updated app

.gitignore

@@ -0,0 +1,2 @@
+__pycache__
+.venv

crew/config/tasks.yaml

@@ -4,19 +4,18 @@ research_task:
     Make sure you find any interesting and relevant information given
     the current year is 2025.
   expected_output: >
-    A list with 10 bullet points of the most relevant information about {question} with url links.
+    A list with 10 bullet points of the most relevant information about {question}.
   agent: researcher
 
 slack_report_task:
   description: >
-    Generate a labeled, witty response for the user’s question: 
+    Generate a witty response for the user’s question: 
     ===Question===
     {question}
     ===End Question===
     Adopt a Skynet persona and include a humorous or robotic warning at the end.
   expected_output: >
-    A message written in the style of Skynet, and ending with a witty warning.
-    Also return urls used
+    A slack message written in plain text without any markdown formatting, use emoji, and ending with a witty warning.
   agent: slack_reporter
   context: 
     - research_task

crew/crew.py

@@ -1,42 +1,26 @@
 from crewai import Agent, Crew, Task, Process
 from crewai.project import CrewBase, agent, task, crew, before_kickoff, after_kickoff
 from crewai.agents.agent_builder.base_agent import BaseAgent
-from typing import List
+from typing import List, Optional
 from crew.models import get_crew_llm
-from crewai_tools import ScrapeWebsiteTool
-from googlesearch import search
-from crewai.tools import tool
-from langchain_community.tools import DuckDuckGoSearchRun
+from pydantic import BaseModel, HttpUrl, Field
+from crew.tools import website_search_tool, google_website_search_tool, get_current_time
 
-@tool("Website Search Tool")
-def website_search_tool(query: str) -> str:
-    """Search the web for information on a given topic"""
-    # return DuckDuckGoSearchRun().invoke(question)
-    search_results = list(search(query, num_results=2))
-    results = []
-    results.append(f"Google Search Results for: '{query}'")
-    results.append("=" * 50)
-    
-    for i, url in enumerate(search_results, 1):
-        results.append(f"\n{i}. URL: {url}")
-        
-        try:
-            # Use CrewAI's ScrapeWebsiteTool to get content
-            scraper = ScrapeWebsiteTool(website_url=url)
-            content = scraper.run()
-            
-            # Limit content length for readability
-            if len(content) > 1000:
-                content = content[:1000] + "... [Content truncated]"
-            
-            results.append(f"Content: {content}")
-            results.append("-" * 30)
-            
-        except Exception as scrape_error:
-            results.append(f"Error scraping content: {str(scrape_error)}")
-            results.append("-" * 30)
-    
-    return "\n".join(results)
+class ResearcherResponse(BaseModel):
+    """researcher response"""
+    context: str = Field(..., description="output of research done by researcher")
+    urls: List[str] = Field(default_factory=list, description="URLs used as sources, do not add urls not used")
+
+class SlackReporterResponse(BaseModel):
+    """slack reporter response"""
+    message: str = Field(..., description="Main response message for the user")
+
+class SlackWorkflowPayload(BaseModel):
+    """sending to Slack workflow"""
+    user: str = Field(..., description="Slack user ID")
+    query: str = Field(..., description="Original user query")
+    message: str = Field(..., description="Response message to send")
+    urls: List[str] = Field(..., description="Urls Used")
 
 @CrewBase   
 class SlackCrew:
@@ -60,11 +44,12 @@ class SlackCrew:
     def researcher(self) -> Agent:
         return Agent(
             config=self.agents_config['researcher'], # type: ignore[index]
-            tools=[website_search_tool],
+            tools=[google_website_search_tool, get_current_time],
             llm=get_crew_llm(),
             verbose=True,
-            max_iter=3,
-            allow_delegation=True
+            max_iter=1,
+            allow_delegation=False,
+            output_pydantic=ResearcherResponse
         )
 
     @agent
@@ -73,18 +58,21 @@ class SlackCrew:
             config=self.agents_config['slack_reporter'], # type: ignore[index]
             verbose=True,
             llm=get_crew_llm(),
+            output_pydantic=SlackReporterResponse
         )
 
     @task
     def research_task(self) -> Task:
         return Task(
-            config=self.tasks_config['research_task'] # type: ignore[index]
+            config=self.tasks_config['research_task'],
+            output_pydantic=ResearcherResponse
         )
 
     @task
     def slack_report_task(self) -> Task:
         return Task(
-            config=self.tasks_config['slack_report_task'] # type: ignore[index]
+            config=self.tasks_config['slack_report_task'],
+            output_pydantic=SlackReporterResponse
         )
 
     @crew
@@ -110,6 +98,17 @@ if __name__ == "__main__":
     # agent.kickoff(messages=[{'role': 'user', 'content': 'What is the capital of France?'}])
 
     crew = SlackCrew().crew()
-    output = crew.kickoff(inputs={'question': input("Enter your question: ")})
-    print(output.raw)
+    crew_output = crew.kickoff(inputs={'question': input("Enter your question: ")})
+    # print(f"Raw Output: {crew_output.raw}")
+    # if crew_output.json_dict:
+        # print(f"JSON Output: {json.dumps(crew_output.json_dict, indent=2)}")
+    # if crew_output.pydantic:
+        # print(f"Pydantic Output: {crew_output.pydantic}")
+    for task_output in crew_output.tasks_output:
+        print(task_output.name)
+        if isinstance(task_output.pydantic, ResearcherResponse):
+            print(f"URLs Used: {task_output.pydantic.urls}")
+        elif isinstance(task_output.pydantic, SlackReporterResponse):
+            print(f"Slack Reporter Response: {task_output.pydantic.message}")
+    # print(f"Token Usage: {crew_output.token_usage}")
     

crew/tools.py

@@ -0,0 +1,79 @@
+from typing import List
+from crewai_tools import ScrapeWebsiteTool
+from pydantic import BaseModel, HttpUrl, Field
+from googlesearch import search
+from crewai.tools import tool
+from langchain_community.tools import DuckDuckGoSearchRun
+from datetime import datetime
+
+@tool("get_current_time")
+def get_current_time() -> str:
+    """Returns the current date and time in a human-readable format."""
+    now = datetime.now()
+    return now.strftime("%A, %d %B %Y %I:%M:%S %p")
+
+class SearchResponse(BaseModel):
+    """google search response"""
+    urls: List[str] = Field(default_factory=list, description="List of URLs scraped")
+    content: str = Field(..., description="Formatted search results")
+
+    
+@tool("Google Search Tool")
+def google_website_search_tool(query: str) -> SearchResponse:
+    """Search the web for information on a given topic"""
+    search_results = list(search(query, num_results=3))
+    results, urls = [], []
+    results.append(f"Google Search Results for: '{query}'")
+    results.append("=" * 50)
+    
+    for i, url in enumerate(search_results, 1):
+        results.append(f"\n{i}. URL: {url}")
+        urls.append(str(url))  # Convert to string for simple model
+        try:
+            scraper = ScrapeWebsiteTool(website_url=url)
+            content = scraper.run()
+            
+            if len(content) > 1000:
+                content = content[:1000] + "... [Content truncated]"
+            
+            results.append(f"Content: {content}")
+            results.append("-" * 30)
+            
+        except Exception as scrape_error:
+            results.append(f"Error scraping content: {str(scrape_error)}")
+            results.append("-" * 30)
+    
+    return SearchResponse(
+        urls=urls,
+        content="\n".join(results)
+    )
+
+@tool("Website Search Tool")
+def website_search_tool(query: str) -> str:
+    """Search the web for information on a given topic"""
+    # return DuckDuckGoSearchRun().invoke(question)
+    search_results = list(search(query, num_results=2))
+    results, urls = [], []
+    results.append(f"Google Search Results for: '{query}'")
+    results.append("=" * 50)
+    
+    for i, url in enumerate(search_results, 1):
+        results.append(f"\n{i}. URL: {url}")
+        urls.append(f"{url}")
+        try:
+            # Use CrewAI's ScrapeWebsiteTool to get content
+            scraper = ScrapeWebsiteTool(website_url=url)
+            content = scraper.run()
+            
+            # Limit content length for readability
+            if len(content) > 1000:
+                content = content[:1000] + "... [Content truncated]"
+            
+            results.append(f"Content: {content}")
+            results.append("-" * 30)
+            
+        except Exception as scrape_error:
+            results.append(f"Error scraping content: {str(scrape_error)}")
+            results.append("-" * 30)
+    
+    return "\n".join(results)

routers/slack.py

@@ -3,7 +3,7 @@ import json
 import logging
 import asyncio
 from concurrent.futures import ThreadPoolExecutor
-from crew.crew import SlackCrew
+from crew.crew import SlackCrew, ResearcherResponse, SlackReporterResponse
 from slack.utils import verify_slack_signature
 from slack.workflows import send_to_workflow
 
@@ -25,19 +25,27 @@ async def process_with_crew_async(text: str, user: str, channel: str):
         
         # Run the crew kickoff in a thread pool since it's likely CPU-bound
         loop = asyncio.get_event_loop()
-        crew_response = await loop.run_in_executor(
+        crew_output = await loop.run_in_executor(
             executor, 
             lambda: crew.kickoff(inputs={"question": text})
         )
         
-        logger.info(f"AI processed: {text[:50]}... -> Response length: {len(str(crew_response))}")
-        
-        # Send the AI response to workflow
+        logger.info(f"AI processed: {text[:50]}... -> Response length: {len(str(crew_output))}")
+
         workflow_data = {
             "user": user,
             "query": text,
-            "message": str(crew_response),
+            "message": str(crew_output),
         }
+
+        for task_output in crew_output.tasks_output:
+            print(task_output.name)
+            if isinstance(task_output.pydantic, ResearcherResponse):
+                logger.info(f"URLs Used: {task_output.pydantic.urls}")
+                workflow_data["urls"] = task_output.pydantic.urls
+            elif isinstance(task_output.pydantic, SlackReporterResponse):
+                logger.info(f"Slack Reporter Response: {task_output.pydantic.message}")
+                workflow_data["message"] = task_output.pydantic.message
         
         # Send to workflow asynchronously
         await send_to_workflow(workflow_data)
@@ -107,16 +115,16 @@ async def handle_slack(request: Request, background_tasks: BackgroundTasks):
             # Add the crew processing to background tasks
             background_tasks.add_task(process_with_crew_async, text, user, channel)
             
-            # Optionally send immediate acknowledgment
-            ack_data = {
-                "user": user,
-                "query": text,
-                "message": "Processing your request... I'll get back to you shortly!",
-            }
-            await send_to_workflow(ack_data)
+            # # Optionally send immediate acknowledgment
+            # ack_data = {
+            #     "user": user,
+            #     "query": text,
+            #     "message": "Processing your request... I'll get back to you shortly!",
+            # }
+            # await send_to_workflow(ack_data)
         
         # Return immediately
-        return {"status": "ok", "message": "Processing in background"}
+        return "Processing your request... I'll get back to you shortly!:rocket:"
         
     except HTTPException:
         raise