Restore Dockerfile

.env.example

@@ -0,0 +1,32 @@
+# MCP Summary Server - Environment Variables
+
+# Server Configuration
+PORT=8080
+
+# Authentication (optional)
+# If set, requests must include: Authorization: Bearer <API_KEY>
+API_KEY=
+
+# LLM Configuration
+OPENAPI_URL=http://localhost:8080/v1
+OPENAPI_API_KEY=
+MODEL_NAME=gpt-4o
+
+# LLM Call Timeout in seconds (increase for large documents)
+LLM_TIMEOUT=120
+
+# Summarization Configuration
+# Characters per chunk when splitting long text
+CHUNK_SIZE=4000
+
+# Characters of overlap between chunks to maintain context
+OVERLAP=200
+
+# Target length for intermediate chunk summaries (words)
+TARGET_INTERMEDIATE_SUMMARY_LENGTH=150
+
+# Maximum length for final synthesized summary (words)
+MAX_DIRECT_SUMMARY_LENGTH=100
+
+# Maximum text length (characters) before chunking is triggered
+MAX_DIRECT_TEXT_LENGTH=8000

Dockerfile

@@ -0,0 +1,37 @@
+# Dockerfile for MCP Summary Server
+#
+# Usage (from directory containing this Dockerfile and mcp_summary_server.py):
+#
+#   docker build -t mcp-summary .
+#   docker run -p 8080:8080 --env-file .env mcp-summary
+#
+
+FROM python:3.12-slim
+
+WORKDIR /app
+
+# Install runtime dependencies
+COPY requirements.txt .
+RUN pip install --no-cache-dir -r requirements.txt && rm requirements.txt
+
+# Copy the server script
+COPY mcp_summary_server.py /app/mcp_summary_server.py
+
+# Expose HTTP port
+EXPOSE 8080
+
+# Environment variables
+ENV PORT=8080
+ENV OPENAPI_URL=http://localhost:8080/v1
+ENV OPENAPI_API_KEY=
+ENV MODEL_NAME=gpt-4o
+ENV CHUNK_SIZE=4000
+ENV OVERLAP=200
+ENV TARGET_INTERMEDIATE_SUMMARY_LENGTH=150
+ENV MAX_DIRECT_SUMMARY_LENGTH=100
+ENV MAX_DIRECT_TEXT_LENGTH=8000
+ENV LLM_TIMEOUT=120
+ENV API_KEY=
+
+# Start the MCP summary server
+ENTRYPOINT ["python", "-u", "/app/mcp_summary_server.py"]

README.md

@@ -0,0 +1,137 @@
+# MCP Summary Server
+
+An MCP (Model Context Protocol) server for document summarization that keeps full text out of the chat context window.
+
+## Features
+
+- Automatically determines whether to summarize directly or use chunked summarization
+- All processing happens server-side
+- Returns only the summary to the client
+- Configurable chunking parameters
+- Bearer token authentication (optional)
+
+## Setup
+
+### Environment Variables
+
+Copy `.env.example` to `.env` and configure:
+
+```bash
+cp .env.example .env
+```
+
+| Variable | Default | Description |
+|----------|---------|-------------|
+| PORT | 8080 | HTTP server port |
+| API_KEY | (empty) | Bearer token for authentication |
+| OPENAPI_URL | http://localhost:8080/v1 | LLM API endpoint |
+| OPENAPI_API_KEY | (empty) | LLM API key |
+| MODEL_NAME | gpt-4o | LLM model to use |
+| LLM_TIMEOUT | 120 | LLM call timeout in seconds |
+| CHUNK_SIZE | 4000 | Characters per chunk |
+| OVERLAP | 200 | Characters of overlap between chunks |
+| TARGET_INTERMEDIATE_SUMMARY_LENGTH | 150 | Words per chunk summary |
+| MAX_DIRECT_SUMMARY_LENGTH | 100 | Max final summary length |
+| MAX_DIRECT_TEXT_LENGTH | 8000 | Max text length before chunking |
+
+## Running
+
+### Docker
+
+```bash
+# Build
+docker build -t mcp-summary .
+
+# Run with environment file
+docker run -p 8080:8080 --env-file .env mcp-summary
+
+# Run with inline environment variables
+docker run -p 8080:8080 \
+  -e OPENAPI_URL=http://localhost:8080/v1 \
+  -e OPENAPI_API_KEY=your-key \
+  -e MODEL_NAME=gpt-4o \
+  mcp-summary
+```
+
+### Python
+
+```bash
+pip install -r requirements.txt
+python mcp_summary_server.py
+```
+
+## Connecting to OpenWebUI
+
+### In OpenWebUI Admin Settings
+
+1. Go to **Admin Settings → External Tools**
+2. Click **+ (Add Server)**
+3. Set **Type** to **MCP (Streamable HTTP)**
+4. Enter your **Server URL**
+5. Set **Authentication**:
+   - **None** if no API key is configured
+   - **Bearer** if API_KEY is set (provide the key)
+6. Save
+
+### Docker Networking
+
+If running both OpenWebUI and MCP Summary in Docker:
+
+```bash
+# Use host.docker.internal to reach host machine
+docker run -p 8080:8080 \
+  -e OPENAPI_URL=http://host.docker.internal:3000/v1 \
+  -e OPENAPI_API_KEY=your-key \
+  mcp-summary
+```
+
+If both containers are on the same Docker network, use the container name directly:
+
+```bash
+docker run --network mynetwork -p 8080:8080 \
+  -e OPENAPI_URL=http://openwebui-container:8080/v1 \
+  -e OPENAPI_API_KEY=your-key \
+  mcp-summary
+```
+
+## MCP Tool
+
+### summarize_document
+
+Summarizes a document, automatically handling chunking for long text.
+
+**Parameters:**
+- `text` (string, required): The document text to summarize
+- `max_length` (integer, optional): Maximum summary length in words (default: 100)
+
+**Returns:**
+```json
+{
+  "summary": "The summarized text...",
+  "original_length": 12345,
+  "method": "direct",  // or "chunked"
+  "chunks": 1  // number of chunks used
+}
+```
+
+## Troubleshooting
+
+### "Failed to connect to MCP server"
+
+1. **Check authentication**: Ensure you haven't selected `Bearer` without a key. Switch to `None` if no token is needed.
+2. **Check network connectivity**: Ensure OpenWebUI can reach the MCP server URL
+3. **Check LLM connectivity**: Ensure the MCP server can reach the LLM at OPENAPI_URL
+4. **Check timeouts**: Increase LLM_TIMEOUT if summarization takes too long
+
+### Infinite loading screen
+
+This may occur if you configured the server as OpenAPI instead of MCP. Fix by:
+
+1. Opening Admin Settings → External Tools
+2. Disabling/deleting the problematic connection
+3. Re-adding with **Type** set to **MCP (Streamable HTTP)**
+
+### Slow initialization
+
+If the server takes longer than 10 seconds to initialize:
+- Increase `MCP_INITIALIZE_TIMEOUT` in OpenWebUI (default: 10 seconds)

diagnose.sh

@@ -0,0 +1,34 @@
+#!/bin/bash
+# Diagnostic script for MCP Summary Server
+
+echo "================================"
+echo "MCP Summary Server Diagnostics"
+echo "================================"
+
+# Check if server is running
+echo -e "\n1. Checking if server process is running..."
+ps aux | grep mcp_summary_server || echo "Server process not found"
+
+# Check if port is listening
+echo -e "\n2. Checking if port is listening..."
+netstat -tlnp 2>/dev/null | grep 8080 || echo "Port 8080 not listening"
+
+# Test basic connectivity
+echo -e "\n3. Testing basic connectivity..."
+curl -s http://localhost:8080/ || echo "Cannot connect to localhost:8080"
+
+# Test MCP initialize
+echo -e "\n4. Testing MCP initialize..."
+curl -s -X POST http://localhost:8080/ \
+  -H "Content-Type: application/json" \
+  -d '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2025-11-25","capabilities":{},"clientInfo":{"name":"test","version":"1.0.0"}}}' | jq .
+
+# Test tools list
+echo -e "\n5. Testing tools list..."
+curl -s -X POST http://localhost:8080/ \
+  -H "Content-Type: application/json" \
+  -d '{"jsonrpc":"2.0","id":2,"method":"tools/list","params":{}}' | jq .
+
+echo -e "\n================================"
+echo "Diagnostics complete"
+echo "================================"

mcp_summary_server.py

@@ -24,32 +24,11 @@ Auth:
 import json
 import os
 import sys
-import logging
 from http.server import HTTPServer, BaseHTTPRequestHandler
-from typing import Any, Dict, List, Optional
+from typing import Any, Dict, Optional
 import requests
-from requests.exceptions import RequestException
 
-# Configure logging
-logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
-logger = logging.getLogger("mcp-summary")
-
-# MCP Server Configuration
 API_KEY = os.environ.get("API_KEY", "").strip()
-PORT = int(os.environ.get("PORT", "8080"))
-
-# LLM Configuration
-OPENAPI_URL = os.environ.get("OPENAPI_URL", "http://localhost:8080/v1")
-OPENAPI_API_KEY = os.environ.get("OPENAPI_API_KEY", "")
-MODEL_NAME = os.environ.get("MODEL_NAME", "gpt-4o")
-
-# Summarization Configuration
-CHUNK_SIZE = int(os.environ.get("CHUNK_SIZE", "4000"))
-OVERLAP = int(os.environ.get("OVERLAP", "200"))
-TARGET_INTERMEDIATE_SUMMARY_LENGTH = int(os.environ.get("TARGET_INTERMEDIATE_SUMMARY_LENGTH", "150"))
-MAX_DIRECT_SUMMARY_LENGTH = int(os.environ.get("MAX_DIRECT_SUMMARY_LENGTH", "100"))
-MAX_DIRECT_TEXT_LENGTH = int(os.environ.get("MAX_DIRECT_TEXT_LENGTH", "8000"))
-LLM_TIMEOUT = int(os.environ.get("LLM_TIMEOUT", "120"))
 
 # Tool definitions
 TOOLS_LIST: Dict[str, Any] = {
@@ -85,7 +64,8 @@ def get_bearer_token(headers: Any) -> Optional[str]:
 
 
 def require_auth(headers: Any) -> bool:
-    """Check authentication. Returns True if auth passes or is not required."""
+    """Check authentication if API key is configured."""
+    # If API_KEY is not set, allow unauthenticated access
     if not API_KEY:
         return True
 
@@ -95,52 +75,55 @@ def require_auth(headers: Any) -> bool:
     return True
 
 
-def call_llm(messages: List[Dict], temperature: float = 0.3) -> str:
+def call_llm(text: str, system_prompt: str, max_tokens: int = 2000) -> str:
-    """Make an OpenAPI-compatible LLM call with error handling."""
+    """Make an OpenAPI-compatible LLM call."""
-    url = f"{OPENAPI_URL}/chat/completions"
+    openapi_url = os.environ.get("OPENAPI_URL", "http://localhost:8080/v1")
+    openapi_api_key = os.environ.get("OPENAPI_API_KEY", "")
+    model_name = os.environ.get("MODEL_NAME", "gpt-4o")
+    timeout = int(os.environ.get("LLM_TIMEOUT", "120"))
+    
+    url = f"{openapi_url}/chat/completions"
     headers = {
         "Content-Type": "application/json",
-        "Authorization": f"Bearer {OPENAPI_API_KEY}"
+        "Authorization": f"Bearer {openapi_api_key}"
     }
     
     payload = {
-        "model": MODEL_NAME,
+        "model": model_name,
-        "messages": messages,
+        "messages": [
-        "temperature": temperature,
+            {"role": "system", "content": system_prompt},
-        "max_tokens": 2000,
+            {"role": "user", "content": text}
+        ],
+        "temperature": 0.3,
+        "max_tokens": max_tokens,
         "top_p": 0.9
     }
     
-    try:
+    response = requests.post(url, headers=headers, json=payload, timeout=timeout)
-        logger.info(f"Calling LLM at {OPENAPI_URL} with model {MODEL_NAME}")
+    response.raise_for_status()
-        response = requests.post(url, headers=headers, json=payload, timeout=LLM_TIMEOUT)
-        response.raise_for_status()
     
-        data = response.json()
+    data = response.json()
-        return data["choices"][0]["message"]["content"]
+    return data["choices"][0]["message"]["content"]
-    
-    except RequestException as e:
-        logger.error(f"LLM request failed: {e}")
-        raise RuntimeError(f"Failed to connect to LLM at {OPENAPI_URL}: {str(e)}")
-    except Exception as e:
-        logger.error(f"LLM call failed: {e}")
-        raise RuntimeError(f"LLM call failed: {str(e)}")
 
 
-def chunk_text(text: str) -> List[str]:
+def chunk_text(text: str) -> list:
     """Split text into chunks with overlap for summarization."""
-    if len(text) <= CHUNK_SIZE:
+    chunk_size = int(os.environ.get("CHUNK_SIZE", "4000"))
+    overlap = int(os.environ.get("OVERLAP", "200"))
+    
+    if len(text) <= chunk_size:
         return [text]
     
     chunks = []
     start = 0
     
     while start < len(text):
-        end = min(start + CHUNK_SIZE, len(text))
+        end = min(start + chunk_size, len(text))
         
+        # Try to break at sentence/paragraph boundary
         break_point = end
         for marker in ["\n\n", "\n", ". ", "! ", "? "]:
-            pos = text.rfind(marker, start + CHUNK_SIZE // 2, end)
+            pos = text.rfind(marker, start + chunk_size // 2, end)
             if pos > start:
                 break_point = pos
                 break
@@ -149,46 +132,84 @@ def chunk_text(text: str) -> List[str]:
         if chunk.strip():
             chunks.append(chunk)
         
-        start = break_point - OVERLAP if break_point < len(text) else len(text)
+        start = break_point - overlap if break_point < len(text) else len(text)
         if start >= len(text):
             break
     
-    logger.info(f"Split text into {len(chunks)} chunks")
     return chunks
 
 
-def summarize_chunk(chunk_text: str, chunk_num: int, total_chunks: int) -> str:
+def summarize_document(text: str, max_length: int = 100) -> dict:
-    """Summarize a single chunk of text."""
+    """
-    system_prompt = f"""You are a precise legal assistant creating concise, accurate summaries.
+    Main summarization function.
     
-You are processing chunk {chunk_num} of {total_chunks} from a larger document.
+    - If text is short, summarize directly
+    - If text is long, chunk and summarize each chunk, then synthesize
+    """
+    original_length = len(text)
+    
+    text = text.strip()
+    if not text:
+        raise ValueError("Empty text provided")
+    
+    max_direct_length = int(os.environ.get("MAX_DIRECT_TEXT_LENGTH", "8000"))
+    intermediate_length = int(os.environ.get("TARGET_INTERMEDIATE_SUMMARY_LENGTH", "150"))
+    
+    # Direct summarization for shorter texts
+    if len(text) <= max_direct_length:
+        system_prompt = f"""You are a precise legal assistant creating concise, accurate summaries.
+
+Create a summary that:
+- Is approximately {max_length} words
+- Captures key points and important details
+- Uses clear, professional language
+- Preserves names, dates, and specific facts
+
+Format as plain text without bullet points."""
+        
+        user_prompt = f"""Summarize the following document:
+
+{text}
+
+Summary:"""
+        
+        summary = call_llm(user_prompt, system_prompt)
+        
+        return {
+            "summary": summary,
+            "original_length": original_length,
+            "method": "direct",
+            "chunks": 1
+        }
+    
+    # Chunked summarization for longer texts
+    chunks = chunk_text(text)
+    
+    chunk_summaries = []
+    for i, chunk in enumerate(chunks, 1):
+        system_prompt = f"""You are a precise legal assistant creating concise, accurate summaries.
+
+You are processing chunk {i} of {len(chunks)} from a larger document.
 
 Create a focused summary that:
 - Captures key points and important details
-- Is approximately {TARGET_INTERMEDIATE_SUMMARY_LENGTH} words
+- Is approximately {intermediate_length} words
 - Can be combined with other chunk summaries
 - Uses clear, professional language
 - Preserves names, dates, and specific facts
 
 Respond as plain text without bullet points."""
         
-    user_prompt = f"""Summarize this text (chunk {chunk_num} of {total_chunks}):
+        user_prompt = f"""Summarize this text (chunk {i} of {len(chunks)}):
 
-{chunk_text}
+{chunk}
 
 Summary:"""
         
-    messages = [
+        chunk_summary = call_llm(user_prompt, system_prompt)
-        {"role": "system", "content": system_prompt},
+        chunk_summaries.append(chunk_summary)
-        {"role": "user", "content": user_prompt}
-    ]
     
-    logger.info(f"Summarizing chunk {chunk_num}/{total_chunks}")
+    # Synthesize into final summary
-    return call_llm(messages)
-
-
-def synthesize_summaries(chunk_summaries: List[str]) -> str:
-    """Synthesize multiple chunk summaries into a single final summary."""
     combined = "\n\n".join(chunk_summaries)
     
     system_prompt = """You are a precise legal assistant creating executive-level summaries.
@@ -209,71 +230,7 @@ Format as a single paragraph of plain text."""
 
 Final summary:"""
     
-    messages = [
+    final_summary = call_llm(user_prompt, system_prompt)
-        {"role": "system", "content": system_prompt},
-        {"role": "user", "content": user_prompt}
-    ]
-    
-    logger.info(f"Synthesizing {len(chunk_summaries)} chunk summaries")
-    return call_llm(messages)
-
-
-def summarize_document(text: str, max_length: int = MAX_DIRECT_SUMMARY_LENGTH) -> Dict[str, Any]:
-    """
-    Main summarization function.
-    
-    - If text is short, summarize directly
-    - If text is long, chunk and summarize each chunk, then synthesize
-    """
-    original_length = len(text)
-    
-    text = text.strip()
-    if not text:
-        raise ValueError("Empty text provided")
-    
-    logger.info(f"Summarizing text of {original_length} characters")
-    
-    # Direct summarization for shorter texts
-    if len(text) <= MAX_DIRECT_TEXT_LENGTH:
-        system_prompt = f"""You are a precise legal assistant creating concise, accurate summaries.
-
-Create a summary that:
-- Is approximately {max_length} words
-- Captures key points and important details
-- Uses clear, professional language
-- Preserves names, dates, and specific facts
-
-Format as plain text without bullet points."""
-        
-        user_prompt = f"""Summarize the following document:
-
-{text}
-
-Summary:"""
-        
-        messages = [
-            {"role": "system", "content": system_prompt},
-            {"role": "user", "content": user_prompt}
-        ]
-        
-        summary = call_llm(messages)
-        
-        return {
-            "summary": summary,
-            "original_length": original_length,
-            "method": "direct",
-            "chunks": 1
-        }
-    
-    # Chunked summarization for longer texts
-    chunks = chunk_text(text)
-    
-    chunk_summaries = []
-    for i, chunk in enumerate(chunks, 1):
-        chunk_summary = summarize_chunk(chunk, i, len(chunks))
-        chunk_summaries.append(chunk_summary)
-    
-    final_summary = synthesize_summaries(chunk_summaries)
     
     return {
         "summary": final_summary,
@@ -287,7 +244,8 @@ class MCPSummaryHandler(BaseHTTPRequestHandler):
     """HTTP handler for MCP summary server."""
     
     def log_message(self, format, *args):
-        logger.info(format % args)
+        # Quiet logs by default
+        pass
 
     def _send_json(self, status: int, payload: Any):
         """Send JSON response."""
@@ -298,8 +256,8 @@ class MCPSummaryHandler(BaseHTTPRequestHandler):
         self.end_headers()
         self.wfile.write(body)
 
-    def _auth_or_401(self):
+    def _auth_or_401(self) -> bool:
-        """Check authentication. Returns False if auth fails."""
+        """Check authentication if API key is configured."""
         try:
             return require_auth(self.headers)
         except PermissionError:
@@ -312,8 +270,6 @@ class MCPSummaryHandler(BaseHTTPRequestHandler):
             self._send_json(200, {
                 "service": "mcp-summary",
                 "transport": "streamable-http",
-                "model": MODEL_NAME,
-                "status": "running",
                 "docs": "Use POST / with MCP JSON-RPC (initialize, tools/list, tools/call)."
             })
             return
@@ -322,7 +278,6 @@ class MCPSummaryHandler(BaseHTTPRequestHandler):
 
     def do_POST(self):
         """Handle MCP JSON-RPC requests."""
-        # Streamable HTTP MCP endpoint
         if self.path not in ("/", "/mcp"):
             self.send_error(404, "Not Found")
             return
@@ -346,8 +301,6 @@ class MCPSummaryHandler(BaseHTTPRequestHandler):
         params = req.get("params") or {}
         req_id = req.get("id")
 
-        logger.info(f"MCP request: method={method}, id={req_id}")
-        
         # MCP: initialize
         if method == "initialize":
             self._send_json(200, {
@@ -366,15 +319,6 @@ class MCPSummaryHandler(BaseHTTPRequestHandler):
             })
             return
 
-        # MCP: ping
-        if method == "ping":
-            self._send_json(200, {
-                "jsonrpc": "2.0",
-                "id": req_id,
-                "result": {}
-            })
-            return
-        
         # MCP: tools/list
         if method == "tools/list":
             self._send_json(200, {
@@ -400,7 +344,6 @@ class MCPSummaryHandler(BaseHTTPRequestHandler):
                     }
                 })
             except Exception as e:
-                logger.error(f"Tool call failed: {e}", exc_info=True)
                 self._send_json(200, {
                     "jsonrpc": "2.0",
                     "id": req_id,
@@ -421,7 +364,7 @@ class MCPSummaryHandler(BaseHTTPRequestHandler):
             if not text:
                 raise ValueError("Text parameter is required")
             
-            max_length = args.get("max_length", MAX_DIRECT_SUMMARY_LENGTH)
+            max_length = args.get("max_length", 100)
             return summarize_document(text, max_length)
 
         raise ValueError(f"Unknown tool: {name}")
@@ -433,11 +376,6 @@ def main():
     server = HTTPServer(("0.0.0.0", port), MCPSummaryHandler)
     mode = "auth enabled (Bearer)" if API_KEY else "no auth (API_KEY not set)"
     print(f"MCP Summary Server listening on 0.0.0.0:{port} [{mode}]")
-    print(f"  - Model: {MODEL_NAME}")
-    print(f"  - LLM URL: {OPENAPI_URL}")
-    print(f"  - Chunk size: {CHUNK_SIZE} characters")
-    print(f"  - Max direct text: {MAX_DIRECT_TEXT_LENGTH} characters")
-    print(f"  - LLM timeout: {LLM_TIMEOUT} seconds")
     try:
         server.serve_forever()
     except KeyboardInterrupt:

requirements.txt

@@ -0,0 +1,4 @@
+# requirements.txt for MCP Summary Server
+
+# HTTP requests for LLM communication
+requests>=2.31.0