File size: 14,613 Bytes
5f78432
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
# services/mcp_server.py
"""
Model Context Protocol (MCP) server for MasterLLM.
Exposes CrewAI tools via standardized MCP protocol for external integration.
"""

import json
import os
from typing import Any, Dict, List, Optional
from mcp.server import Server
from mcp.types import Tool, TextContent, ImageContent, EmbeddedResource
from mcp.server.stdio import stdio_server

# Import CrewAI tools
from services.agent_crewai import (
    ExtractTextTool,
    ExtractTablesTool,
    DescribeImagesTool,
    SummarizeTextTool,
    ClassifyTextTool,
    ExtractEntitesTool,
    TranslateTextTool,
    SignatureVerificationTool,
    StampDetectionTool,
    get_master_tools,
    run_agent,
)


# ========================
# MCP SERVER SETUP
# ========================

class MasterLLMMCPServer:
    """MCP Server for MasterLLM document processing tools."""
    
    def __init__(self, name: str = "masterllm-orchestrator"):
        self.server = Server(name)
        self.tools = get_master_tools()
        self._setup_handlers()
    
    def _setup_handlers(self):
        """Register MCP protocol handlers."""
        
        @self.server.list_tools()
        async def list_tools() -> List[Tool]:
            """List all available tools exposed via MCP."""
            mcp_tools = []
            
            for tool in self.tools:
                # Convert CrewAI tool to MCP tool format
                mcp_tool = Tool(
                    name=tool.name,
                    description=tool.description,
                    inputSchema={
                        "type": "object",
                        "properties": self._get_tool_schema(tool.name),
                        "required": self._get_required_fields(tool.name),
                    }
                )
                mcp_tools.append(mcp_tool)
            
            return mcp_tools
        
        @self.server.call_tool()
        async def call_tool(name: str, arguments: dict) -> List[TextContent]:
            """Execute a tool and return results."""
            # Find the matching CrewAI tool
            matching_tool = None
            for tool in self.tools:
                if tool.name == name:
                    matching_tool = tool
                    break
            
            if not matching_tool:
                return [TextContent(
                    type="text",
                    text=json.dumps({"error": f"Tool '{name}' not found"})
                )]
            
            try:
                # Execute the CrewAI tool
                result = matching_tool._run(**arguments)
                
                # Parse result if it's a JSON string
                if isinstance(result, str):
                    try:
                        result = json.loads(result)
                    except json.JSONDecodeError:
                        pass
                
                return [TextContent(
                    type="text",
                    text=json.dumps(result, indent=2)
                )]
            
            except Exception as e:
                return [TextContent(
                    type="text",
                    text=json.dumps({
                        "error": str(e),
                        "tool": name,
                        "arguments": arguments
                    })
                )]
        
        @self.server.list_resources()
        async def list_resources() -> List[Any]:
            """List available resources (e.g., workflow templates, history)."""
            # Can be extended to expose MongoDB records, S3 files, etc.
            return [
                {
                    "uri": "workflow://templates",
                    "name": "Workflow Templates",
                    "description": "Pre-configured document processing workflows",
                    "mimeType": "application/json"
                },
                {
                    "uri": "workflow://history",
                    "name": "Execution History",
                    "description": "Recent workflow execution history",
                    "mimeType": "application/json"
                }
            ]
        
        @self.server.read_resource()
        async def read_resource(uri: str) -> str:
            """Read a specific resource."""
            if uri == "workflow://templates":
                templates = {
                    "document_analysis": {
                        "pipeline": "text-table-summarize",
                        "description": "Extract text and tables, then summarize"
                    },
                    "multilingual_processing": {
                        "pipeline": "text-translate-summarize",
                        "description": "Extract, translate, and summarize document"
                    },
                    "verification": {
                        "pipeline": "signature_verification-stamp_detection",
                        "description": "Verify signatures and detect stamps"
                    }
                }
                return json.dumps(templates, indent=2)
            
            elif uri == "workflow://history":
                # This could query MongoDB for recent executions
                # For now, return placeholder
                return json.dumps({
                    "message": "Connect to MongoDB to view execution history",
                    "recent_workflows": []
                }, indent=2)
            
            return json.dumps({"error": f"Resource not found: {uri}"})
        
        @self.server.list_prompts()
        async def list_prompts() -> List[Any]:
            """List available prompt templates."""
            return [
                {
                    "name": "analyze_document",
                    "description": "Comprehensive document analysis workflow",
                    "arguments": [
                        {
                            "name": "file_path",
                            "description": "Path to the document file",
                            "required": True
                        },
                        {
                            "name": "analysis_depth",
                            "description": "Level of analysis: basic, standard, or comprehensive",
                            "required": False
                        }
                    ]
                },
                {
                    "name": "extract_and_summarize",
                    "description": "Extract content and generate summary",
                    "arguments": [
                        {
                            "name": "file_path",
                            "description": "Path to the document file",
                            "required": True
                        },
                        {
                            "name": "include_tables",
                            "description": "Whether to include tables in summary",
                            "required": False
                        }
                    ]
                }
            ]
        
        @self.server.get_prompt()
        async def get_prompt(name: str, arguments: dict) -> Any:
            """Get a specific prompt with filled arguments."""
            if name == "analyze_document":
                file_path = arguments.get("file_path", "")
                depth = arguments.get("analysis_depth", "standard")
                
                if depth == "comprehensive":
                    instruction = f"Perform comprehensive analysis on {file_path}: extract text, tables, describe images, classify content, extract entities, verify signatures, and detect stamps. Then provide a detailed summary."
                elif depth == "basic":
                    instruction = f"Perform basic analysis on {file_path}: extract text and provide a brief summary."
                else:  # standard
                    instruction = f"Analyze {file_path}: extract text and tables, then provide a summary of the content."
                
                return {
                    "messages": [
                        {
                            "role": "user",
                            "content": {
                                "type": "text",
                                "text": instruction
                            }
                        }
                    ]
                }
            
            elif name == "extract_and_summarize":
                file_path = arguments.get("file_path", "")
                include_tables = arguments.get("include_tables", "true").lower() == "true"
                
                if include_tables:
                    instruction = f"Extract text and tables from {file_path}, then create a comprehensive summary including the table data."
                else:
                    instruction = f"Extract text from {file_path} and create a summary."
                
                return {
                    "messages": [
                        {
                            "role": "user",
                            "content": {
                                "type": "text",
                                "text": instruction
                            }
                        }
                    ]
                }
            
            return {"error": f"Prompt not found: {name}"}
    
    def _get_tool_schema(self, tool_name: str) -> Dict[str, Any]:
        """Get JSON schema for tool parameters."""
        base_file_schema = {
            "file_path": {
                "type": "string",
                "description": "Absolute or relative path to the file"
            },
            "start_page": {
                "type": "integer",
                "description": "Start page (1-indexed)",
                "default": 1
            },
            "end_page": {
                "type": "integer",
                "description": "End page (inclusive, 1-indexed)",
                "default": 1
            }
        }
        
        text_or_file_schema = {
            "text": {
                "type": "string",
                "description": "Raw text to process (alternative to file_path)"
            },
            "file_path": {
                "type": "string",
                "description": "Path to document file (alternative to text)"
            },
            "start_page": {
                "type": "integer",
                "description": "Start page for file processing",
                "default": 1
            },
            "end_page": {
                "type": "integer",
                "description": "End page for file processing",
                "default": 1
            }
        }
        
        schemas = {
            "extract_text": base_file_schema,
            "extract_tables": base_file_schema,
            "describe_images": base_file_schema,
            "summarize_text": text_or_file_schema,
            "classify_text": text_or_file_schema,
            "extract_entities": text_or_file_schema,
            "translate_text": {
                **text_or_file_schema,
                "target_lang": {
                    "type": "string",
                    "description": "Target language code (e.g., 'es', 'fr', 'de') or name (e.g., 'Spanish')"
                }
            },
            "signature_verification": base_file_schema,
            "stamp_detection": base_file_schema,
        }
        
        return schemas.get(tool_name, {})
    
    def _get_required_fields(self, tool_name: str) -> List[str]:
        """Get required fields for each tool."""
        file_based_tools = [
            "extract_text",
            "extract_tables", 
            "describe_images",
            "signature_verification",
            "stamp_detection"
        ]
        
        if tool_name in file_based_tools:
            return ["file_path"]
        elif tool_name == "translate_text":
            return ["target_lang"]
        else:
            return []  # text or file_path required, but either is acceptable
    
    async def run(self):
        """Run the MCP server using stdio transport."""
        async with stdio_server() as (read_stream, write_stream):
            await self.server.run(
                read_stream,
                write_stream,
                self.server.create_initialization_options()
            )


# ========================
# FASTAPI INTEGRATION
# ========================

def create_mcp_fastapi_routes(app):
    """
    Add MCP SSE (Server-Sent Events) endpoints to FastAPI app.
    This allows MCP clients to connect via HTTP instead of stdio.
    """
    from mcp.server.sse import SseServerTransport
    from fastapi import Request
    from fastapi.responses import StreamingResponse
    from sse_starlette import EventSourceResponse
    
    mcp_server = MasterLLMMCPServer()
    
    @app.get("/mcp/sse")
    async def mcp_sse_endpoint(request: Request):
        """SSE endpoint for MCP protocol."""
        from mcp.server.sse import sse_transport
        
        async def event_generator():
            async with sse_transport() as (read_stream, write_stream):
                await mcp_server.server.run(
                    read_stream,
                    write_stream,
                    mcp_server.server.create_initialization_options()
                )
        
        return EventSourceResponse(event_generator())
    
    @app.post("/mcp/message")
    async def mcp_post_endpoint(request: Request):
        """POST endpoint for MCP messages (alternative to SSE)."""
        data = await request.json()
        
        # Handle MCP JSON-RPC requests
        method = data.get("method")
        params = data.get("params", {})
        
        if method == "tools/list":
            tools = await mcp_server.server._tool_list_handler()
            return {"jsonrpc": "2.0", "result": tools, "id": data.get("id")}
        
        elif method == "tools/call":
            name = params.get("name")
            arguments = params.get("arguments", {})
            result = await mcp_server.server._tool_call_handler(name, arguments)
            return {"jsonrpc": "2.0", "result": result, "id": data.get("id")}
        
        return {"jsonrpc": "2.0", "error": {"code": -32601, "message": "Method not found"}, "id": data.get("id")}


# ========================
# STANDALONE SERVER
# ========================

async def main():
    """Run MCP server in standalone mode (stdio transport)."""
    server = MasterLLMMCPServer()
    await server.run()


if __name__ == "__main__":
    import asyncio
    asyncio.run(main())