feat(summarization): LLM-based code summaries via local OpenAI

SUMMARIZATION_README.md

@@ -0,0 +1,99 @@
+# Code Summarization Feature
+
+## Overview
+
+LLM-based code summarization service for AST-RAG that generates structured summaries of functions, methods, and classes.
+
+## Files
+
+- `ast_rag/summarizer.py` - Main service implementation
+- `ast_rag/cli.py` - CLI command (`ast-rag summarize`)
+- `ast_rag/ast_rag_mcp.py` - MCP tool (`summarize_code`)
+- `docs/SUMMARIZATION.md` - Full documentation
+- `tests/test_summarizer.py` - Unit tests
+- `demo_summarizer.py` - Demo script
+- `summarizer_config_example.json` - Example configuration
+
+## Quick Start
+
+### 1. Install Ollama
+
+```bash
+curl -fsSL https://ollama.ai/install.sh | sh
+ollama pull qwen2.5-coder:14b
+```
+
+### 2. Summarize Code
+
+```bash
+# CLI usage
+ast-rag summarize com.example.MyFunction
+
+# JSON output
+ast-rag summarize MyFunction --output json
+
+# Force regeneration
+ast-rag summarize MyFunction --force
+```
+
+### 3. Programmatic Usage
+
+```python
+from ast_rag.summarizer import SummarizerService
+
+summarizer = SummarizerService(
+    base_url="http://localhost:11434/v1",
+    model="qwen2.5-coder:14b",
+)
+
+summary = summarizer.summarize_node(node_id="abc123", api=rag_api)
+print(summary.summary)
+print(summary.to_markdown())
+```
+
+## Output Format
+
+```json
+{
+  "node_id": "abc123",
+  "summary": "Processes HTTP requests...",
+  "inputs": [{"name": "request", "type": "HttpRequest", "description": "..."}],
+  "outputs": [{"name": "return", "type": "HttpResponse", "description": "..."}],
+  "side_effects": ["Database write", "Logging"],
+  "calls": ["com.example.repo.save"],
+  "called_by": ["com.example.controller.handle"],
+  "complexity": "medium",
+  "tags": ["async", "io"]
+}
+```
+
+## Features
+
+- ✅ Structured JSON output
+- ✅ Caller/callee context
+- ✅ Complexity estimation
+- ✅ Auto-tagging
+- ✅ Summary caching
+- ✅ Multiple output formats (Markdown, JSON, text)
+- ✅ CLI and MCP integration
+- ✅ Batch summarization
+
+## Configuration
+
+See `summarizer_config_example.json` for configuration options.
+
+Key settings:
+- `llm.base_url`: OpenAI-compatible API endpoint
+- `llm.model`: Model name for summarization
+- `cache.enabled`: Enable/disable caching
+- `context.max_callers/callees`: Context size
+
+## Testing
+
+```bash
+pytest tests/test_summarizer.py -v
+```
+
+## Documentation
+
+See `docs/SUMMARIZATION.md` for full documentation.