Update app.py

app.py

@@ -1,58 +1,66 @@
+# app.py
+
 import gradio as gr
 from inference import get_evo_response, get_gpt_response
-from logger import log_feedback
-import retrain
 import pandas as pd
 import os
 
-def advisor_interface(question, context, options_text, feedback_choice):
-    options = [opt.strip() for opt in options_text.split("\n") if opt.strip()]
-    evo_output, evo_reasoning = get_evo_response(question, context, options)
-    gpt_output = get_gpt_response(question, context, options)
+LOG_PATH = "feedback_log.csv"
+os.makedirs(os.path.dirname(LOG_PATH), exist_ok=True)
+if not os.path.exists(LOG_PATH):
+    pd.DataFrame(columns=["question", "context", "option1", "option2", "evo_answer", "feedback"]).to_csv(LOG_PATH, index=False)
+
+def advisor_interface(question, context, options_text, feedback=None):
+    options = [o.strip() for o in options_text.split("\n") if o.strip()]
+    if len(options) != 2:
+        return "Please enter exactly two options (one per line).", "", ""
 
-    if feedback_choice != "No feedback":
-        label = 1 if feedback_choice == "👍 Helpful" else 0
-        log_feedback(question, context, evo_output, label)
+    option1, option2 = options
+    evo_answer, confidence, s1, s2 = get_evo_response(question, option1, option2)
+    gpt_output = get_gpt_response(question, option1, option2)
 
-    return evo_reasoning, gpt_output, load_history()
+    # Log feedback if given
+    if feedback:
+        df = pd.read_csv(LOG_PATH)
+        df = df.append({
+            "question": question,
+            "context": context,
+            "option1": option1,
+            "option2": option2,
+            "evo_answer": evo_answer,
+            "feedback": feedback
+        }, ignore_index=True)
+        df.to_csv(LOG_PATH, index=False)
 
-def retrain_evo():
-    retrain.fine_tune_on_feedback()
-    return "✅ Evo retrained on feedback.", load_history()
+    evo_msg = f"Evo suggests: **{evo_answer}** (Confidence: {confidence:.2f})"
+    context_used = f"Option 1 Score: {s1:.2f}\nOption 2 Score: {s2:.2f}"
+    return evo_msg + "\n\n" + context_used, gpt_output, load_history()
 
 def load_history():
-    if os.path.exists("feedback_log.csv"):
-        df = pd.read_csv("feedback_log.csv")
-        try:
-            return df.tail(10).to_markdown(index=False)
-        except ImportError:
-            return "📜 History available, but markdown rendering failed. Install `tabulate` for full view."
-    return "No history available yet."
+    df = pd.read_csv(LOG_PATH)
+    if df.empty:
+        return "No feedback yet."
+    return df.tail(10).to_markdown(index=False)
 
 with gr.Blocks() as demo:
-    gr.Markdown("## 🧠 EvoRAG – General-Purpose Adaptive AI with RAG + Web Search")
+    gr.Markdown("🧠 **EvoRAG – General-Purpose Adaptive AI with Web Reasoning**")
 
     with gr.Row():
-        question = gr.Textbox(label="📝 Ask anything", placeholder="e.g. Who is the current president of the US?")
-    with gr.Row():
-        options_text = gr.Textbox(label="🔢 Options (one per line)", placeholder="Option 1\nOption 2\nOption 3", lines=4)
-    with gr.Row():
-        context = gr.Textbox(label="📂 Optional Context or Notes", placeholder="Paste memo, article, or info here...", lines=4)
+        with gr.Column():
+            question = gr.Textbox(label="📝 Ask anything")
+            context = gr.Textbox(label="📂 Optional Context or Notes", lines=2)
+            options = gr.Textbox(label="🧠 Options (Enter two options, one per line)", placeholder="Option 1\nOption 2")
+            run_btn = gr.Button("🔍 Run Advisors")
 
-    feedback = gr.Radio(["👍 Helpful", "👎 Not Helpful", "No feedback"], label="Was Evo’s answer useful?", value="No feedback")
+            feedback = gr.Radio(["👍 Helpful", "👎 Not Helpful", "No feedback"], label="Was Evo’s answer useful?")
+            evo_out = gr.Markdown()
+            gpt_out = gr.Markdown()
+            history = gr.Markdown()
 
-    with gr.Row():
-        evo_out = gr.Textbox(label="🔬 EvoRAG Suggestion (with reasoning)")
-        gpt_out = gr.Textbox(label="🤖 GPT-3.5 Suggestion")
-
-    run_button = gr.Button("Run Advisors")
-    run_button.click(fn=advisor_interface, inputs=[question, context, options_text, feedback], outputs=[evo_out, gpt_out, gr.Textbox(label="📜 Recent History")])
-
-    gr.Markdown("---")
-    gr.Markdown("### 🔁 Retrain Evo from Feedback")
-    retrain_button = gr.Button("📚 Retrain Evo")
-    retrain_output = gr.Textbox(label="🛠️ Retrain Status")
-    history_output = gr.Textbox(label="📜 Recent History")
-    retrain_button.click(fn=retrain_evo, inputs=[], outputs=[retrain_output, history_output])
+        run_btn.click(
+            advisor_interface,
+            inputs=[question, context, options, feedback],
+            outputs=[evo_out, gpt_out, history]
+        )
 
 demo.launch()