app.py

import os
import json
import gradio as gr
import pandas as pd
import numpy as np
import faiss
from sklearn.preprocessing import StandardScaler
from groq import Groq
import plotly.express as px

# Set up Groq client
client = Groq(api_key="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")

# Initialize FAISS index and scaler
scaler = StandardScaler()
index = None
data_columns = []
project_ids = []

# Function to load historical project data
def load_data(file):
    global index, scaler, data_columns, project_ids
    try:
        # Load the CSV file
        data = pd.read_csv(file.name)
        data_columns = data.columns.tolist()
        project_ids = data.iloc[:, 0].tolist()  # First column assumed as project IDs
        
        # Process numerical data
        numerical_data = data.select_dtypes(include=[np.number])
        if numerical_data.empty:
            return "Error: No numeric columns found in the dataset."

        scaled_data = scaler.fit_transform(numerical_data.values)
        faiss_data = scaled_data.astype("float32")
        dimension = faiss_data.shape[1]
        index = faiss.IndexFlatL2(dimension)
        index.add(faiss_data)
        return f"Data loaded successfully! Columns: {', '.join(data_columns)}"
    except Exception as e:
        return f"Error loading data: {e}"

# Function to generate construction schedule
def generate_schedule(input_parameters):
    global index, scaler, project_ids, data_columns
    try:
        # Parse input parameters
        parameters = list(map(float, input_parameters.split(",")))
        input_features = np.array([parameters]).astype("float32")
        scaled_features = scaler.transform(input_features)

        # Query FAISS for similar projects
        k = 5  # Number of similar projects to retrieve
        distances, indices = index.search(scaled_features, k)
        similar_projects = [{"project_id": project_ids[i]} for i in indices[0]]

        # Construct prompt for Groq API
        prompt = f"""
        Based on these similar projects:
        {similar_projects}
        Generate an optimized construction schedule for a project with the following parameters:
        Duration: {parameters[0]} days, Milestones: {parameters[1]}, Delays: {parameters[2]} days.
        Output a JSON object for a Gantt chart, including tasks, start and end dates, and dependencies.
        """
        response = client.chat.completions.create(
            messages=[{"role": "user", "content": prompt}],
            model="llama3-8b-8192",
        )
        generated_schedule = response.choices[0].message.content
        
        # Parse the Groq API response
        schedule_dict = json.loads(generated_schedule)
        schedule_data = pd.DataFrame(schedule_dict)

        # Generate Gantt chart
        fig = px.timeline(
            schedule_data,
            x_start="start_date",
            x_end="end_date",
            y="task",
            title="Optimized Construction Schedule",
            labels={"task": "Task", "start_date": "Start Date", "end_date": "End Date"},
        )
        fig.update_yaxes(categoryorder="total ascending")
        return "Schedule generated successfully!", fig
    except Exception as e:
        return f"Error generating schedule: {e}", None

# Gradio app definition
def main():
    with gr.Blocks() as app:
        gr.Markdown("# Intelligent Construction Scheduling Assistant")

        with gr.Tab("Upload Data"):
            data_file = gr.File(label="Upload Historical Project Data (CSV)")
            upload_button = gr.Button("Upload")
            upload_output = gr.Textbox(label="Output", interactive=False)
            upload_button.click(load_data, inputs=data_file, outputs=upload_output)

        with gr.Tab("Generate Schedule"):
            input_params = gr.Textbox(
                label="Input Parameters (comma-separated, e.g., '120, 10, 5')",
                placeholder="Enter duration, milestones, and delays (in days)",
            )
            generate_button = gr.Button("Generate Schedule")
            output_text = gr.Textbox(label="Output", interactive=False)
            output_plot = gr.Plot(label="Schedule Gantt Chart")
            generate_button.click(
                generate_schedule, inputs=input_params, outputs=[output_text, output_plot]
            )

    return app

# Run app
if __name__ == "__main__":
    app = main()
    app.launch()