up-cli 0.1.1__py3-none-any.whl → 0.2.0__py3-none-any.whl

up/summarizer.py

@@ -0,0 +1,407 @@
+"""Conversation summarizer for Claude and Cursor chat history.
+
+Extracts patterns, learnings, and actionable insights from AI conversation history.
+"""
+
+import json
+import re
+from dataclasses import dataclass, field
+from datetime import datetime
+from pathlib import Path
+from typing import Optional
+from collections import Counter
+
+
+@dataclass
+class ConversationPattern:
+    """A pattern extracted from conversations."""
+    name: str
+    description: str
+    frequency: int
+    examples: list[str] = field(default_factory=list)
+    category: str = "general"
+
+
+@dataclass 
+class ConversationInsight:
+    """An insight or learning from conversations."""
+    title: str
+    description: str
+    source_count: int
+    confidence: str  # high, medium, low
+    actionable: bool = False
+    action: str = ""
+
+
+@dataclass
+class SummaryReport:
+    """Summary report from conversation analysis."""
+    total_conversations: int
+    total_messages: int
+    date_range: tuple[str, str]
+    top_topics: list[tuple[str, int]]
+    patterns: list[ConversationPattern]
+    insights: list[ConversationInsight]
+    code_snippets: list[dict]
+    errors_encountered: list[str]
+
+
+class ConversationSummarizer:
+    """Analyzes and summarizes AI conversation history."""
+    
+    # Common coding topics to detect
+    TOPIC_PATTERNS = {
+        "testing": r"\b(test|pytest|jest|unit test|integration test|coverage)\b",
+        "debugging": r"\b(debug|error|exception|traceback|fix|bug)\b",
+        "refactoring": r"\b(refactor|clean up|improve|optimize|restructure)\b",
+        "documentation": r"\b(document|readme|docstring|comment|explain)\b",
+        "api": r"\b(api|endpoint|route|http|rest|graphql)\b",
+        "database": r"\b(database|sql|query|migration|schema|model)\b",
+        "authentication": r"\b(auth|login|jwt|session|permission|role)\b",
+        "deployment": r"\b(deploy|docker|kubernetes|ci/cd|pipeline)\b",
+        "performance": r"\b(performance|optimize|cache|slow|fast|memory)\b",
+        "security": r"\b(security|vulnerability|sanitize|escape|inject)\b",
+    }
+    
+    # Error patterns
+    ERROR_PATTERNS = [
+        r"error:\s*(.+?)(?:\n|$)",
+        r"exception:\s*(.+?)(?:\n|$)",
+        r"failed:\s*(.+?)(?:\n|$)",
+        r"TypeError:\s*(.+?)(?:\n|$)",
+        r"ValueError:\s*(.+?)(?:\n|$)",
+        r"ImportError:\s*(.+?)(?:\n|$)",
+    ]
+    
+    def __init__(self, conversations: list[dict]):
+        """Initialize with conversation data.
+        
+        Args:
+            conversations: List of conversation dicts with messages
+        """
+        self.conversations = conversations
+        self.all_messages = []
+        self._extract_messages()
+    
+    def _extract_messages(self) -> None:
+        """Extract all messages from conversations."""
+        for conv in self.conversations:
+            for msg in conv.get("messages", []):
+                self.all_messages.append({
+                    "role": msg.get("role", "unknown"),
+                    "content": msg.get("content", ""),
+                    "timestamp": msg.get("timestamp"),
+                    "conversation_id": conv.get("id"),
+                    "project": conv.get("project"),
+                })
+    
+    def analyze(self) -> SummaryReport:
+        """Analyze conversations and generate summary report."""
+        # Basic stats
+        total_convs = len(self.conversations)
+        total_msgs = len(self.all_messages)
+        
+        # Date range
+        timestamps = [
+            m["timestamp"] for m in self.all_messages 
+            if m.get("timestamp")
+        ]
+        if timestamps:
+            date_range = (
+                self._format_timestamp(min(timestamps)),
+                self._format_timestamp(max(timestamps)),
+            )
+        else:
+            date_range = ("Unknown", "Unknown")
+        
+        # Analyze topics
+        top_topics = self._analyze_topics()
+        
+        # Extract patterns
+        patterns = self._extract_patterns()
+        
+        # Generate insights
+        insights = self._generate_insights(top_topics, patterns)
+        
+        # Extract code snippets
+        code_snippets = self._extract_code_snippets()
+        
+        # Extract errors
+        errors = self._extract_errors()
+        
+        return SummaryReport(
+            total_conversations=total_convs,
+            total_messages=total_msgs,
+            date_range=date_range,
+            top_topics=top_topics[:10],
+            patterns=patterns[:10],
+            insights=insights[:10],
+            code_snippets=code_snippets[:20],
+            errors_encountered=errors[:20],
+        )
+    
+    def _format_timestamp(self, ts: int) -> str:
+        """Format timestamp to readable string."""
+        try:
+            dt = datetime.fromtimestamp(ts / 1000)
+            return dt.strftime("%Y-%m-%d")
+        except (ValueError, TypeError):
+            return "Unknown"
+    
+    def _analyze_topics(self) -> list[tuple[str, int]]:
+        """Analyze topic frequency in conversations."""
+        topic_counts = Counter()
+        
+        for msg in self.all_messages:
+            content = msg.get("content", "").lower()
+            for topic, pattern in self.TOPIC_PATTERNS.items():
+                if re.search(pattern, content, re.IGNORECASE):
+                    topic_counts[topic] += 1
+        
+        return topic_counts.most_common()
+    
+    def _extract_patterns(self) -> list[ConversationPattern]:
+        """Extract common patterns from conversations."""
+        patterns = []
+        
+        # Pattern: Common user request types
+        request_patterns = {
+            "implementation": r"(implement|create|add|build)\s+(\w+)",
+            "fix": r"(fix|resolve|solve)\s+(\w+)",
+            "explain": r"(explain|what is|how does)\s+(\w+)",
+            "refactor": r"(refactor|improve|clean)\s+(\w+)",
+        }
+        
+        request_counts = Counter()
+        request_examples = {}
+        
+        for msg in self.all_messages:
+            if msg.get("role") != "user":
+                continue
+            content = msg.get("content", "")
+            
+            for pattern_name, pattern in request_patterns.items():
+                matches = re.findall(pattern, content, re.IGNORECASE)
+                if matches:
+                    request_counts[pattern_name] += len(matches)
+                    if pattern_name not in request_examples:
+                        request_examples[pattern_name] = []
+                    if len(request_examples[pattern_name]) < 3:
+                        request_examples[pattern_name].append(content[:100])
+        
+        for name, count in request_counts.most_common(10):
+            patterns.append(ConversationPattern(
+                name=f"Request: {name}",
+                description=f"User frequently requests {name} actions",
+                frequency=count,
+                examples=request_examples.get(name, []),
+                category="request_type",
+            ))
+        
+        return patterns
+    
+    def _generate_insights(
+        self, 
+        topics: list[tuple[str, int]], 
+        patterns: list[ConversationPattern]
+    ) -> list[ConversationInsight]:
+        """Generate actionable insights from analysis."""
+        insights = []
+        
+        # Insight from top topics
+        if topics:
+            top_topic = topics[0][0]
+            insights.append(ConversationInsight(
+                title=f"Primary Focus: {top_topic.title()}",
+                description=f"Most conversations involve {top_topic} ({topics[0][1]} mentions)",
+                source_count=topics[0][1],
+                confidence="high",
+                actionable=True,
+                action=f"Consider creating documentation or templates for {top_topic}",
+            ))
+        
+        # Insight from error patterns
+        error_count = sum(1 for m in self.all_messages if "error" in m.get("content", "").lower())
+        if error_count > 10:
+            insights.append(ConversationInsight(
+                title="Frequent Debugging Sessions",
+                description=f"{error_count} messages contain error-related content",
+                source_count=error_count,
+                confidence="high",
+                actionable=True,
+                action="Consider improving error handling or adding better logging",
+            ))
+        
+        # Insight from conversation length
+        avg_msgs = len(self.all_messages) / max(len(self.conversations), 1)
+        if avg_msgs > 20:
+            insights.append(ConversationInsight(
+                title="Long Conversation Sessions",
+                description=f"Average {avg_msgs:.0f} messages per conversation",
+                source_count=len(self.conversations),
+                confidence="medium",
+                actionable=True,
+                action="Consider breaking complex tasks into smaller sessions",
+            ))
+        
+        return insights
+    
+    def _extract_code_snippets(self) -> list[dict]:
+        """Extract code snippets from conversations."""
+        snippets = []
+        code_pattern = r"```(\w+)?\n(.*?)```"
+        
+        for msg in self.all_messages:
+            if msg.get("role") != "assistant":
+                continue
+            content = msg.get("content", "")
+            
+            matches = re.findall(code_pattern, content, re.DOTALL)
+            for lang, code in matches:
+                if len(code.strip()) > 20:  # Skip trivial snippets
+                    snippets.append({
+                        "language": lang or "unknown",
+                        "code": code.strip()[:500],  # Limit size
+                        "project": msg.get("project"),
+                    })
+        
+        return snippets
+    
+    def _extract_errors(self) -> list[str]:
+        """Extract unique errors from conversations."""
+        errors = set()
+        
+        for msg in self.all_messages:
+            content = msg.get("content", "")
+            for pattern in self.ERROR_PATTERNS:
+                matches = re.findall(pattern, content, re.IGNORECASE)
+                for match in matches:
+                    error_text = match.strip()[:100]
+                    if len(error_text) > 10:
+                        errors.add(error_text)
+        
+        return sorted(errors)
+    
+    def to_markdown(self) -> str:
+        """Generate markdown report."""
+        report = self.analyze()
+        
+        lines = [
+            "# Conversation Analysis Report",
+            "",
+            f"**Generated**: {datetime.now().strftime('%Y-%m-%d %H:%M')}",
+            f"**Period**: {report.date_range[0]} to {report.date_range[1]}",
+            "",
+            "---",
+            "",
+            "## Summary",
+            "",
+            f"- **Total Conversations**: {report.total_conversations}",
+            f"- **Total Messages**: {report.total_messages}",
+            "",
+            "## Top Topics",
+            "",
+        ]
+        
+        for topic, count in report.top_topics:
+            lines.append(f"- {topic.title()}: {count} mentions")
+        
+        lines.extend([
+            "",
+            "## Key Insights",
+            "",
+        ])
+        
+        for insight in report.insights:
+            lines.append(f"### {insight.title}")
+            lines.append(f"")
+            lines.append(f"{insight.description}")
+            if insight.actionable:
+                lines.append(f"")
+                lines.append(f"**Action**: {insight.action}")
+            lines.append("")
+        
+        if report.errors_encountered:
+            lines.extend([
+                "## Common Errors",
+                "",
+            ])
+            for error in report.errors_encountered[:10]:
+                lines.append(f"- `{error}`")
+        
+        lines.extend([
+            "",
+            "---",
+            "",
+            "*Generated by up-cli conversation summarizer*",
+        ])
+        
+        return "\n".join(lines)
+    
+    def to_json(self) -> str:
+        """Generate JSON report."""
+        report = self.analyze()
+        
+        return json.dumps({
+            "generated_at": datetime.now().isoformat(),
+            "summary": {
+                "total_conversations": report.total_conversations,
+                "total_messages": report.total_messages,
+                "date_range": report.date_range,
+            },
+            "top_topics": [
+                {"topic": t, "count": c} for t, c in report.top_topics
+            ],
+            "insights": [
+                {
+                    "title": i.title,
+                    "description": i.description,
+                    "actionable": i.actionable,
+                    "action": i.action if i.actionable else None,
+                }
+                for i in report.insights
+            ],
+            "errors": report.errors_encountered,
+        }, indent=2)
+
+
+def summarize_cursor_history(output_format: str = "markdown") -> str:
+    """Summarize Cursor chat history.
+    
+    Args:
+        output_format: 'markdown' or 'json'
+        
+    Returns:
+        Formatted summary report
+    """
+    # Import the export script
+    try:
+        from scripts.export_cursor_history import load_all_data
+    except ImportError:
+        # Try relative import
+        import sys
+        sys.path.insert(0, str(Path(__file__).parent.parent.parent / "scripts"))
+        from export_cursor_history import load_all_data
+    
+    conversations = load_all_data()
+    summarizer = ConversationSummarizer(conversations)
+    
+    if output_format == "json":
+        return summarizer.to_json()
+    return summarizer.to_markdown()
+
+
+# CLI
+if __name__ == "__main__":
+    import sys
+    
+    output_format = "markdown"
+    if len(sys.argv) > 1 and sys.argv[1] == "--json":
+        output_format = "json"
+    
+    try:
+        result = summarize_cursor_history(output_format)
+        print(result)
+    except Exception as e:
+        print(f"Error: {e}", file=sys.stderr)
+        sys.exit(1)

up/templates/__init__.py

@@ -18,13 +18,20 @@ def scaffold_project(
     from up.templates.loop import create_loop_system
     from up.templates.docs_skill import create_docs_skill
     from up.templates.config import create_config_files
+    from up.templates.mcp import create_mcp_config
+    from up.context import create_context_budget_file
 
     # Create base structure
     _create_base_structure(target_dir, ai_target)
 
-    # Create config files
+    # Create config files (CLAUDE.md, .cursor/rules/, etc.)
     create_config_files(target_dir, ai_target, force)
 
+    # Create context budget tracking
+    if ai_target in ("claude", "both"):
+        console.print("  [dim]Creating context budget tracking...[/]")
+        create_context_budget_file(target_dir)
+
     # Create selected systems
     if "docs" in systems:
         console.print("  [dim]Creating docs system...[/]")
@@ -39,6 +46,13 @@ def scaffold_project(
         console.print("  [dim]Creating product-loop system...[/]")
         create_loop_system(target_dir, ai_target, force)
 
+    if "mcp" in systems:
+        console.print("  [dim]Creating MCP configuration...[/]")
+        create_mcp_config(target_dir, ai_target, force)
+
+    # Create handoff file
+    _create_handoff_file(target_dir, force)
+
 
 def _create_base_structure(target_dir: Path, ai_target: str) -> None:
     """Create base directory structure."""
@@ -48,7 +62,61 @@ def _create_base_structure(target_dir: Path, ai_target: str) -> None:
     if ai_target in ("claude", "both"):
         dirs.append(".claude/skills")
     if ai_target in ("cursor", "both"):
-        dirs.append(".cursor")
+        dirs.append(".cursor/rules")
 
     for d in dirs:
         (target_dir / d).mkdir(parents=True, exist_ok=True)
+
+
+def _create_handoff_file(target_dir: Path, force: bool) -> None:
+    """Create initial handoff file for session continuity."""
+    from datetime import date
+    
+    handoff_dir = target_dir / "docs/handoff"
+    handoff_dir.mkdir(parents=True, exist_ok=True)
+    
+    content = f"""# Latest Session Handoff
+
+**Date**: {date.today().isoformat()}
+**Status**: 🟢 Ready
+
+---
+
+## Session Summary
+
+Project initialized with up-cli.
+
+## What Was Done
+
+- Initialized project structure
+- Set up documentation system
+- Configured AI assistant integration
+
+## Current State
+
+- All systems initialized and ready
+- No blockers
+
+## Next Steps
+
+1. Define project vision in `docs/roadmap/vision/PRODUCT_VISION.md`
+2. Run `/learn auto` to analyze project
+3. Begin development with `/product-loop`
+
+## Files Modified
+
+- CLAUDE.md
+- .cursorrules
+- docs/* (initial structure)
+
+## Notes
+
+Ready for development.
+
+---
+
+*Update this file at the end of each session*
+"""
+    filepath = handoff_dir / "LATEST.md"
+    if not filepath.exists() or force:
+        filepath.write_text(content)