import streamlit as st
import tempfile
import os
import pandas as pd
import base64
from inference import InferencePipeline


# === Streamlit App Configuration ===
st.set_page_config(page_title="Solar Panel Fault Detection", layout="wide")
st.title("🔍 Solar Panel Fault Detection (Roboflow Workflow)")
st.write("Upload a thermal video (MP4). Faults will be detected using your custom Roboflow model.")

# === Globals for Fault Logging ===
fault_log = []

# === Fault Handler Callback ===
def my_sink(result, video_frame):
    global fault_log
    if result.get("predictions"):
        frame_idx = result.get("frame_id", -1)
        timestamp = result.get("timestamp", -1.0)

        for pred in result["predictions"]:
            label = pred["class"]
            conf = round(pred["confidence"], 2)
            x, y, w, h = pred["x"], pred["y"], pred["width"], pred["height"]
            x1, y1 = int(x - w / 2), int(y - h / 2)
            x2, y2 = int(x + w / 2), int(y + h / 2)

            fault_log.append({
                "Frame": frame_idx,
                "Time (s)": round(timestamp, 2),
                "Fault Type": label,
                "Confidence": conf,
                "Box": f"({x1},{y1},{x2},{y2})"
            })

# === CSV Conversion ===
def convert_df(df):
    return df.to_csv(index=False).encode("utf-8")

# === Streamlit UI ===
uploaded_file = st.file_uploader("📤 Upload thermal video", type=["mp4"])

if uploaded_file:
    st.video(uploaded_file)

    # Save uploaded file temporarily
    temp_input_path = tempfile.NamedTemporaryFile(delete=False, suffix=".mp4").name
    with open(temp_input_path, "wb") as f:
        f.write(uploaded_file.read())

    st.info("⏳ Running fault detection...")

    # Initialize and start the inference pipeline
    pipeline = InferencePipeline.init_with_workflow(
        api_key="dxkgGGHSZ3DI8XzVn29U",
        workspace_name="naveen-kumar-hnmil",
        workflow_id="custom-workflow",
        video_reference=temp_input_path,
        max_fps=0.5,
        on_prediction=my_sink
        # max_duration_seconds=30  # Optional: limit for testing
    )

    pipeline.start()
    pipeline.join()

    # Display results
    if fault_log:
        df = pd.DataFrame(fault_log)
        st.subheader("📊 Detected Faults Table")
        st.dataframe(df)

        st.download_button(
            "📥 Download Fault Log CSV",
            convert_df(df),
            "fault_log.csv",
            "text/csv"
        )
    else:
        st.success("✅ No faults detected.")

    os.unlink(temp_input_path)

st.markdown("---")
st.caption("Built with Streamlit + Roboflow Inference SDK")