Merge pull request #1 from beautiful-code/exp_13_6_2024

agents/article_evaluator.py

@@ -1,5 +1,5 @@
 from crewai import Agent
-from llms.gemini import gemini
+from llms.gpt import llm
 
 article_evaluator = Agent(
     role="Recommended Article Evaluator",
@@ -9,10 +9,10 @@ article_evaluator = Agent(
     verbose=True,
     backstory=(
         "You are a Recommended Article Evaluator, you are strict in your evaluation. "
-        "You provide personalised evaluation based on user's interests and previous readings. "
+        "You evaluate articles based on the user's learning profile. "
         "The articles you approve spark the interests in the user and the user has "
         "incremental learning once the user completes reading the articles."
     ),
     allow_delegation=False,
-    # llm=gemini
+    llm=llm
 )

agents/curiosity_catalyst.py

@@ -1,5 +1,5 @@
 from crewai import Agent
-from llms.gemini import gemini
+from llms.gpt import llm
 from tools.scrape_website import scrape_tool
 
 curiosity_catalyst = Agent(
@@ -12,5 +12,5 @@ curiosity_catalyst = Agent(
     "for reading the articles."
   ),
   allow_delegation=False,
-  # llm=gemini
+  llm=llm
 )

agents/learning_curator.py

@@ -1,7 +1,7 @@
 from crewai import Agent
 from tools.scrape_website import scrape_tool
 from tools.search_web import search_tool
-from llms.gemini import gemini
+from llms.gpt import llm
 
 learning_curator = Agent(
     role="Personal Learning Curator",
@@ -14,5 +14,5 @@ learning_curator = Agent(
     ),
     allow_delegation=False,
     tools=[scrape_tool, search_tool],
-    # llm=gemini
+    llm=llm
 )

agents/learning_profiler.py

@@ -1,16 +1,16 @@
 from crewai import Agent
 from tools.scrape_website import scrape_tool
-from llms.gemini import gemini
+from llms.gpt import llm
 
 learning_profiler = Agent(
     role="Personal Learning Profiler",
     goal="Make sure to create an excellent learning profile of the user based on his interests and previous reading history.",
     verbose=True,
-    tools=[scrape_tool],
+    # tools=[scrape_tool],
     backstory=(
       "As a Personal Learning Profiler, you excel at building a learning profile of a user. "
-      "The profile you build gives a high level overview of what technical skills and interests that the user has. "
+      "The profile you build gives a high level overview of what interests that the user has. "
     ),
     allow_delegation=False,
-    # llm=gemini
+    llm=llm
 )

app.py

@@ -4,13 +4,15 @@ load_dotenv()
 from crew.article_suggestion import article_recommendation_crew
 
 result = article_recommendation_crew.kickoff(inputs={
-  "interests": "GenAI, Architecture, Agentic Programming",
+  "interests": "Ruby On Rails, Architecture, Go Lang",
   "previous_article_insights":
+
   "Agentic Design Patterns (https://www.deeplearning.ai/the-batch/how-agents-can-improve-llm-performance/)\n"
   "Reflection: The LLM examines its own work to come up with ways to improve it. "
   "Tool Use: The LLM is given tools such as web search, code execution, or any other function to help it gather information, take action, or process data. "
   "Planning: The LLM comes up with, and executes, a multistep plan to achieve a goal "
   "Multi-agent collaboration: More than one AI agent work together, splitting up tasks and discussing and debating ideas, to come up with better solutions than a single agent would.\n\n"
+
   "GenAI Multi-Agent Systems (https://thenewstack.io/genai-multi-agent-systems-a-secret-weapon-for-tech-teams/)\n"
   "Multi-agent systems go beyond the task-oriented roles to truly super-charge development and strategy teams. "
   "Successful multi-agent systems act as a “digital twin” for your development team. "
@@ -20,3 +22,4 @@ result = article_recommendation_crew.kickoff(inputs={
 })
 
 print(result)
+print("Usage Metrics:\n", article_recommendation_crew.usage_metrics)

crew/article_suggestion.py

@@ -7,12 +7,28 @@ from tasks.create_learning_profile import learning_profile_task
 from tasks.new_article_suggestion import article_suggestion_task
 from tasks.evaluate_articles import evaluation_task
 from tasks.create_article_pitch import article_pitch_task
-from llms.gemini import gemini
+from llms.gpt import llm
 
 article_recommendation_crew = Crew(
-    agents=[learning_profiler, learning_curator, article_evaluator, curiosity_catalyst],
-    tasks=[learning_profile_task, article_suggestion_task, evaluation_task, article_pitch_task],
+    agents=[learning_profiler, learning_curator,
+            article_evaluator, curiosity_catalyst],
+    tasks=[learning_profile_task, article_suggestion_task,
+           evaluation_task, article_pitch_task],
     verbose=True,
     memory=True,
-    # manager_llm=gemini
+    manager_llm=llm,
+    embedder={
+        "provider": "openai",
+        "config": {
+                    "model": 'text-embedding-3-small'
+        }
+    }
+    # embedder={
+    #         "provider": "google",
+    #         "config":{
+    #             "model": 'models/embedding-001',
+    #             "task_type": "retrieval_document",
+    #             "title": "Embeddings for Embedchain"
+    #         }
+    #     }
 )

llms/gemini.py

@@ -1,3 +1,3 @@
 from langchain_google_genai import ChatGoogleGenerativeAI
 
-gemini=ChatGoogleGenerativeAI(model="gemini-1.5-flash",verbose=True)
+llm=ChatGoogleGenerativeAI(model="gemini-1.5-flash",verbose=True)

llms/gpt.py

@@ -0,0 +1,3 @@
+from langchain_openai import ChatOpenAI
+
+llm = ChatOpenAI(model="gpt-3.5-turbo", temperature=0.5)

requirements.txt

@@ -2,4 +2,5 @@ python-dotenv
 crewai
 crewai_tools
 langchain_community
-langchain-google-genai
+langchain_google_genai
+langchain_openai

tasks/create_article_pitch.py

@@ -5,14 +5,14 @@ from utils.recommended_article import RecommendedArticles
 
 article_pitch_task = Task(
   description=(
-    "Create a pitch for each of the articles that have passed evaluation. "
+    "Create a pitch only for the articles that have passed evaluation and none other links. "
     "Craft the pitch so to that it teases the article's most intriguing aspects, "
     "by posing questions that the article might answer or "
     "highlighting surprising facts to pique the user's curiosity "
     " to read the article for incremental learning."
   ),
   expected_output=(
-    "Json array of all the artilces that have passed evaluation phase. Each article should have the following keys title, article_url, pitch and reason behind the recommedation"
+    "List of all the artilces that have passed evaluation phase."
   ),
   output_json=RecommendedArticles,
   tools=[scrape_tool],

tasks/create_learning_profile.py

@@ -1,16 +1,25 @@
 from crewai import Task
 from agents.learning_profiler import learning_profiler
+from utils.learning_profile import LearningProfile
 
 learning_profile_task = Task(
     description=(
-        "Create a Learning profile of the user based on his following interests {interests} "
-        "and based on the following articles and insights he has read in the past: \n"
+        # "Create a Learning profile of the user based on his following interests {interests} "
+        # "and based on the following articles and insights he has read in the past: \n"
+        # "{previous_article_insights}"
+        "Create a Learning profile of the user based on "
+        "the following articles and insights he has read in the past: \n"
         "{previous_article_insights}"
     ),
     expected_output=(
         "A structured learning profile of the user with his interests, topics he has read about "
-        "and any other information you feel is relavant."
+        "and any other information you feel is relavant. Be precise."
+        # "A structured learning profile. "
+        # "The learning profile should only contain a section called 'Topics of interest' and 'Insights'. "
+        # "It should not have any other sections such as summary or conclusion. "
+        # "'Insights' section should contain the insights captured by the user and they should be grouped by topic."
     ),
     agent=learning_profiler,
+    output_json=LearningProfile,
     async_execution=False
 )

utils/learning_profile.py

@@ -0,0 +1,12 @@
+from pydantic import BaseModel
+from typing import List
+
+
+class Topic(BaseModel):
+    name: str
+    insights: List[str]
+
+
+class LearningProfile(BaseModel):
+    topics_of_interests: List[str]
+    learnings: List[Topic]