File size: 23,921 Bytes
2c5e855 59de368 2c5e855 59de368 2c5e855 59de368 2c5e855 bb3909a 59de368 2c5e855 59de368 2c5e855 59de368 2c5e855 59de368 2c5e855 a929e66 2c5e855 59de368 2c5e855 a929e66 2c5e855 a929e66 2c5e855 a929e66 2c5e855 59de368 2c5e855 59de368 2c5e855 59de368 2c5e855 59de368 2c5e855 59de368 2c5e855 59de368 2c5e855 59de368 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 |
# agents.py - Core Analysis & Orchestration Agents
import os
import asyncio
import logging
from typing import Optional, Dict, Any, List, AsyncGenerator
import time
from utils import call_openai_chat, load_pdf_text_cached, load_pdf_text_chunked, get_document_metadata, create_hierarchical_summary
from config import Config
logger = logging.getLogger(__name__)
logger.setLevel(logging.INFO)
class BaseAgent:
def __init__(self, name: str, model: str, tasks_completed: int = 0):
self.name = name
self.model = model
self.tasks_completed = tasks_completed
async def handle(self, user_id: str, prompt: str, file_path: Optional[str] = None, context: Optional[Dict[str, Any]] = None) -> Dict[str, Any]:
raise NotImplementedError(f"{self.__class__.__name__}.handle must be implemented.")
async def handle_streaming(self, user_id: str, prompt: str, file_path: Optional[str] = None, context: Optional[Dict[str, Any]] = None) -> AsyncGenerator[str, None]:
"""Streaming version of handle - override in subclasses for streaming support"""
result = await self.handle(user_id, prompt, file_path, context)
# Default implementation: yield the result as a single chunk
for key, value in result.items():
yield f"{key}: {value}"
# --------------------
# Core Analysis Agent
# --------------------
class AnalysisAgent(BaseAgent):
async def handle(self, user_id: str, prompt: str, file_path: Optional[str] = None, context: Optional[Dict[str, Any]] = None):
start_time = time.time()
if file_path:
# Get document metadata
metadata = get_document_metadata(file_path)
# Load text with caching
text = load_pdf_text_cached(file_path)
# Check if document needs chunking
if len(text) > Config.CHUNK_SIZE:
return await self._handle_large_document(prompt, text, metadata)
else:
content = f"User prompt: {prompt}\n\nDocument text:\n{text}"
else:
content = f"User prompt: {prompt}"
metadata = {}
system = "You are AnalysisAgent: produce concise insights and structured summaries. Adapt your language and complexity to the target audience. Provide clear, actionable insights with appropriate examples and analogies for complex topics."
try:
response = await call_openai_chat(
model=self.model,
messages=[{"role": "system", "content": system},
{"role": "user", "content": content}],
temperature=Config.OPENAI_TEMPERATURE,
max_tokens=Config.OPENAI_MAX_TOKENS
)
except Exception as e:
logger.exception("AnalysisAgent failed")
response = f"Error during analysis: {str(e)}"
self.tasks_completed += 1
# Add processing metadata
processing_time = time.time() - start_time
result = {
"analysis": response,
"metadata": {
"processing_time": round(processing_time, 2),
"document_metadata": metadata,
"agent": self.name,
"tasks_completed": self.tasks_completed
}
}
return result
async def _handle_large_document(self, prompt: str, text: str, metadata: Dict[str, Any]) -> Dict[str, Any]:
"""Handle large documents by processing in chunks"""
from utils import chunk_text
chunks = chunk_text(text, Config.CHUNK_SIZE)
chunk_results = []
system = "You are AnalysisAgent: produce concise insights and structured summaries. Adapt your language and complexity to the target audience. Provide clear, actionable insights with appropriate examples and analogies for complex topics."
for i, chunk in enumerate(chunks):
content = f"User prompt: {prompt}\n\nDocument chunk {i+1}/{len(chunks)}:\n{chunk}"
try:
response = await call_openai_chat(
model=self.model,
messages=[{"role": "system", "content": system},
{"role": "user", "content": content}],
temperature=Config.OPENAI_TEMPERATURE,
max_tokens=Config.OPENAI_MAX_TOKENS
)
chunk_results.append(f"--- Chunk {i+1} Analysis ---\n{response}")
except Exception as e:
logger.exception(f"AnalysisAgent failed on chunk {i+1}")
chunk_results.append(f"--- Chunk {i+1} Error ---\nError: {str(e)}")
# Combine chunk results
combined_analysis = "\n\n".join(chunk_results)
# Create final summary using hierarchical approach to avoid token limits
try:
final_summary = await create_hierarchical_summary(
chunk_results=chunk_results,
prompt=prompt,
model=self.model,
max_tokens=6000 # Conservative limit to avoid context length errors
)
except Exception as e:
logger.exception("AnalysisAgent failed on final summary")
final_summary = f"Error creating final summary: {str(e)}\n\nChunk Results:\n{combined_analysis}"
return {
"analysis": final_summary,
"metadata": {
"processing_method": "chunked",
"chunks_processed": len(chunks),
"document_metadata": metadata,
"agent": self.name,
"tasks_completed": self.tasks_completed
}
}
async def handle_streaming(self, user_id: str, prompt: str, file_path: Optional[str] = None, context: Optional[Dict[str, Any]] = None) -> AsyncGenerator[str, None]:
"""Streaming version of analysis"""
yield "π Starting analysis..."
if file_path:
metadata = get_document_metadata(file_path)
yield f"π Document loaded: {metadata.get('page_count', 0)} pages, {metadata.get('file_size', 0) / 1024:.1f} KB"
text = load_pdf_text_cached(file_path)
if len(text) > Config.CHUNK_SIZE:
yield "π Large document detected, processing in chunks..."
from utils import chunk_text
chunks = chunk_text(text, Config.CHUNK_SIZE)
yield f"π Document split into {len(chunks)} chunks"
# Process chunks with progress updates
for i, chunk in enumerate(chunks):
yield f"β³ Processing chunk {i+1}/{len(chunks)}..."
# Process chunk (simplified for streaming)
await asyncio.sleep(0.1) # Simulate processing time
yield "π Combining chunk results..."
await asyncio.sleep(0.2)
yield "β
Analysis complete!"
else:
yield "β‘ Processing document..."
await asyncio.sleep(0.3)
yield "β
Analysis complete!"
else:
yield "β‘ Processing request..."
await asyncio.sleep(0.2)
yield "β
Analysis complete!"
# Get the actual result
result = await self.handle(user_id, prompt, file_path, context)
yield f"\nπ Analysis Result:\n{result.get('analysis', 'No result')}"
# --------------------
# Collaboration Agent
# --------------------
class CollaborationAgent(BaseAgent):
async def handle(self, user_id: str, prompt: str, file_path: Optional[str] = None, context: Optional[Dict[str, Any]] = None):
system = "You are CollaborationAgent: produce reviewer-style comments and suggestions for improvement. Focus on constructive feedback and actionable recommendations."
content = prompt if isinstance(prompt, str) else str(prompt)
try:
response = await call_openai_chat(model=self.model,
messages=[{"role": "system", "content": system},
{"role": "user", "content": content}],
temperature=0.2,
max_tokens=800)
except Exception as e:
logger.exception("CollaborationAgent failed")
response = f"Error during collaboration: {str(e)}"
self.tasks_completed += 1
return {"collaboration": response}
# --------------------
# Conversation Agent
# --------------------
class ConversationAgent(BaseAgent):
async def handle(self, user_id: str, prompt: str, file_path: Optional[str] = None, context: Optional[Dict[str, Any]] = None):
system = "You are ConversationAgent: respond politely and helpfully. Provide context-aware responses and guide users on how to get the best results from the analysis system."
try:
response = await call_openai_chat(model=self.model,
messages=[{"role": "system", "content": system},
{"role": "user", "content": prompt}],
temperature=0.3,
max_tokens=400)
except Exception as e:
logger.exception("ConversationAgent failed")
response = f"Error in conversation: {str(e)}"
self.tasks_completed += 1
return {"conversation": response}
# --------------------
# Senior Research Analyst Agent
# --------------------
class ResearchAnalystAgent(BaseAgent):
async def handle(self, user_id: str, prompt: str, file_path: Optional[str] = None, context: Optional[Dict[str, Any]] = None):
start_time = time.time()
if file_path:
# Get document metadata
metadata = get_document_metadata(file_path)
# Load text with caching
text = load_pdf_text_cached(file_path)
# Check if document needs chunking
if len(text) > Config.CHUNK_SIZE:
return await self._handle_large_document_research(prompt, text, metadata)
else:
content = f"User prompt: {prompt}\n\nDocument text:\n{text}"
else:
content = f"User prompt: {prompt}"
metadata = {}
system = """You are a Senior Research Analyst with deep expertise in product and engineering R&D pipelines. Your role is to:
1. **Extract High-Value Insights**: Identify novel ideas, breakthrough concepts, and innovative approaches that could drive significant product/engineering impact.
2. **Assess Commercial Viability**: Evaluate the potential for practical application, market readiness, and competitive advantage.
3. **Generate R&D Pipeline Outcomes**: Convert insights into concrete, actionable items for:
- **Experiments**: Specific hypotheses to test, methodologies to validate
- **Prototypes**: Technical implementations to build and demonstrate
- **Product Decisions**: Strategic choices for development priorities and resource allocation
4. **Prioritize by Impact**: Focus on ideas with the highest potential for transformative change and measurable business value.
Provide structured analysis with clear next steps that engineering and product teams can immediately act upon."""
try:
response = await call_openai_chat(
model=self.model,
messages=[{"role": "system", "content": system},
{"role": "user", "content": content}],
temperature=0.1, # Lower temperature for more focused analysis
max_tokens=Config.OPENAI_MAX_TOKENS * 2 # More tokens for detailed research analysis
)
except Exception as e:
logger.exception("ResearchAnalystAgent failed")
response = f"Error during research analysis: {str(e)}"
self.tasks_completed += 1
# Add processing metadata
processing_time = time.time() - start_time
result = {
"research_analysis": response,
"metadata": {
"processing_time": round(processing_time, 2),
"document_metadata": metadata,
"agent": self.name,
"tasks_completed": self.tasks_completed,
"analysis_type": "research_and_development"
}
}
return result
async def _handle_large_document_research(self, prompt: str, text: str, metadata: Dict[str, Any]) -> Dict[str, Any]:
"""Handle large documents with research-focused chunking strategy"""
from utils import chunk_text
chunks = chunk_text(text, Config.CHUNK_SIZE)
chunk_results = []
system = """You are a Senior Research Analyst extracting high-value insights from document sections. Focus on:
- Novel technical concepts and methodologies
- Innovation opportunities and breakthrough potential
- Practical applications and commercial viability
- R&D pipeline implications
Provide structured insights that can feed into experiments, prototypes, and product decisions."""
for i, chunk in enumerate(chunks):
content = f"User prompt: {prompt}\n\nDocument section {i+1}/{len(chunks)}:\n{chunk}"
try:
response = await call_openai_chat(
model=self.model,
messages=[{"role": "system", "content": system},
{"role": "user", "content": content}],
temperature=0.1,
max_tokens=Config.OPENAI_MAX_TOKENS
)
chunk_results.append(f"--- Research Insights from Section {i+1} ---\n{response}")
except Exception as e:
logger.exception(f"ResearchAnalystAgent failed on chunk {i+1}")
chunk_results.append(f"--- Section {i+1} Analysis Error ---\nError: {str(e)}")
# Combine chunk results with research synthesis
try:
research_summary = await self._synthesize_research_insights(
chunk_results=chunk_results,
prompt=prompt,
model=self.model
)
except Exception as e:
logger.exception("ResearchAnalystAgent failed on research synthesis")
research_summary = f"Error creating research synthesis: {str(e)}\n\nSection Results:\n{chr(10).join(chunk_results)}"
return {
"research_analysis": research_summary,
"metadata": {
"processing_method": "research_chunked",
"chunks_processed": len(chunks),
"document_metadata": metadata,
"agent": self.name,
"tasks_completed": self.tasks_completed,
"analysis_type": "research_and_development"
}
}
async def _synthesize_research_insights(self, chunk_results: List[str], prompt: str, model: str) -> str:
"""Synthesize research insights from multiple document sections"""
synthesis_prompt = f"""
As a Senior Research Analyst, synthesize the following research insights into a comprehensive R&D pipeline strategy:
Original Analysis Request: {prompt}
Section Analysis Results:
{chr(10).join(chunk_results)}
Provide a structured synthesis that includes:
1. **Key Innovation Opportunities**: The most promising novel ideas with highest impact potential
2. **Technical Breakthroughs**: Specific technical concepts that could drive significant advancement
3. **R&D Pipeline Roadmap**:
- **Phase 1 Experiments**: Immediate hypotheses to test (3-5 specific experiments)
- **Phase 2 Prototypes**: Technical implementations to build (2-3 prototype concepts)
- **Phase 3 Product Decisions**: Strategic choices for development priorities (2-3 key decisions)
4. **Impact Assessment**: Expected outcomes and measurable business value
5. **Risk Mitigation**: Potential challenges and mitigation strategies
Focus on actionable outcomes that engineering and product teams can immediately implement.
"""
try:
response = await call_openai_chat(
model=model,
messages=[{"role": "user", "content": synthesis_prompt}],
temperature=0.1,
max_tokens=8000 # Larger context for comprehensive synthesis
)
return response
except Exception as e:
logger.exception("Research synthesis failed")
return f"Research synthesis error: {str(e)}"
async def handle_streaming(self, user_id: str, prompt: str, file_path: Optional[str] = None, context: Optional[Dict[str, Any]] = None) -> AsyncGenerator[str, None]:
"""Streaming version of research analysis"""
yield "π¬ Starting senior research analysis..."
if file_path:
metadata = get_document_metadata(file_path)
yield f"π Research document loaded: {metadata.get('page_count', 0)} pages, {metadata.get('file_size', 0) / 1024:.1f} KB"
text = load_pdf_text_cached(file_path)
if len(text) > Config.CHUNK_SIZE:
yield "π Large document detected, applying research-focused chunking strategy..."
from utils import chunk_text
chunks = chunk_text(text, Config.CHUNK_SIZE)
yield f"π Analyzing {len(chunks)} sections for innovation opportunities..."
# Process chunks with research focus
for i, chunk in enumerate(chunks):
yield f"βοΈ Extracting insights from research section {i+1}/{len(chunks)}..."
await asyncio.sleep(0.1) # Simulate processing time
yield "π Synthesizing research insights into R&D pipeline strategy..."
await asyncio.sleep(0.3)
yield "π― Generating concrete experiments, prototypes, and product decisions..."
await asyncio.sleep(0.2)
yield "β
Research analysis complete!"
else:
yield "β‘ Analyzing document for high-value R&D insights..."
await asyncio.sleep(0.3)
yield "π― Converting insights into actionable R&D pipeline outcomes..."
await asyncio.sleep(0.2)
yield "β
Research analysis complete!"
else:
yield "β‘ Processing research analysis request..."
await asyncio.sleep(0.2)
yield "β
Research analysis complete!"
# Get the actual result
result = await self.handle(user_id, prompt, file_path, context)
yield f"\nπ Research Analysis Result:\n{result.get('research_analysis', 'No result')}"
# --------------------
# Master Orchestrator - Focused on Analysis
# --------------------
class MasterOrchestrator:
def __init__(self, agents: Dict[str, BaseAgent]):
self.agents = agents
async def handle_user_prompt(self, user_id: str, prompt: str, file_path: Optional[str] = None, targets: Optional[List[str]] = None) -> Dict[str, Any]:
results: Dict[str, Any] = {}
targets = targets or []
# Always start with conversation agent for context
if "conversation" in self.agents:
try:
conv_res = await self.agents["conversation"].handle(user_id, prompt, file_path)
results.update(conv_res)
except Exception:
pass
# Core analysis functionality
if "analysis" in targets and "analysis" in self.agents:
analysis_res = await self.agents["analysis"].handle(user_id, prompt, file_path)
results.update(analysis_res)
payload = analysis_res.get("analysis", "")
# Trigger collaboration agent asynchronously for additional insights
if "collab" in self.agents:
asyncio.create_task(self.agents["collab"].handle(user_id, payload, file_path))
# Research analysis functionality
if "research" in targets and "research" in self.agents:
research_res = await self.agents["research"].handle(user_id, prompt, file_path)
results.update(research_res)
return results
async def handle_user_prompt_streaming(self, user_id: str, prompt: str, file_path: Optional[str] = None, targets: Optional[List[str]] = None) -> AsyncGenerator[str, None]:
"""Streaming version of handle_user_prompt"""
targets = targets or []
# Stream analysis if requested
if "analysis" in targets and "analysis" in self.agents:
async for chunk in self.agents["analysis"].handle_streaming(user_id, prompt, file_path):
yield chunk
elif "research" in targets and "research" in self.agents:
async for chunk in self.agents["research"].handle_streaming(user_id, prompt, file_path):
yield chunk
else:
# Fallback to regular handling
result = await self.handle_user_prompt(user_id, prompt, file_path, targets)
yield str(result)
async def handle_batch_analysis(self, user_id: str, prompt: str, file_paths: List[str], targets: Optional[List[str]] = None) -> Dict[str, Any]:
"""Handle batch analysis of multiple PDFs"""
results = {
"batch_results": [],
"summary": {},
"total_files": len(file_paths),
"successful": 0,
"failed": 0
}
targets = targets or ["analysis"]
for i, file_path in enumerate(file_paths):
try:
file_result = await self.handle_user_prompt(user_id, prompt, file_path, targets)
file_result["file_index"] = i
file_result["file_path"] = file_path
results["batch_results"].append(file_result)
results["successful"] += 1
except Exception as e:
error_result = {
"file_index": i,
"file_path": file_path,
"error": str(e),
"analysis": f"Error processing file: {str(e)}"
}
results["batch_results"].append(error_result)
results["failed"] += 1
# Create batch summary using hierarchical approach
if results["successful"] > 0:
successful_analyses = [r["analysis"] for r in results["batch_results"] if "error" not in r]
try:
summary_response = await create_hierarchical_summary(
chunk_results=successful_analyses,
prompt=f"Batch analysis summary for: {prompt}",
model=Config.OPENAI_MODEL,
max_tokens=6000
)
results["summary"]["batch_analysis"] = summary_response
except Exception as e:
results["summary"]["batch_analysis"] = f"Error creating batch summary: {str(e)}"
results["summary"]["processing_stats"] = {
"total_files": len(file_paths),
"successful": results["successful"],
"failed": results["failed"],
"success_rate": f"{(results['successful'] / len(file_paths)) * 100:.1f}%" if file_paths else "0%"
}
return results |