File size: 25,542 Bytes
26fe3e3 ab78f68 26fe3e3 ab78f68 26fe3e3 ab78f68 26fe3e3 ab78f68 26fe3e3 ab78f68 26fe3e3 ab78f68 26fe3e3 ab78f68 26fe3e3 ab78f68 26fe3e3 ab78f68 26fe3e3 ab78f68 26fe3e3 ab78f68 26fe3e3 ab78f68 26fe3e3 ab78f68 26fe3e3 ab78f68 26fe3e3 ab78f68 26fe3e3 ab78f68 26fe3e3 ab78f68 26fe3e3 ab78f68 26fe3e3 ab78f68 26fe3e3 ab78f68 26fe3e3 ab78f68 26fe3e3 5c67c52 26fe3e3 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 |
from agents import Runner, trace, gen_trace_id, Agent, function_tool
from search_agent import search_agent, create_search_agent
from planner_agent import planner_agent, create_planner_agent, WebSearchItem, WebSearchPlan
from writer_agent import writer_agent, create_writer_agent, ReportData
from email_agent import email_agent, create_email_agent
from clarifier_agent import clarifier_agent, create_clarifier_agent, ClarificationData
from evaluator_agent import evaluator_agent, optimizer_agent, create_evaluator_agent, create_optimizer_agent, EvaluationResult, OptimizedReport
import asyncio
from typing import Dict, Any, AsyncGenerator
# Legacy ResearchManager class for backward compatibility
class ResearchManager:
def __init__(self, api_key: str = None, model: str = "gpt-4o-mini"):
"""Initialize ResearchManager with optional API key and model"""
self.api_key = api_key
self.model = model
# Set the API key in environment if provided
if api_key:
import os
os.environ["OPENAI_API_KEY"] = api_key
async def run_with_clarification(self, query: str):
""" Run the clarification step and return clarifying questions """
trace_id = gen_trace_id()
with trace("Clarification trace", trace_id=trace_id):
print(f"View trace: https://platform.openai.com/traces/trace?trace_id={trace_id}")
print("Generating clarifying questions...")
clarifier = create_clarifier_agent(self.model)
result = await Runner.run(
clarifier,
f"Query: {query}",
)
clarification_data = result.final_output_as(ClarificationData)
print(f"Generated {len(clarification_data.questions)} clarifying questions")
return {
"questions": clarification_data.questions,
"trace_id": trace_id
}
async def run_research_with_answers(self, query: str, answers: list[str]):
""" Run the full research process with clarification answers """
trace_id = gen_trace_id()
with trace("Research with clarification trace", trace_id=trace_id):
print(f"View trace: https://platform.openai.com/traces/trace?trace_id={trace_id}")
print("Starting research with clarifications...")
# Use the new manager agent instead
clarified_query = self._format_clarified_query(query, answers)
result = await Runner.run(
ResearchManagerAgent,
f"Research Query: {clarified_query}",
)
return {
"report": result.final_output,
"trace_id": trace_id
}
def _format_clarified_query(self, original_query: str, answers: list[str]) -> str:
""" Format the original query with clarification answers """
clarifications = []
for i, answer in enumerate(answers, 1):
if answer.strip():
clarifications.append(f"{i}. {answer.strip()}")
if clarifications:
clarified_query = f"""Original query: {original_query}
Clarifications provided:
{chr(10).join(clarifications)}
Please use these clarifications to focus and refine the research approach."""
else:
clarified_query = original_query
return clarified_query
async def run(self, query: str):
""" Run the deep research process, yielding the status updates and the final report"""
trace_id = gen_trace_id()
with trace("Research trace", trace_id=trace_id):
print(f"View trace: https://platform.openai.com/traces/trace?trace_id={trace_id}")
yield f"View trace: https://platform.openai.com/traces/trace?trace_id={trace_id}"
print("Starting research...")
# Use the new manager agent
result = await Runner.run(ResearchManagerAgent, f"Research Query: {query}")
yield "Research complete"
yield result.final_output
# Global variable to store current model
_current_model = "gpt-4o-mini"
def set_current_model(model: str):
"""Set the current model for function tools"""
global _current_model
_current_model = model
# Function tools for the manager agent to orchestrate the research process
@function_tool
async def plan_research(query: str) -> Dict[str, Any]:
""" Plan the research searches for a given query """
print("Planning searches...")
planner = create_planner_agent(_current_model)
result = await Runner.run(planner, f"Query: {query}")
search_plan = result.final_output_as(WebSearchPlan)
print(f"Will perform {len(search_plan.searches)} searches")
return {
"searches": [{"query": item.query, "reason": item.reason} for item in search_plan.searches],
"plan_ready": True
}
@function_tool
async def perform_search(search_query: str, reason: str) -> str:
""" Perform a single web search and return summarized results """
print(f"Searching for: {search_query}")
input_text = f"Search term: {search_query}\nReason for searching: {reason}"
try:
searcher = create_search_agent(_current_model)
result = await Runner.run(searcher, input_text)
return str(result.final_output)
except Exception as e:
print(f"Search failed for '{search_query}': {e}")
return f"Search failed for '{search_query}': {str(e)}"
@function_tool
async def write_initial_report(query: str, search_results: str) -> Dict[str, Any]:
""" Generate an initial research report from search results """
print("Writing initial report...")
input_text = f"Original query: {query}\nSummarized search results: {search_results}"
writer = create_writer_agent(_current_model)
result = await Runner.run(writer, input_text)
report_data = result.final_output_as(ReportData)
print("Initial report completed")
return {
"markdown_report": report_data.markdown_report,
"short_summary": report_data.short_summary,
"follow_up_questions": report_data.follow_up_questions
}
@function_tool
async def evaluate_report(query: str, report: str) -> Dict[str, Any]:
""" Evaluate the quality of a research report """
print("Evaluating report quality...")
input_text = f"Original Query: {query}\n\nReport to Evaluate:\n{report}"
evaluator = create_evaluator_agent(_current_model)
result = await Runner.run(evaluator, input_text)
evaluation = result.final_output_as(EvaluationResult)
print(f"Evaluation complete - Score: {evaluation.overall_score}/10, Needs refinement: {evaluation.needs_refinement}")
return {
"score": evaluation.overall_score,
"strengths": evaluation.strengths,
"weaknesses": evaluation.weaknesses,
"suggestions": evaluation.suggestions,
"needs_refinement": evaluation.needs_refinement,
"refinement_requirements": evaluation.refined_requirements
}
@function_tool
async def optimize_report(query: str, original_report: str, evaluation_feedback: str) -> str:
""" Optimize and improve a research report based on evaluation feedback """
print("Optimizing report...")
input_text = f"""Original Query: {query}
Original Report:
{original_report}
Evaluation Feedback:
{evaluation_feedback}
Please improve the report based on this feedback."""
optimizer = create_optimizer_agent(_current_model)
result = await Runner.run(optimizer, input_text)
optimized = result.final_output_as(OptimizedReport)
print("Report optimization complete")
return optimized.improved_markdown_report
# Regular function that can be called directly
async def _send_report_email_to_address(report: str, recipient_email: str) -> Dict[str, str]:
""" Send the final research report via email to a specific address """
import os
import sendgrid
from sendgrid.helpers.mail import Email, Mail, Content, To
print(f"Sending email to: {recipient_email}")
try:
sg = sendgrid.SendGridAPIClient(api_key=os.environ.get('SENDGRID_API_KEY'))
from_email = Email("mantomarchi300@outlook.com") # Verified sender
to_email = To(recipient_email) # User-provided email
# Create a nice subject line
subject = "π Your Research Report is Ready"
# Convert markdown to HTML for better email formatting
import re
# Basic markdown to HTML conversion
html_report = report
# Convert markdown links to HTML links with styling
html_report = re.sub(r'\[([^\]]+)\]\(([^)]+)\)', r'<a href="\2" style="color: #2563eb; text-decoration: none; border-bottom: 1px solid #2563eb;" target="_blank">\1</a>', html_report)
# Convert headers
html_report = re.sub(r'^### (.*$)', r'<h3 style="color: #2563eb; margin-top: 25px; margin-bottom: 10px;">\1</h3>', html_report, flags=re.MULTILINE)
html_report = re.sub(r'^## (.*$)', r'<h2 style="color: #1d4ed8; margin-top: 30px; margin-bottom: 15px;">\1</h2>', html_report, flags=re.MULTILINE)
html_report = re.sub(r'^# (.*$)', r'<h1 style="color: #1e40af; margin-top: 35px; margin-bottom: 20px;">\1</h1>', html_report, flags=re.MULTILINE)
# Convert bold text
html_report = re.sub(r'\*\*(.*?)\*\*', r'<strong style="color: #374151;">\1</strong>', html_report)
# Convert numbered lists (for sources)
html_report = re.sub(r'^(\d+\.\s)(.*$)', r'<li style="margin-bottom: 8px; list-style-type: decimal;">\2</li>', html_report, flags=re.MULTILINE)
# Convert bullet points
html_report = re.sub(r'^- (.*$)', r'<li style="margin-bottom: 8px;">\1</li>', html_report, flags=re.MULTILINE)
# Wrap consecutive list items in ul/ol tags
html_report = re.sub(r'(<li style="margin-bottom: 8px; list-style-type: decimal;">.*?</li>)', r'<ol style="margin: 15px 0; padding-left: 25px;">\1</ol>', html_report, flags=re.DOTALL)
html_report = re.sub(r'(<li style="margin-bottom: 8px;">.*?</li>)', r'<ul style="margin: 15px 0; padding-left: 25px;">\1</ul>', html_report, flags=re.DOTALL)
# Convert line breaks
html_report = html_report.replace('\n\n', '</p><p style="margin-bottom: 15px; line-height: 1.6;">')
html_report = '<p style="margin-bottom: 15px; line-height: 1.6;">' + html_report + '</p>'
html_content = f"""
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>Your Research Report</title>
</head>
<body style="font-family: 'Segoe UI', Tahoma, Geneva, Verdana, sans-serif; line-height: 1.6; color: #374151; background-color: #f9fafb; margin: 0; padding: 20px;">
<div style="max-width: 800px; margin: 0 auto; background: #ffffff; border-radius: 12px; box-shadow: 0 4px 6px rgba(0, 0, 0, 0.07); overflow: hidden;">
<!-- Header -->
<div style="background: linear-gradient(135deg, #2563eb 0%, #1d4ed8 100%); color: white; padding: 30px; text-align: center;">
<h1 style="margin: 0; font-size: 28px; font-weight: 600;">
π Your Research Report
</h1>
<p style="margin: 10px 0 0 0; opacity: 0.9; font-size: 16px;">
Comprehensive AI-powered research analysis
</p>
</div>
<!-- Content -->
<div style="padding: 40px 30px;">
<div style="background: #f8fafc; padding: 30px; border-radius: 8px; border-left: 4px solid #2563eb; margin-bottom: 30px;">
{html_report}
</div>
</div>
<!-- Footer -->
<div style="background: #f8fafc; padding: 25px 30px; border-top: 1px solid #e5e7eb;">
<div style="text-align: center; color: #6b7280; font-size: 14px;">
<p style="margin: 0 0 10px 0;">
<strong>π€ Generated by Deep Research Assistant</strong>
</p>
<p style="margin: 0;">
This report was created using advanced AI with multi-step quality assurance
</p>
<div style="margin-top: 15px; padding-top: 15px; border-top: 1px solid #d1d5db;">
<p style="margin: 0; font-size: 12px; color: #9ca3af;">
Thank you for using our research service β’ Generated on {__import__('datetime').datetime.now().strftime('%B %d, %Y at %I:%M %p')}
</p>
</div>
</div>
</div>
</div>
</body>
</html>
"""
content = Content("text/html", html_content)
mail = Mail(from_email, to_email, subject, content).get()
response = sg.client.mail.send.post(request_body=mail)
print(f"Email response: {response.status_code}")
if response.status_code == 202:
return {"status": f"Email sent successfully to {recipient_email}"}
else:
return {"status": f"Email sending failed with status {response.status_code}"}
except Exception as e:
print(f"Email sending error: {e}")
return {"status": f"Email sending failed: {str(e)}"}
@function_tool
async def send_report_email_to_address(report: str, recipient_email: str) -> Dict[str, str]:
""" Send the final research report via email to a specific address """
return await _send_report_email_to_address(report, recipient_email)
@function_tool
async def send_report_email(report: str) -> Dict[str, str]:
""" Send the final research report via email (legacy function - uses hardcoded email) """
print("Sending email to default address...")
result = await Runner.run(email_agent, report)
print("Email sent to default address")
return {"status": "Email sent successfully to default address"}
# Manager Agent Instructions
MANAGER_INSTRUCTIONS = """
You are the Research Manager Agent responsible for orchestrating a comprehensive research process with quality assurance.
Your workflow:
1. **PLAN**: Use plan_research to create a search strategy for the query
2. **SEARCH**: Use perform_search for each planned search item to gather information
3. **INITIAL REPORT**: Use write_initial_report to create a first draft from all search results
4. **EVALUATE**: Use evaluate_report to assess the quality of the initial report
5. **OPTIMIZE** (if needed): If evaluation shows needs_refinement=True, use optimize_report to improve it
6. **FINALIZE**: Use send_report_email_to_address to deliver the final report
Quality Standards:
- Only proceed to email if the report scores 7+ or has been optimized
- If a report needs refinement, always optimize it before sending
- Ensure comprehensive coverage of the original query
- Maintain high standards for accuracy and completeness
Be methodical and ensure each step completes successfully before proceeding to the next.
"""
# Function to create custom research agent with email options
def create_custom_research_agent(email_address: str = None, send_email: bool = False, api_key: str = None, model: str = "gpt-4o-mini"):
"""Create a research manager agent with custom email settings"""
# Set API key in environment if provided
if api_key:
import os
os.environ["OPENAI_API_KEY"] = api_key
# Set the current model for all function tools
set_current_model(model)
if send_email and email_address:
# Include email sending in tools
tools = [
plan_research,
perform_search,
write_initial_report,
evaluate_report,
optimize_report,
send_report_email_to_address
]
instructions = f"""
You are the Research Manager Agent responsible for orchestrating a comprehensive research process with quality assurance.
Your workflow:
1. **PLAN**: Use plan_research to create a search strategy for the query
2. **SEARCH**: Use perform_search for each planned search item to gather information
3. **INITIAL REPORT**: Use write_initial_report to create a first draft from all search results
4. **EVALUATE**: Use evaluate_report to assess the quality of the initial report
5. **OPTIMIZE** (if needed): If evaluation shows needs_refinement=True, use optimize_report to improve it
6. **FINALIZE**: Use send_report_email_to_address with the report and recipient email "{email_address}" to deliver the final report
Quality Standards:
- Only proceed to email if the report scores 7+ or has been optimized
- If a report needs refinement, always optimize it before sending
- Ensure comprehensive coverage of the original query
- Maintain high standards for accuracy and completeness
IMPORTANT: When using send_report_email_to_address, you must provide both:
- The final report text as the first parameter
- The recipient email address "{email_address}" as the second parameter
Be methodical and ensure each step completes successfully before proceeding to the next.
The user has requested the report be emailed to: {email_address}
"""
else:
# Exclude email sending from tools
tools = [
plan_research,
perform_search,
write_initial_report,
evaluate_report,
optimize_report
]
instructions = """
You are the Research Manager Agent responsible for orchestrating a comprehensive research process with quality assurance.
Your workflow:
1. **PLAN**: Use plan_research to create a search strategy for the query
2. **SEARCH**: Use perform_search for each planned search item to gather information
3. **INITIAL REPORT**: Use write_initial_report to create a first draft from all search results
4. **EVALUATE**: Use evaluate_report to assess the quality of the initial report
5. **OPTIMIZE** (if needed): If evaluation shows needs_refinement=True, use optimize_report to improve it
6. **COMPLETE**: Return the final optimized report (do NOT send email - user chose not to receive email)
Quality Standards:
- Complete the process when report scores 7+ or has been optimized
- If a report needs refinement, always optimize it before completing
- Ensure comprehensive coverage of the original query
- Maintain high standards for accuracy and completeness
Be methodical and ensure each step completes successfully before proceeding to the next.
The user has chosen NOT to receive the report via email.
"""
return Agent(
name=f"Custom Research Manager Agent",
instructions=instructions,
tools=tools,
model=model,
handoff_description="Orchestrate comprehensive research with quality assurance and optional email delivery"
)
# Create the Research Manager Agent with agents-as-tools
ResearchManagerAgent = Agent(
name="Research Manager Agent",
instructions=MANAGER_INSTRUCTIONS,
tools=[
plan_research,
perform_search,
write_initial_report,
evaluate_report,
optimize_report,
send_report_email_to_address
],
model="gpt-4o-mini",
handoff_description="Orchestrate comprehensive research with quality assurance and optimization"
)
async def run_research_with_progress(query: str, email_address: str = None, send_email: bool = False, api_key: str = None, model: str = "gpt-4o-mini") -> AsyncGenerator[str, None]:
"""Run research with step-by-step progress updates"""
# Set API key if provided
if api_key:
import os
os.environ["OPENAI_API_KEY"] = api_key
# Set current model for function tools
set_current_model(model)
trace_id = gen_trace_id()
yield f"π **Starting Enhanced Research**\n\n**Query:** {query}\n\n**Trace ID:** {trace_id}\n\n**Model:** {model}\n\n---\n\n"
try:
with trace("Enhanced Research with Progress", trace_id=trace_id):
# Step 1: Planning
yield "π **Step 1/6:** Planning research strategy...\n\n*Analyzing your query and determining the best search approach*"
planner = create_planner_agent(model)
result = await Runner.run(planner, f"Query: {query}")
search_plan = result.final_output_as(WebSearchPlan)
yield f"β
**Planning Complete** - Will perform {len(search_plan.searches)} targeted searches\n\n---\n\n"
# Step 2: Searching
yield "π **Step 2/6:** Conducting web searches...\n\n*Gathering information from multiple sources*"
search_results = []
for i, search_item in enumerate(search_plan.searches, 1):
yield f"π **Search {i}/{len(search_plan.searches)}:** {search_item.query}\n\n*{search_item.reason}*"
try:
input_text = f"Search term: {search_item.query}\nReason for searching: {search_item.reason}"
searcher = create_search_agent(model)
result = await Runner.run(searcher, input_text)
search_results.append(str(result.final_output))
yield f"β
**Search {i} Complete**\n\n"
except Exception as e:
yield f"β οΈ **Search {i} Failed:** {str(e)}\n\n"
search_results.append(f"Search failed: {str(e)}")
yield "β
**All Searches Complete**\n\n---\n\n"
# Step 3: Writing Initial Report
yield "βοΈ **Step 3/6:** Writing initial research report...\n\n*Analyzing and synthesizing all gathered information*"
combined_results = "\n\n".join(search_results)
input_text = f"Original query: {query}\nSummarized search results: {combined_results}"
writer = create_writer_agent(model)
result = await Runner.run(writer, input_text)
report_data = result.final_output_as(ReportData)
yield "β
**Initial Report Complete**\n\n---\n\n"
# Step 4: Evaluating Quality
yield "π **Step 4/6:** Evaluating report quality...\n\n*AI quality assessment in progress*"
input_text = f"Original Query: {query}\n\nReport to Evaluate:\n{report_data.markdown_report}"
evaluator = create_evaluator_agent(model)
result = await Runner.run(evaluator, input_text)
evaluation = result.final_output_as(EvaluationResult)
yield f"β
**Evaluation Complete** - Score: {evaluation.overall_score}/10\n\n"
final_report = report_data.markdown_report
# Step 5: Optimization (if needed)
if evaluation.needs_refinement:
yield "π§ **Step 5/6:** Optimizing report quality...\n\n*Improving report based on evaluation feedback*"
feedback = f"Score: {evaluation.overall_score}/10\nWeaknesses: {evaluation.weaknesses}\nSuggestions: {evaluation.suggestions}"
input_text = f"""Original Query: {query}
Original Report:
{report_data.markdown_report}
Evaluation Feedback:
{feedback}
Please improve the report based on this feedback."""
optimizer = create_optimizer_agent(model)
result = await Runner.run(optimizer, input_text)
optimized = result.final_output_as(OptimizedReport)
final_report = optimized.improved_markdown_report
yield "β
**Optimization Complete** - Report quality improved\n\n---\n\n"
else:
yield "β
**No Optimization Needed** - Report quality is excellent\n\n---\n\n"
# Step 6: Email Delivery (if requested)
if send_email and email_address:
yield f"π§ **Step 6/6:** Sending report to {email_address}...\n\n*Preparing and delivering your research report*"
try:
# Call the regular function directly
email_result = await _send_report_email_to_address(final_report, email_address)
email_status = email_result.get("status", "Email status unknown")
yield f"β
**Email Delivery:** {email_status}\n\n---\n\n"
except Exception as e:
yield f"β **Email Failed:** {str(e)}\n\n---\n\n"
else:
yield "π **Step 6/6:** Finalizing report...\n\n*Email delivery not requested*\n\n---\n\n"
# Final result
yield f"""π **Research Complete!**
**π Final Report:**
{final_report}
**π View Full Trace:** https://platform.openai.com/traces/trace?trace_id={trace_id}
---
*Enhanced research completed with quality assurance*"""
except Exception as e:
yield f"β **Error during research:** {str(e)}" |