Create app.py

app.py

@@ -0,0 +1,185 @@
+import streamlit as st
+
+
+st.set_page_config(page_title="AI Data Analysis & Chatbot", layout="wide")
+
+import pandas as pd
+import numpy as np
+import matplotlib.pyplot as plt
+import seaborn as sns
+import networkx as nx
+from pycaret.classification import setup as setup_clf, compare_models as compare_clf
+from pycaret.regression import setup as setup_reg, compare_models as compare_reg
+from sklearn.linear_model import LinearRegression
+from sklearn.cluster import KMeans
+from langchain_openai import ChatOpenAI
+from langchain_experimental.agents.agent_toolkits import create_pandas_dataframe_agent
+from langchain.agents.agent_types import AgentType
+from langchain.schema import HumanMessage, AIMessage
+from fuzzywuzzy import process  # For fuzzy name matching
+from crewai import Crew, Agent, Task
+import os
+import threading
+from dotenv import load_dotenv
+
+# Load environment variables
+load_dotenv()
+openai_api_key = os.getenv("OPENAI_API_KEY")
+if not openai_api_key:
+    st.error("❌ OpenAI API Key is missing! Set it in your .env file.")
+
+# Title
+st.title("📊 AI-Powered Data Analysis & Chatbot")
+
+# 📂 File Upload Section
+uploaded_file = st.file_uploader("📂 Upload your CSV or Excel file", type=["csv", "xlsx"], key="file_uploader_1")
+
+if uploaded_file is not None:
+    # Read File
+    file_extension = uploaded_file.name.split(".")[-1]
+    if file_extension == "csv":
+        df = pd.read_csv(uploaded_file)
+    elif file_extension == "xlsx":
+        df = pd.read_excel(uploaded_file, engine="openpyxl")
+
+    # Display the dataset
+    st.subheader("📋 Data Preview")
+    st.dataframe(df.head())
+
+    # ✅ CrewAI Agents & Tasks
+    cleaning_agent = Agent(
+        role="Data Cleaning Agent",
+        goal="Clean and preprocess dataset.",
+        backstory="An expert data scientist specializing in handling missing values, cleaning data, and ensuring data quality.",
+        llm=ChatOpenAI(model="gpt-3.5-turbo-0125"),
+        allow_delegation=True
+    )
+
+    analysis_agent = Agent(
+        role="Data Analysis Expert",
+        goal="Perform exploratory data analysis and generate insights.",
+        backstory="An AI analyst skilled in identifying patterns, correlations, and visualizing trends.",
+        llm=ChatOpenAI(model="gpt-3.5-turbo-0125"),
+        allow_delegation=True
+    )
+
+    prediction_agent = Agent(
+        role="Prediction Specialist",
+        goal="Run AutoML to identify the best model for predicting future trends.",
+        backstory="An AI-powered machine learning specialist that automates model selection and predictions.",
+        llm=ChatOpenAI(model="gpt-3.5-turbo-0125"),
+        allow_delegation=True
+    )
+
+    # Tasks
+    cleaning_task = Task(
+        description="Clean dataset and handle missing values.",
+        agent=cleaning_agent,
+        expected_output="A cleaned dataset with missing values handled."
+    )
+
+    analysis_task = Task(
+        description="Analyze dataset, detect trends, and generate visualizations.",
+        agent=analysis_agent,
+        expected_output="A report summarizing key trends, correlations, and insights."
+    )
+
+    prediction_task = Task(
+        description="Run AutoML to identify the best predictive model.",
+        agent=prediction_agent,
+        expected_output="A machine learning model recommendation with performance metrics."
+    )
+
+    # CrewAI Execution
+    crew = Crew(
+        agents=[cleaning_agent, analysis_agent, prediction_agent],
+        tasks=[cleaning_task, analysis_task, prediction_task]
+    )
+
+    # Run CrewAI in a separate thread to prevent blocking
+    thread = threading.Thread(target=crew.kickoff, daemon=True)
+    thread.start()
+
+    # ✅ AI Chatbot Agent with Name Matching
+    pandas_agent = create_pandas_dataframe_agent(
+        ChatOpenAI(model="gpt-3.5-turbo-0125"),
+        df,
+        verbose=True,
+        agent_type=AgentType.OPENAI_FUNCTIONS,
+        allow_dangerous_code=True,
+    )
+
+    def find_closest_match(user_input, column_name):
+        """Find the closest matching value in a specific column using fuzzy matching."""
+        if column_name in df.columns:
+            choices = df[column_name].dropna().unique().tolist()
+            best_match, score = process.extractOne(user_input, choices)
+            return best_match if score > 80 else None  # Only return if confidence is high
+        return None
+
+    def process_ai_response(user_input):
+        """AI response processing with fuzzy name matching."""
+        for column in df.columns:  # Search for closest match in all columns
+            closest_match = find_closest_match(user_input, column)
+            if closest_match:
+                st.info(f"🔍 Did you mean **{closest_match}**?")
+                user_input = user_input.replace(user_input, closest_match)
+
+        try:
+            ai_response = pandas_agent.invoke(user_input)
+            graph_result = render_ai_graph(ai_response)
+
+            if graph_result:
+                ai_response = graph_result
+            elif isinstance(ai_response, dict) and "plot" in ai_response:
+                fig = ai_response["plot"]
+                st.pyplot(fig)
+            else:
+                if isinstance(ai_response, dict):
+                    ai_response = str(ai_response)
+                st.markdown(ai_response)
+
+            return ai_response
+        except Exception as e:
+            return f"❌ Error processing request: {str(e)}"
+
+    def render_ai_graph(ai_response):
+        """Detects AI-generated graph code and executes it properly inside Streamlit."""
+        if isinstance(ai_response, dict) and "graph_code" in ai_response:
+            graph_code = ai_response["graph_code"]
+
+            try:
+                exec(graph_code, globals(), locals())  # Execute the AI-generated graph
+                fig = plt.gcf()
+
+                if fig and fig.get_axes():
+                    st.pyplot(fig)
+                    return "📊 AI-Generated Graph Displayed Successfully!"
+                else:
+                    return "⚠ AI generated an empty graph, please refine your request."
+            except Exception as e:
+                return f"❌ Error executing AI graph: {str(e)}"
+
+        return None  # If no graph was generated, return None
+
+st.subheader("💬 Chat with AI")
+if "chat_history" not in st.session_state:
+    st.session_state.chat_history = []
+
+for msg in st.session_state.chat_history:
+    if isinstance(msg, HumanMessage):
+        st.markdown(f"**You:** {msg.content}")
+    elif isinstance(msg, AIMessage):
+        st.markdown(f"🤖 **AI:** {msg.content}")
+
+# ✅ User Input for Chatbot
+user_input = st.text_input("Ask AI About Your Data:")
+
+if user_input:
+    st.session_state.chat_history.append(HumanMessage(content=user_input))
+    ai_response = process_ai_response(user_input)
+    st.session_state.chat_history.append(AIMessage(content=ai_response))
+
+# Footer
+st.markdown("---")
+st.markdown("💡 Built with ❤️ using **Streamlit, CrewAI, LangChain, PyCaret & GPT-3.5-Turbo-0125**")