repr-cli 0.2.16__py3-none-any.whl → 0.2.17__py3-none-any.whl

repr/timeline.py

@@ -0,0 +1,710 @@
+"""
+Timeline management for unified commit + session context.
+
+Handles:
+- Timeline storage (.repr/timeline.json)
+- Commit extraction from git
+- Session loading and context extraction
+- Commit-session matching and merging
+"""
+
+import json
+import os
+from dataclasses import asdict
+from datetime import datetime, timedelta, timezone
+from pathlib import Path
+from typing import Any, Callable
+
+from git import Repo
+from git.exc import InvalidGitRepositoryError
+
+from .models import (
+    CommitData,
+    CommitSessionMatch,
+    ReprTimeline,
+    SessionContext,
+    TimelineEntry,
+    TimelineEntryType,
+    match_commits_to_sessions,
+)
+
+
+# =============================================================================
+# Constants
+# =============================================================================
+
+REPR_DIR = ".repr"
+TIMELINE_FILE = "timeline.json"
+
+
+# =============================================================================
+# Helpers
+# =============================================================================
+
+def get_repr_dir(project_path: Path) -> Path:
+    """Get the .repr directory for a project."""
+    return project_path / REPR_DIR
+
+
+def get_timeline_path(project_path: Path) -> Path:
+    """Get the timeline.json path for a project."""
+    return get_repr_dir(project_path) / TIMELINE_FILE
+
+
+def ensure_repr_dir(project_path: Path) -> Path:
+    """Ensure .repr directory exists and return its path."""
+    repr_dir = get_repr_dir(project_path)
+    repr_dir.mkdir(parents=True, exist_ok=True)
+    
+    # Add .gitignore to .repr if it doesn't exist
+    gitignore_path = repr_dir / ".gitignore"
+    if not gitignore_path.exists():
+        gitignore_path.write_text("# Ignore everything in .repr\n*\n!.gitignore\n")
+    
+    return repr_dir
+
+
+def is_initialized(project_path: Path) -> bool:
+    """Check if a project has been initialized with repr timeline."""
+    return get_timeline_path(project_path).exists()
+
+
+def detect_project_root(start_path: Path | None = None) -> Path | None:
+    """
+    Detect the git repository root from a starting path.
+    
+    Args:
+        start_path: Starting path (defaults to cwd)
+    
+    Returns:
+        Repository root path or None if not in a git repo
+    """
+    start_path = start_path or Path.cwd()
+    
+    try:
+        repo = Repo(start_path, search_parent_directories=True)
+        return Path(repo.working_tree_dir)
+    except InvalidGitRepositoryError:
+        return None
+
+
+# =============================================================================
+# Timeline Serialization
+# =============================================================================
+
+def _datetime_to_iso(dt: datetime) -> str:
+    """Convert datetime to ISO string with timezone."""
+    if dt.tzinfo is None:
+        dt = dt.replace(tzinfo=timezone.utc)
+    return dt.isoformat()
+
+
+def _serialize_commit(commit: CommitData | None) -> dict | None:
+    """Serialize CommitData to JSON-safe dict."""
+    if commit is None:
+        return None
+    return {
+        "sha": commit.sha,
+        "message": commit.message,
+        "author": commit.author,
+        "timestamp": _datetime_to_iso(commit.timestamp),
+        "files": commit.files,
+        "insertions": commit.insertions,
+        "deletions": commit.deletions,
+    }
+
+
+def _serialize_session_context(ctx: SessionContext | None) -> dict | None:
+    """Serialize SessionContext to JSON-safe dict."""
+    if ctx is None:
+        return None
+    return {
+        "session_id": ctx.session_id,
+        "timestamp": _datetime_to_iso(ctx.timestamp),
+        "problem": ctx.problem,
+        "approach": ctx.approach,
+        "decisions": ctx.decisions,
+        "files_modified": ctx.files_modified,
+        "tools_used": ctx.tools_used,
+        "outcome": ctx.outcome,
+        "lessons": ctx.lessons,
+        "linked_commits": ctx.linked_commits,
+    }
+
+
+def _serialize_entry(entry: TimelineEntry) -> dict:
+    """Serialize TimelineEntry to JSON-safe dict."""
+    return {
+        "timestamp": _datetime_to_iso(entry.timestamp),
+        "type": entry.type.value,
+        "commit": _serialize_commit(entry.commit),
+        "session_context": _serialize_session_context(entry.session_context),
+        "story": entry.story,
+    }
+
+
+def _deserialize_commit(data: dict | None) -> CommitData | None:
+    """Deserialize CommitData from JSON dict."""
+    if data is None:
+        return None
+    return CommitData(
+        sha=data["sha"],
+        message=data["message"],
+        author=data["author"],
+        timestamp=datetime.fromisoformat(data["timestamp"]),
+        files=data["files"],
+        insertions=data.get("insertions", 0),
+        deletions=data.get("deletions", 0),
+        author_email=data.get("author_email", ""),
+    )
+
+
+def _deserialize_session_context(data: dict | None) -> SessionContext | None:
+    """Deserialize SessionContext from JSON dict."""
+    if data is None:
+        return None
+    return SessionContext(
+        session_id=data["session_id"],
+        timestamp=datetime.fromisoformat(data["timestamp"]),
+        problem=data["problem"],
+        approach=data["approach"],
+        decisions=data.get("decisions", []),
+        files_modified=data.get("files_modified", []),
+        tools_used=data.get("tools_used", []),
+        outcome=data["outcome"],
+        lessons=data.get("lessons", []),
+        linked_commits=data.get("linked_commits", []),
+    )
+
+
+def _deserialize_entry(data: dict) -> TimelineEntry:
+    """Deserialize TimelineEntry from JSON dict."""
+    return TimelineEntry(
+        timestamp=datetime.fromisoformat(data["timestamp"]),
+        type=TimelineEntryType(data["type"]),
+        commit=_deserialize_commit(data.get("commit")),
+        session_context=_deserialize_session_context(data.get("session_context")),
+        story=data.get("story"),
+    )
+
+
+def serialize_timeline(timeline: ReprTimeline) -> dict:
+    """Serialize ReprTimeline to JSON-safe dict."""
+    return {
+        "project_path": timeline.project_path,
+        "initialized_at": _datetime_to_iso(timeline.initialized_at),
+        "entries": [_serialize_entry(e) for e in timeline.entries],
+        "session_sources": timeline.session_sources,
+        "last_updated": _datetime_to_iso(timeline.last_updated) if timeline.last_updated else None,
+    }
+
+
+def deserialize_timeline(data: dict) -> ReprTimeline:
+    """Deserialize ReprTimeline from JSON dict."""
+    timeline = ReprTimeline(
+        project_path=data["project_path"],
+        initialized_at=datetime.fromisoformat(data["initialized_at"]),
+        entries=[_deserialize_entry(e) for e in data.get("entries", [])],
+        session_sources=data.get("session_sources", []),
+        last_updated=datetime.fromisoformat(data["last_updated"]) if data.get("last_updated") else None,
+    )
+    return timeline
+
+
+# =============================================================================
+# Timeline Storage
+# =============================================================================
+
+def save_timeline(timeline: ReprTimeline, project_path: Path) -> Path:
+    """
+    Save timeline to .repr/timeline.json.
+    
+    Args:
+        timeline: Timeline to save
+        project_path: Project root path
+    
+    Returns:
+        Path to the saved timeline file
+    """
+    ensure_repr_dir(project_path)
+    timeline_path = get_timeline_path(project_path)
+    
+    timeline.last_updated = datetime.now(timezone.utc)
+    
+    with open(timeline_path, "w") as f:
+        json.dump(serialize_timeline(timeline), f, indent=2)
+    
+    return timeline_path
+
+
+def load_timeline(project_path: Path) -> ReprTimeline | None:
+    """
+    Load timeline from .repr/timeline.json.
+    
+    Args:
+        project_path: Project root path
+    
+    Returns:
+        ReprTimeline or None if not found
+    """
+    timeline_path = get_timeline_path(project_path)
+    
+    if not timeline_path.exists():
+        return None
+    
+    with open(timeline_path) as f:
+        data = json.load(f)
+    
+    return deserialize_timeline(data)
+
+
+# =============================================================================
+# Commit Extraction
+# =============================================================================
+
+def extract_commits_from_git(
+    project_path: Path,
+    days: int = 90,
+    max_commits: int = 500,
+    since: datetime | None = None,
+    progress_callback: Callable[[int, int], None] | None = None,
+) -> list[CommitData]:
+    """
+    Extract commits from git history.
+    
+    Args:
+        project_path: Path to git repository
+        days: Number of days to look back
+        max_commits: Maximum number of commits to extract
+        since: Only get commits after this time
+        progress_callback: Optional callback(current, total) for progress
+    
+    Returns:
+        List of CommitData objects sorted by timestamp
+    """
+    repo = Repo(project_path)
+    commits = []
+    
+    # Calculate cutoff
+    if since:
+        cutoff = since
+    else:
+        cutoff = datetime.now(timezone.utc) - timedelta(days=days)
+    
+    cutoff_timestamp = cutoff.timestamp()
+    
+    # Collect commits
+    commit_count = 0
+    for commit in repo.iter_commits(max_count=max_commits):
+        if commit.committed_date < cutoff_timestamp:
+            break
+        
+        # Skip merge commits
+        if len(commit.parents) > 1:
+            continue
+        
+        # Get files changed
+        files = list(commit.stats.files.keys())
+        
+        commits.append(CommitData(
+            sha=commit.hexsha,
+            message=commit.message.strip(),
+            author=commit.author.name,
+            timestamp=datetime.fromtimestamp(commit.committed_date, tz=timezone.utc),
+            files=files,
+            insertions=commit.stats.total.get("insertions", 0),
+            deletions=commit.stats.total.get("deletions", 0),
+            author_email=commit.author.email or "",
+        ))
+        
+        commit_count += 1
+        if progress_callback:
+            progress_callback(commit_count, max_commits)
+    
+    # Sort by timestamp (oldest first)
+    commits.sort(key=lambda c: c.timestamp)
+    
+    return commits
+
+
+# =============================================================================
+# Timeline Initialization
+# =============================================================================
+
+def init_timeline_commits_only(
+    project_path: Path,
+    days: int = 90,
+    max_commits: int = 500,
+    progress_callback: Callable[[str, int, int], None] | None = None,
+) -> ReprTimeline:
+    """
+    Initialize timeline with commits only (no session extraction).
+    
+    Args:
+        project_path: Path to git repository
+        days: Number of days to look back
+        max_commits: Maximum number of commits
+        progress_callback: Optional callback(stage, current, total)
+    
+    Returns:
+        Initialized ReprTimeline
+    """
+    project_path = Path(project_path).resolve()
+    
+    # Create timeline
+    timeline = ReprTimeline(
+        project_path=str(project_path),
+        initialized_at=datetime.now(timezone.utc),
+        entries=[],
+        session_sources=[],
+    )
+    
+    # Extract commits
+    def commit_progress(current: int, total: int) -> None:
+        if progress_callback:
+            progress_callback("commits", current, total)
+    
+    commits = extract_commits_from_git(
+        project_path,
+        days=days,
+        max_commits=max_commits,
+        progress_callback=commit_progress,
+    )
+    
+    # Create timeline entries from commits
+    for commit in commits:
+        entry = TimelineEntry(
+            timestamp=commit.timestamp,
+            type=TimelineEntryType.COMMIT,
+            commit=commit,
+            session_context=None,
+            story=None,
+        )
+        timeline.entries.append(entry)
+    
+    # Save timeline
+    save_timeline(timeline, project_path)
+    
+    return timeline
+
+
+async def init_timeline_with_sessions(
+    project_path: Path,
+    days: int = 90,
+    max_commits: int = 500,
+    session_sources: list[str] | None = None,
+    api_key: str | None = None,
+    base_url: str | None = None,
+    model: str = "openai/gpt-4.1-mini",
+    extraction_concurrency: int = 3,
+    progress_callback: Callable[[str, int, int], None] | None = None,
+) -> ReprTimeline:
+    """
+    Initialize timeline with commits and session context extraction.
+
+    Args:
+        project_path: Path to git repository
+        days: Number of days to look back
+        max_commits: Maximum number of commits
+        session_sources: Sources to use (None = auto-detect)
+        api_key: API key for LLM extraction
+        base_url: Base URL for LLM API (e.g., for local Ollama)
+        model: Model for extraction
+        extraction_concurrency: Max concurrent LLM extractions
+        progress_callback: Optional callback(stage, current, total)
+
+    Returns:
+        Initialized ReprTimeline with session context
+    """
+    from .loaders import load_sessions_for_project, detect_session_source
+    from .session_extractor import SessionExtractor
+    
+    project_path = Path(project_path).resolve()
+    
+    # Create timeline
+    timeline = ReprTimeline(
+        project_path=str(project_path),
+        initialized_at=datetime.now(timezone.utc),
+        entries=[],
+        session_sources=[],
+    )
+    
+    # Extract commits
+    def commit_progress(current: int, total: int) -> None:
+        if progress_callback:
+            progress_callback("commits", current, total)
+    
+    commits = extract_commits_from_git(
+        project_path,
+        days=days,
+        max_commits=max_commits,
+        progress_callback=commit_progress,
+    )
+    
+    # Load sessions
+    if session_sources is None:
+        session_sources = detect_session_source(project_path)
+    
+    timeline.session_sources = session_sources
+    
+    sessions = load_sessions_for_project(
+        project_path,
+        sources=session_sources,
+        days_back=days,
+    )
+    
+    if progress_callback:
+        progress_callback("sessions_loaded", len(sessions), len(sessions))
+    
+    # Match commits to sessions
+    matches = match_commits_to_sessions(commits, sessions)
+    
+    # Build lookup of commit SHA -> matched sessions
+    commit_to_sessions: dict[str, list[str]] = {}
+    for match in matches:
+        if match.commit_sha not in commit_to_sessions:
+            commit_to_sessions[match.commit_sha] = []
+        commit_to_sessions[match.commit_sha].append(match.session_id)
+    
+    # Extract context from sessions
+    session_contexts: dict[str, SessionContext] = {}
+    
+    if sessions:
+        async with SessionExtractor(api_key=api_key, base_url=base_url, model=model) as extractor:
+            # Extract in batches with progress
+            total_sessions = len(sessions)
+            for i, session in enumerate(sessions):
+                if progress_callback:
+                    progress_callback("extracting", i + 1, total_sessions)
+                
+                try:
+                    # Find linked commits for this session
+                    linked_commits = [
+                        m.commit_sha for m in matches
+                        if m.session_id == session.id
+                    ]
+                    
+                    context = await extractor.extract_context(session, linked_commits=linked_commits)
+                    session_contexts[session.id] = context
+                except Exception as e:
+                    # Log but continue
+                    print(f"Warning: Failed to extract context from session {session.id}: {e}")
+    
+    # Create timeline entries
+    # First, add all commits
+    commit_entries: dict[str, TimelineEntry] = {}
+    for commit in commits:
+        entry = TimelineEntry(
+            timestamp=commit.timestamp,
+            type=TimelineEntryType.COMMIT,
+            commit=commit,
+            session_context=None,
+            story=None,
+        )
+        commit_entries[commit.sha] = entry
+        timeline.entries.append(entry)
+    
+    # Merge session context into matching commits
+    for match in matches:
+        if match.session_id in session_contexts and match.commit_sha in commit_entries:
+            entry = commit_entries[match.commit_sha]
+            ctx = session_contexts[match.session_id]
+            
+            # Upgrade to merged entry if we have session context
+            if entry.session_context is None:
+                entry.session_context = ctx
+                entry.type = TimelineEntryType.MERGED
+    
+    # Add standalone sessions (those not matched to any commit)
+    matched_session_ids = {m.session_id for m in matches}
+    for session_id, ctx in session_contexts.items():
+        if session_id not in matched_session_ids:
+            entry = TimelineEntry(
+                timestamp=ctx.timestamp,
+                type=TimelineEntryType.SESSION,
+                commit=None,
+                session_context=ctx,
+                story=None,
+            )
+            timeline.entries.append(entry)
+    
+    # Sort entries by timestamp
+    timeline.entries.sort(key=lambda e: e.timestamp)
+    
+    # Save timeline
+    save_timeline(timeline, project_path)
+    
+    return timeline
+
+
+def init_timeline_with_sessions_sync(
+    project_path: Path,
+    **kwargs,
+) -> ReprTimeline:
+    """Synchronous wrapper for init_timeline_with_sessions."""
+    import asyncio
+    return asyncio.run(init_timeline_with_sessions(project_path, **kwargs))
+
+
+# =============================================================================
+# Timeline Queries
+# =============================================================================
+
+def get_timeline_stats(timeline: ReprTimeline) -> dict[str, Any]:
+    """
+    Get statistics about a timeline.
+    
+    Returns dict with:
+        - total_entries
+        - commit_count
+        - session_count
+        - merged_count
+        - date_range (first, last)
+        - session_sources
+    """
+    stats = {
+        "total_entries": len(timeline.entries),
+        "commit_count": 0,
+        "session_count": 0,
+        "merged_count": 0,
+        "date_range": {"first": None, "last": None},
+        "session_sources": timeline.session_sources,
+    }
+    
+    for entry in timeline.entries:
+        if entry.type == TimelineEntryType.COMMIT:
+            stats["commit_count"] += 1
+        elif entry.type == TimelineEntryType.SESSION:
+            stats["session_count"] += 1
+        elif entry.type == TimelineEntryType.MERGED:
+            stats["merged_count"] += 1
+    
+    if timeline.entries:
+        stats["date_range"]["first"] = timeline.entries[0].timestamp.isoformat()
+        stats["date_range"]["last"] = timeline.entries[-1].timestamp.isoformat()
+    
+    return stats
+
+
+def query_timeline(
+    timeline: ReprTimeline,
+    since: datetime | None = None,
+    until: datetime | None = None,
+    entry_types: list[TimelineEntryType] | None = None,
+    files: list[str] | None = None,
+) -> list[TimelineEntry]:
+    """
+    Query timeline entries with filters.
+    
+    Args:
+        timeline: Timeline to query
+        since: Only entries after this time
+        until: Only entries before this time
+        entry_types: Filter by entry types
+        files: Filter by files touched
+    
+    Returns:
+        Filtered list of timeline entries
+    """
+    results = []
+    
+    for entry in timeline.entries:
+        # Time filters
+        if since and entry.timestamp < since:
+            continue
+        if until and entry.timestamp > until:
+            continue
+        
+        # Type filter
+        if entry_types and entry.type not in entry_types:
+            continue
+        
+        # File filter
+        if files:
+            entry_files = set()
+            if entry.commit:
+                entry_files.update(entry.commit.files)
+            if entry.session_context:
+                entry_files.update(entry.session_context.files_modified)
+            
+            if not any(f in entry_files for f in files):
+                continue
+        
+        results.append(entry)
+    
+    return results
+
+
+async def get_session_contexts_for_commits(
+    project_path: Path,
+    commits: list[CommitData],
+    days: int = 30,
+    session_sources: list[str] | None = None,
+    api_key: str | None = None,
+    base_url: str | None = None,
+    model: str = "openai/gpt-4.1-mini",
+    progress_callback: Callable[[str, int, int], None] | None = None,
+) -> list[SessionContext]:
+    """
+    Load and extract context from sessions linked to a list of commits.
+    
+    Args:
+        project_path: Path to the project
+        commits: List of commits to match against
+        days: How far back to look for sessions
+        session_sources: Sources to check (None = auto-detect)
+        api_key: API key for extraction
+        base_url: Base URL for extraction
+        model: Model to use for extraction
+        progress_callback: Optional progress callback
+        
+    Returns:
+        List of extracted SessionContext objects
+    """
+    from .loaders import load_sessions_for_project, detect_session_source
+    from .session_extractor import SessionExtractor
+    
+    # 1. Detect and load sessions
+    if session_sources is None:
+        session_sources = detect_session_source(project_path)
+    
+    if not session_sources:
+        return []
+        
+    sessions = load_sessions_for_project(
+        project_path,
+        sources=session_sources,
+        days_back=days,
+    )
+    
+    if not sessions:
+        return []
+        
+    if progress_callback:
+        progress_callback("sessions_loaded", len(sessions), len(sessions))
+        
+    # 2. Match commits to sessions
+    matches = match_commits_to_sessions(commits, sessions)
+    
+    if not matches:
+        return []
+        
+    # 3. Extract context from matched sessions
+    matched_session_ids = {m.session_id for m in matches}
+    matched_sessions = [s for s in sessions if s.id in matched_session_ids]
+    
+    session_contexts = []
+    async with SessionExtractor(api_key=api_key, base_url=base_url, model=model) as extractor:
+        total_to_extract = len(matched_sessions)
+        for i, session in enumerate(matched_sessions):
+            if progress_callback:
+                progress_callback("extracting", i + 1, total_to_extract)
+                
+            try:
+                linked_commits = [m.commit_sha for m in matches if m.session_id == session.id]
+                context = await extractor.extract_context(session, linked_commits=linked_commits)
+                session_contexts.append(context)
+            except Exception:
+                continue
+            
+    return session_contexts