updated gradio to call the functions directly

app.py

@@ -1,54 +1,50 @@
 import os
 import gradio as gr
+from Oracle.deepfundingoracle import prepare_dataset, train_predict_weight, create_submission_csv
 import pandas as pd
-import networkx as nx
 import matplotlib.pyplot as plt
-from Oracle.deepfundingoracle import prepare_dataset, train_predict_weight, create_submission_csv, SmolLM
-from Oracle.DataSmolAgent import DataSmolAgent
+import time
+import io
 
-def pipeline(upload):
-    # Load and clean/extract features via SmolAgents
-    df_raw = pd.read_csv(upload.name)
-    df_features = DataSmolAgent(df_raw).run(
-        prompt="Clean and extract features from the uploaded data", output_csv=False)
-    # Save preprocessed features and run dataset preparation on them
-    processed_path = "processed_input.csv"
-    pd.DataFrame(df_features).to_csv(processed_path, index=False)
-    df_prepared = prepare_dataset(processed_path)
-    # Assign base weights and predict final weights
-    df_results = train_predict_weight(df_prepared)
-    # Create submission CSV
-    csv_path = create_submission_csv(df_results, "submission.csv")
-    # Build dependency graph
-    G = nx.DiGraph()
-    for _, row in df_results.iterrows():
-        G.add_edge(row["parent"], row["repo"], weight=row["final_weight"])
-    plt.figure(figsize=(10, 8))
-    pos = nx.spring_layout(G)
-    weights = [G[u][v]["weight"] for u, v in G.edges()]
-    nx.draw(G, pos, with_labels=True, node_size=500, node_color="lightblue",
-            edge_color=weights, edge_cmap=plt.get_cmap('viridis'), width=2)
-    plt.savefig("graph.png")
-    # Generate explanation via SmolLM
-    explanation = SmolLM().predict(
-        "Explain the dependency graph and weight assignments for the dataset.")
-    # Return results
-    return (df_results.head().to_dict("records"), csv_path, "graph.png", explanation)
+def analyze_file(file, progress=gr.Progress(track_tqdm=True)):
+    start_time = time.time()
+    progress(0, desc="Preparing dataset...")
+    df = prepare_dataset(file.name)
+    progress(0.3, desc="Predicting weights...")
+    df = train_predict_weight(df)
+    progress(0.6, desc="Saving results to CSV...")
+    csv_path = create_submission_csv(df, "submission.csv")
+    progress(0.8, desc="Generating graph...")
+    # Example: plot histogram of a column if exists
+    fig, ax = plt.subplots()
+    if 'final_weight' in df.columns:
+        df['final_weight'].hist(ax=ax)
+        ax.set_title('Distribution of Final Weights')
+        ax.set_xlabel('Final Weight')
+        ax.set_ylabel('Count')
+    else:
+        ax.text(0.5, 0.5, 'No final_weight column to plot', ha='center')
+    buf = io.BytesIO()
+    plt.savefig(buf, format='png')
+    buf.seek(0)
+    plt.close(fig)
+    progress(1, desc="Done!")
+    elapsed = time.time() - start_time
+    preview = df.head().to_csv(index=False)
+    return preview, csv_path, buf, f"Analysis completed in {elapsed:.2f} seconds."
 
 iface = gr.Interface(
-    fn=pipeline,
-    inputs=gr.File(label="Upload CSV", type="filepath"),
+    fn=analyze_file,
+    inputs=gr.File(label="Upload CSV"),
     outputs=[
-        gr.Dataframe(label="Preview of Results"),
+        gr.Textbox(label="Preview of Results"),
         gr.File(label="Download CSV"),
-        gr.Image(label="Dependency Graph"),
-        gr.Textbox(label="Explanation")
+        gr.Image(label="Analysis Graph"),
+        gr.Textbox(label="Status/Timing Info")
     ],
     title="DeepFunding Oracle",
-    description=(
-        "Upload a CSV to extract features, assign base weights via LLama, predict final weights with RandomForest, "
-        "and visualize the dependency graph with explanations."
-    )
+    description="Upload a CSV of repo-parent relationships; see analysis progress, get a graph, and download results as CSV.",
+    allow_flagging="never"
 )
 
 if __name__ == "__main__":