ii])on sent a message:)|(?:\\A\\s*# Augment Agent(?:\\s|$))|(?:\\A\\s*\\[Role Rem(?-i:[Ii])nder(?:[\\s:\\]]|$))";
 pub const TASK_NOTIFICATION_MARKER: &str = "<task-notification>";
 pub const TOOL_USE_ID_PREFIX: &str = "<tool-use-id>";
 pub const TOOL_USE_ID_SUFFIX: &str = "</tool-use-id>";

{cc_transcript-14.14.0 → cc_transcript-14.16.0}/rust/crates/core/src/protocol.rs

@@ -167,6 +167,15 @@ mod tests {
         assert!(is_agent_injection(
             "<scheduled-task id='7'>run the suite</scheduled-task>"
         ));
+        assert!(is_agent_injection(
+            "Another Claude session sent a message: <teammate-message teammate_id=\"x\" color=\"orange\">\nhello"
+        ));
+        assert!(is_agent_injection(
+            "<agent-message from='x'>rebase onto main"
+        ));
+        assert!(is_agent_injection(
+            "<cross-session-message from='x'>the build is green"
+        ));
         assert!(is_agent_injection("[Role Reminder: You are a Coordinator."));
         assert!(is_agent_injection("# Augment Agent\nyou have these tools"));
         // Leading whitespace before the marker is tolerated.
@@ -190,6 +199,14 @@ mod tests {
         assert!(!is_agent_injection(
             "we discussed the [Role Reminder] banner mid-sentence"
         ));
+        assert!(!is_agent_injection(
+            "the log says Another Claude session sent a message: further down"
+        ));
+        // The "i" of "session" is ASCII-pinned: the dotless-I (U+0131) Python re folds must
+        // never match on either backend.
+        assert!(!is_agent_injection(
+            "Another Claude sess\u{131}on sent a message: dotless"
+        ));
         // A combining mark (U+0301) after the tag name is not a portable word boundary; and
         // dotless-I must not fold to ASCII "i" — Python re once matched both, Rust never did.
         assert!(!is_agent_injection("<teammate-message\u{0301}>"));