bzzz/slurp/context-intelligence/context_resolver.py

#!/usr/bin/env python3
"""
Cascading Context Resolution System

This system efficiently resolves context for UCXL addresses by:
1. Loading the cascading metadata hierarchy
2. Resolving context through CSS-like inheritance
3. Providing fast lookups with caching
4. Supporting temporal context evolution (decisions change over time)

The temporal component you mentioned is crucial - we need to track how context 
and decisions evolve over time. A temporal graph network could indeed be the 
way to handle this, where each context node has temporal versions.

Usage:
    from context_resolver import CascadingContextResolver
    
    resolver = CascadingContextResolver("~/chorus/project-metadata/BZZZ")
    context = resolver.resolve("ucxl://any:any@BZZZ:RUSTLE-testing/src/main.rs")
"""

import json
from pathlib import Path
from typing import Dict, List, Optional, Any, Tuple
from datetime import datetime, timezone
from dataclasses import dataclass
import logging

logger = logging.getLogger(__name__)

@dataclass
class ResolvedContext:
    """Complete resolved context for a UCXL address"""
    ucxl_address: str
    summary: str
    purpose: str
    technologies: List[str]
    tags: List[str]
    insights: List[str]
    
    # Metadata
    file_type: str
    size_bytes: Optional[int]
    language: Optional[str]
    last_modified: Optional[str]
    content_hash: Optional[str]
    
    # Resolution metadata
    context_source_path: str
    inheritance_chain: List[str]
    resolution_confidence: float
    resolved_at: str
    
    # Temporal metadata (for future enhancement)
    temporal_version: int = 1
    context_evolution: List[str] = None

class CascadingContextResolver:
    """Efficiently resolves cascading context for UCXL addresses"""
    
    def __init__(self, metadata_dir: str, max_hierarchy_depth: int = 10):
        self.metadata_dir = Path(metadata_dir)
        self.context_hierarchy = {}
        self.path_metadata = {}
        self.lookup_index = {}
        self.resolution_cache = {}
        self.global_contexts = []  # Contexts that apply to all paths
        self.max_hierarchy_depth = max_hierarchy_depth
        
        self.load_metadata_system()
    
    def load_metadata_system(self) -> None:
        """Load the cascading metadata system from files"""
        
        try:
            # Load context hierarchy
            hierarchy_file = self.metadata_dir / "context_hierarchy.json"
            if hierarchy_file.exists():
                with open(hierarchy_file, 'r') as f:
                    self.context_hierarchy = json.load(f)
                logger.info(f"Loaded {len(self.context_hierarchy)} context nodes")
            
            # Load path metadata
            paths_file = self.metadata_dir / "path_metadata.json"
            if paths_file.exists():
                with open(paths_file, 'r') as f:
                    self.path_metadata = json.load(f)
                logger.info(f"Loaded {len(self.path_metadata)} path entries")
            
            # Load lookup index
            index_file = self.metadata_dir / "context_lookup_index.json"
            if index_file.exists():
                with open(index_file, 'r') as f:
                    self.lookup_index = json.load(f)
                logger.info(f"Loaded {len(self.lookup_index)} lookup entries")
            
            # Load global contexts
            global_file = self.metadata_dir / "global_contexts.json"
            if global_file.exists():
                with open(global_file, 'r') as f:
                    self.global_contexts = json.load(f)
                logger.info(f"Loaded {len(self.global_contexts)} global contexts")
                
        except Exception as e:
            logger.error(f"Error loading metadata system: {e}")
            raise
    
    def resolve(self, ucxl_address: str) -> Optional[ResolvedContext]:
        """Resolve context for a UCXL address using cascading inheritance"""
        
        # Check cache first
        if ucxl_address in self.resolution_cache:
            return self.resolution_cache[ucxl_address]
        
        # Find the filesystem path from UCXL address
        filesystem_path = self.ucxl_to_filesystem_path(ucxl_address)
        if not filesystem_path:
            return None
        
        # Get path metadata
        path_meta = self.path_metadata.get(filesystem_path)
        if not path_meta:
            return None
        
        # Resolve context through hierarchy with bounded traversal
        resolved_context = self.resolve_cascading_context(filesystem_path, self.max_hierarchy_depth)
        if not resolved_context:
            return None
        
        # Combine path metadata with resolved context
        result = ResolvedContext(
            ucxl_address=ucxl_address,
            summary=resolved_context.get('summary', ''),
            purpose=resolved_context.get('purpose', ''),
            technologies=resolved_context.get('technologies', []),
            tags=resolved_context.get('tags', []),
            insights=resolved_context.get('insights', []),
            
            file_type=path_meta.get('file_type', 'unknown'),
            size_bytes=path_meta.get('size_bytes'),
            language=path_meta.get('language'),
            last_modified=path_meta.get('last_modified'),
            content_hash=path_meta.get('content_hash'),
            
            context_source_path=resolved_context.get('source_path', ''),
            inheritance_chain=resolved_context.get('inheritance_chain', []),
            resolution_confidence=resolved_context.get('confidence', 0.5),
            resolved_at=datetime.now(timezone.utc).isoformat(),
            
            context_evolution=[]  # Placeholder for temporal versioning
        )
        
        # Cache the result
        self.resolution_cache[ucxl_address] = result
        
        return result
    
    def ucxl_to_filesystem_path(self, ucxl_address: str) -> Optional[str]:
        """Convert UCXL address to filesystem path"""
        
        # Parse UCXL address: ucxl://agent:role@project:task/path
        if not ucxl_address.startswith("ucxl://"):
            return None
        
        try:
            # Remove scheme
            remainder = ucxl_address[7:]  # Remove "ucxl://"
            
            # Split by @ to separate authority and path
            if '@' not in remainder:
                return None
            
            authority_part, path_part = remainder.split('@', 1)
            
            # Split by first / to separate project:task from path
            if '/' in path_part:
                project_task, file_path = path_part.split('/', 1)
                
                # Look for matching path in our metadata
                for fs_path, meta in self.path_metadata.items():
                    if meta['ucxl_address'] == ucxl_address:
                        return fs_path
                
                # If not found, try to construct from UCXL path
                # This is project-specific logic for BZZZ
                if project_task == "BZZZ:RUSTLE-testing":
                    base_path = "/home/tony/chorus/project-queues/active/BZZZ"  # Could be configurable
                    constructed_path = str(Path(base_path) / file_path)
                    if constructed_path in self.path_metadata:
                        return constructed_path
            
            return None
            
        except Exception as e:
            logger.warning(f"Error parsing UCXL address {ucxl_address}: {e}")
            return None
    
    def resolve_cascading_context(self, filesystem_path: str, max_depth: int = 10) -> Optional[Dict[str, Any]]:
        """Resolve context by walking up the directory hierarchy with bounded traversal"""
        
        current_path = Path(filesystem_path)
        if not current_path.is_dir():
            current_path = current_path.parent
        
        contexts = []
        inheritance_chain = []
        depth = 0
        
        # First, check for global contexts
        global_contexts = self._get_global_contexts()
        
        # Walk up the directory tree with bounded depth
        while depth < max_depth:
            path_str = str(current_path)
            
            if path_str in self.context_hierarchy:
                context_node = self.context_hierarchy[path_str]
                contexts.append(context_node)
                inheritance_chain.append(path_str)
                
                # Stop if this context doesn't cascade to children
                if not context_node.get('applies_to_children', True):
                    break
                
                # Stop if this context overrides parent
                if context_node.get('overrides_parent', False):
                    break
            
            # Move up one level
            parent = current_path.parent
            if parent == current_path:  # Reached filesystem root
                break
            current_path = parent
            depth += 1
        
        # Add global contexts (they apply everywhere)
        contexts.extend(global_contexts)
        
        if not contexts:
            return None
        
        # Merge contexts (most specific first, global contexts last)
        merged_context = self.merge_contexts(contexts)
        merged_context['inheritance_chain'] = inheritance_chain
        merged_context['source_path'] = inheritance_chain[0] if inheritance_chain else filesystem_path
        merged_context['confidence'] = min(1.0, len(contexts) * 0.2 + 0.6)  # Higher confidence with more context layers
        merged_context['bounded_depth'] = depth
        merged_context['global_contexts_applied'] = len(global_contexts) > 0
        
        return merged_context
    
    def merge_contexts(self, contexts: List[Dict[str, Any]]) -> Dict[str, Any]:
        """Merge multiple contexts using CSS-like specificity"""
        
        if len(contexts) == 1:
            return contexts[0].copy()
        
        # Sort by specificity (higher = more specific)
        contexts.sort(key=lambda c: c.get('context_specificity', 0), reverse=True)
        
        # Start with most specific context
        merged = contexts[0].copy()
        
        # Merge in less specific contexts
        for context in contexts[1:]:
            # Tags are additive (union)
            merged_tags = set(merged.get('tags', []))
            merged_tags.update(context.get('tags', []))
            merged['tags'] = list(merged_tags)
            
            # Technologies are additive (union)
            merged_tech = set(merged.get('technologies', []))
            merged_tech.update(context.get('technologies', []))
            merged['technologies'] = list(merged_tech)
            
            # Insights are additive (append unique)
            merged_insights = merged.get('insights', []).copy()
            for insight in context.get('insights', []):
                if insight not in merged_insights:
                    merged_insights.append(insight)
            merged['insights'] = merged_insights
            
            # Use most specific for summary/purpose unless empty
            if not merged.get('summary'):
                merged['summary'] = context.get('summary', '')
            if not merged.get('purpose'):
                merged['purpose'] = context.get('purpose', '')
        
        return merged
    
    def _get_global_contexts(self) -> List[Dict[str, Any]]:
        """Get contexts marked as global (apply everywhere)"""
        return [context for context in self.global_contexts 
                if context.get('is_global', False)]
    
    def add_global_context(self, context: Dict[str, Any]) -> None:
        """Add a context that applies globally to all paths"""
        context['is_global'] = True
        context['context_specificity'] = -1  # Lowest specificity
        self.global_contexts.append(context)
        self.clear_cache()  # Clear cache since global context affects everything
    
    def remove_global_context(self, context_id: str) -> bool:
        """Remove a global context by its ID"""
        initial_count = len(self.global_contexts)
        self.global_contexts = [ctx for ctx in self.global_contexts 
                               if ctx.get('id') != context_id]
        if len(self.global_contexts) < initial_count:
            self.clear_cache()
            return True
        return False
    
    def set_hierarchy_depth_limit(self, max_depth: int) -> None:
        """Set the maximum depth for hierarchy traversal"""
        self.max_hierarchy_depth = max_depth
        self.clear_cache()  # Clear cache since depth limit affects resolution
    
    def get_context_statistics(self) -> Dict[str, Any]:
        """Get statistics about the loaded context system"""
        
        return {
            'context_nodes': len(self.context_hierarchy),
            'path_entries': len(self.path_metadata),
            'lookup_entries': len(self.lookup_index),
            'cached_resolutions': len(self.resolution_cache),
            'global_contexts': len(self.global_contexts),
            'max_hierarchy_depth': self.max_hierarchy_depth,
            'metadata_directory': str(self.metadata_dir)
        }
    
    def clear_cache(self) -> None:
        """Clear the resolution cache"""
        self.resolution_cache.clear()
    
    def batch_resolve(self, ucxl_addresses: List[str]) -> Dict[str, ResolvedContext]:
        """Efficiently resolve multiple UCXL addresses"""
        
        results = {}
        
        for address in ucxl_addresses:
            context = self.resolve(address)
            if context:
                results[address] = context
        
        return results
    
    def search_by_tag(self, tag: str) -> List[ResolvedContext]:
        """Find all contexts containing a specific tag"""
        
        results = []
        
        # Search through all path metadata
        for fs_path, meta in self.path_metadata.items():
            ucxl_address = meta.get('ucxl_address')
            if not ucxl_address:
                continue
            
            context = self.resolve(ucxl_address)
            if context and tag.lower() in [t.lower() for t in context.tags]:
                results.append(context)
        
        return results
    
    def search_by_technology(self, technology: str) -> List[ResolvedContext]:
        """Find all contexts using a specific technology"""
        
        results = []
        
        for fs_path, meta in self.path_metadata.items():
            ucxl_address = meta.get('ucxl_address')
            if not ucxl_address:
                continue
            
            context = self.resolve(ucxl_address)
            if context and technology.lower() in [t.lower() for t in context.technologies]:
                results.append(context)
        
        return results

# Example usage and demo
def demo_context_resolution():
    """Demonstrate the cascading context resolution system"""
    
    metadata_dir = Path.home() / "chorus" / "project-metadata" / "BZZZ"
    
    if not metadata_dir.exists():
        logger.error(f"Metadata directory not found: {metadata_dir}")
        logger.info("Run cascading_metadata_generator.py first to generate the hierarchy")
        return
    
    resolver = CascadingContextResolver(str(metadata_dir))
    
    # Demo UCXL addresses
    demo_addresses = [
        "ucxl://any:any@BZZZ:RUSTLE-testing/src/main.rs",
        "ucxl://any:any@BZZZ:RUSTLE-testing/src/api/handlers.rs",
        "ucxl://any:any@BZZZ:RUSTLE-testing/tests/unit/core_tests.rs",
        "ucxl://any:any@BZZZ:RUSTLE-testing/assets/fonts/README.md",
        "ucxl://any:any@BZZZ:RUSTLE-testing/config/settings.toml"
    ]
    
    logger.info("🔍 Context Resolution Demo")
    logger.info("=" * 50)
    
    for address in demo_addresses:
        logger.info(f"\n🔗 Resolving: {address}")
        
        context = resolver.resolve(address)
        if context:
            logger.info(f"   📝 Summary: {context.summary}")
            logger.info(f"   🎯 Purpose: {context.purpose}")
            logger.info(f"   🔧 Technologies: {', '.join(context.technologies[:3])}")
            logger.info(f"   🏷️ Tags: {', '.join(context.tags[:5])}")
            logger.info(f"   📊 Source: {Path(context.context_source_path).name}")
            logger.info(f"   🔗 Chain: {' → '.join([Path(p).name for p in context.inheritance_chain])}")
            logger.info(f"   ✨ Confidence: {context.resolution_confidence:.2f}")
        else:
            logger.warning(f"   ❌ Could not resolve context")
    
    # Show statistics
    stats = resolver.get_context_statistics()
    logger.info(f"\n📊 System Statistics:")
    logger.info(f"   🏗️ Context nodes: {stats['context_nodes']}")
    logger.info(f"   📄 Path entries: {stats['path_entries']}")
    logger.info(f"   🔍 Lookup entries: {stats['lookup_entries']}")
    logger.info(f"   💾 Cached resolutions: {stats['cached_resolutions']}")
    
    # Demo search functionality
    logger.info(f"\n🔎 Search Demo:")
    rust_contexts = resolver.search_by_technology("rust")
    logger.info(f"   🦀 Found {len(rust_contexts)} Rust contexts")
    
    source_contexts = resolver.search_by_tag("source-code")
    logger.info(f"   💻 Found {len(source_contexts)} source code contexts")

if __name__ == "__main__":
    logging.basicConfig(level=logging.INFO, format='%(levelname)s - %(message)s')
    demo_context_resolution()