Upload 13 files

.env

@@ -0,0 +1,3 @@
+LITELLM_PROVIDER=ollama
+OLLAMA_MODEL=llama2
+OLLAMA_BASE_URL=http://localhost:11434

app.py

@@ -0,0 +1,20 @@
+import streamlit as st
+import pandas as pd
+from custom_ollama import query_ollama
+
+# App title
+st.title("AI-Powered Data Analysis with Local Ollama (FastAPI)")
+
+# Upload dataset
+uploaded_file = st.file_uploader("Upload your dataset (CSV)", type=["csv"])
+if uploaded_file:
+    df = pd.read_csv(uploaded_file)
+    st.write("### Preview of Your Dataset", df.head())
+
+    # Run analysis using Ollama
+    if st.button("Run Data Analysis"):
+        with st.spinner("Analyzing your data..."):
+            prompt = f"Perform a basic exploratory data analysis on the following dataset:\n{df.head()}"
+            response = query_ollama(prompt)
+            st.success("Analysis complete!")
+            st.write(response)

crews/data_analysis_crew/config/agents.yaml

@@ -0,0 +1,27 @@
+agents:
+  - name: data_cleaning_agent
+    role: "Data Cleaning Specialist"
+    goal: "Preprocess and clean the dataset by handling missing values and outliers."
+    backstory: "An expert in data preprocessing and cleaning, ensuring datasets are free of missing values and inconsistencies."
+    llm:
+      provider: "ollama"
+      model: "llama2"
+      base_url: "http://localhost:11434"
+
+  - name: eda_agent
+    role: "EDA Specialist"
+    goal: "Perform exploratory data analysis and generate insightful visualizations."
+    backstory: "You specialize in finding hidden patterns and correlations within data using various visualization techniques."
+    llm:
+      provider: "ollama"
+      model: "llama2"
+      base_url: "http://localhost:11434"
+
+  - name: ml_insight_agent
+    role: "Machine Learning Specialist"
+    goal: "Train predictive models and provide feature importance analysis."
+    backstory: "An AI expert focused on building, training, and interpreting machine learning models for accurate predictions."
+    llm:
+      provider: "ollama"
+      model: "llama2"
+      base_url: "http://localhost:11434"

crews/data_analysis_crew/config/tasks.yaml

@@ -0,0 +1,15 @@
+tasks:
+  - name: clean_data_task
+    description: "Clean the dataset by handling missing values, outliers, and ensuring data consistency."
+    expected_output: "A cleaned dataset with no missing values or outliers."
+    agent: "data_cleaning_agent"
+
+  - name: eda_task
+    description: "Generate summary statistics, feature distributions, and correlation heatmaps to explore the dataset."
+    expected_output: "Visualizations including summary statistics, distributions, and correlations."
+    agent: "eda_agent"
+
+  - name: ml_task
+    description: "Train machine learning models and analyze feature importance based on the cleaned dataset."
+    expected_output: "A trained machine learning model with feature importance analysis."
+    agent: "ml_insight_agent"

crews/data_analysis_crew/data_analysis_crew.py

@@ -0,0 +1,38 @@
+import yaml
+from crewai import Crew, Agent, Task
+
+# Load agents from YAML
+with open('crews/data_analysis_crew/config/agents.yaml', 'r') as file:
+    agents_config = yaml.safe_load(file)['agents']
+
+# Load tasks from YAML
+with open('crews/data_analysis_crew/config/tasks.yaml', 'r') as file:
+    tasks_config = yaml.safe_load(file)['tasks']
+
+# Create agent instances
+agents = []
+for agent_data in agents_config:
+    agents.append(Agent(
+        role=agent_data['role'],
+        goal=agent_data['goal'],
+        backstory=agent_data['backstory'],
+        llm=None  # LLM configuration will be handled by Ollama integration within CrewAI
+    ))
+
+# Map agent names to actual objects
+agent_dict = {agent_data['name']: agent for agent_data, agent in zip(agents_config, agents)}
+
+# Create task instances
+tasks = []
+for task_data in tasks_config:
+    tasks.append(Task(
+        description=task_data['description'],
+        expected_output=task_data['expected_output'],
+        agent=agent_dict[task_data['agent']]
+    ))
+
+# Define the crew workflow
+crew = Crew(
+    agents=agents,
+    tasks=tasks
+)

custom_ollama.py

@@ -0,0 +1,12 @@
+import requests
+
+# Function to call the FastAPI server
+def query_ollama(prompt):
+    url = "http://localhost:8000/generate/"
+    payload = {"prompt": prompt}
+    response = requests.post(url, json=payload)
+    
+    if response.status_code == 200:
+        return response.json()["response"]
+    else:
+        raise Exception(f"Failed to generate response: {response.text}")

fastapi_app.py

@@ -0,0 +1,22 @@
+from fastapi import FastAPI, HTTPException
+from pydantic import BaseModel
+from transformers import pipeline
+
+# Initialize FastAPI app
+app = FastAPI()
+
+# Pydantic schema for input data
+class Query(BaseModel):
+    prompt: str
+
+# Load a Hugging Face pipeline
+model = pipeline("text-generation", model="gpt2")
+
+# Define a POST endpoint for processing queries
+@app.post("/generate/")
+async def generate_text(query: Query):
+    try:
+        response = model(query.prompt, max_length=150, num_return_sequences=1)
+        return {"response": response[0]["generated_text"]}
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=str(e))

requirements.txt

@@ -0,0 +1,5 @@
+fastapi[all]
+uvicorn
+transformers
+pydantic
+streamlit