Create app.py

app.py

@@ -0,0 +1,408 @@
+import gradio as gr
+import json
+import pandas as pd
+from datetime import datetime
+import os
+
+class PreferenceCollector:
+    def __init__(self, jsonl_file):
+        self.jsonl_file = jsonl_file
+        self.data = []
+        self.current_index = 0
+        self.preferences = {}
+        self.load_data()
+        
+    def load_data(self):
+        """Load data from JSONL file"""
+        try:
+            with open(self.jsonl_file, 'r') as f:
+                for line in f:
+                    self.data.append(json.loads(line))
+        except FileNotFoundError:
+            print(f"File {self.jsonl_file} not found!")
+            self.data = []
+    
+    def get_current_question(self):
+        """Get the current question data"""
+        if 0 <= self.current_index < len(self.data):
+            return self.data[self.current_index]
+        return None
+    
+    def record_preference(self, sample_choice):
+        """Record user's preference for current question"""
+        if sample_choice in ["sample_zero", "sample_one", "sample_two"]:
+            self.preferences[self.current_index] = {
+                "question_index": self.current_index,
+                "preferred_sample": sample_choice,
+                "timestamp": datetime.now().isoformat()
+            }
+            return True
+        return False
+    
+    def save_results(self, filename="preferences_results.json"):
+        """Save preference results to file"""
+        with open(filename, 'w') as f:
+            json.dump(self.preferences, f, indent=2)
+        
+        # Also create a summary
+        summary = {
+            "total_questions": len(self.data),
+            "answered_questions": len(self.preferences),
+            "sample_zero_count": sum(1 for p in self.preferences.values() if p["preferred_sample"] == "sample_zero"),
+            "sample_one_count": sum(1 for p in self.preferences.values() if p["preferred_sample"] == "sample_one"),
+            "sample_two_count": sum(1 for p in self.preferences.values() if p["preferred_sample"] == "sample_two")
+        }
+        
+        with open("preferences_summary.json", 'w') as f:
+            json.dump(summary, f, indent=2)
+        
+        return summary
+
+# Initialize the preference collector
+collector = None
+selected_sample = None
+
+def initialize_collector(file_path):
+    """Initialize the collector with a file path"""
+    global collector
+    if file_path:
+        collector = PreferenceCollector(file_path.name)
+        if collector.data:
+            return update_display()
+        else:
+            return (
+                "No data loaded. Please check your file.",
+                "", "", "", "",
+                gr.update(interactive=False),
+                gr.update(interactive=False),
+                gr.update(interactive=False),
+                gr.update(interactive=False),
+                gr.update(interactive=False),
+                "0 / 0"
+            )
+    return (
+        "Please upload a JSONL file to begin.",
+        "", "", "", "",
+        gr.update(interactive=False),
+        gr.update(interactive=False),
+        gr.update(interactive=False),
+        gr.update(interactive=False),
+        gr.update(interactive=False),
+        "0 / 0"
+    )
+
+def update_display():
+    """Update the display with current question data"""
+    global collector, selected_sample
+    
+    if not collector or not collector.data:
+        return (
+            "No data loaded",
+            "", "", "", "",
+            gr.update(interactive=False),
+            gr.update(interactive=False),
+            gr.update(interactive=False),
+            gr.update(interactive=False),
+            gr.update(interactive=False),
+            "0 / 0"
+        )
+    
+    question = collector.get_current_question()
+    if not question:
+        return (
+            "No more questions",
+            "", "", "", "",
+            gr.update(interactive=False),
+            gr.update(interactive=False),
+            gr.update(interactive=False),
+            gr.update(interactive=False),
+            gr.update(interactive=False),
+            f"{collector.current_index + 1} / {len(collector.data)}"
+        )
+    
+    # Reset selected sample
+    selected_sample = None
+    
+    # Check if this question has been answered before
+    is_answered = collector.current_index in collector.preferences
+    
+    return (
+        question.get("introductory_example", "No introductory example"),
+        question.get("sample_zero", "No sample zero"),
+        question.get("sample_one", "No sample one"),
+        question.get("sample_two", "No sample two"),
+        f"{'✓ Answered' if is_answered else 'Not answered yet'}",
+        gr.update(interactive=True, variant="primary" if selected_sample == "sample_zero" else "secondary"),
+        gr.update(interactive=True, variant="primary" if selected_sample == "sample_one" else "secondary"),
+        gr.update(interactive=True, variant="primary" if selected_sample == "sample_two" else "secondary"),
+        gr.update(interactive=collector.current_index > 0),
+        gr.update(interactive=collector.current_index < len(collector.data) - 1),
+        f"{collector.current_index + 1} / {len(collector.data)}"
+    )
+
+def select_sample(sample_name):
+    """Handle sample selection"""
+    global selected_sample
+    selected_sample = sample_name
+    
+    # Update button variants to show selection
+    return (
+        gr.update(variant="primary" if sample_name == "sample_zero" else "secondary"),
+        gr.update(variant="primary" if sample_name == "sample_one" else "secondary"),
+        gr.update(variant="primary" if sample_name == "sample_two" else "secondary"),
+        gr.update(interactive=True)  # Enable confirm button
+    )
+
+def confirm_selection():
+    """Confirm the current selection"""
+    global collector, selected_sample
+    
+    if not collector or not selected_sample:
+        return (
+            gr.update(value="Please select a sample first"),
+            gr.update(interactive=False)
+        )
+    
+    if collector.record_preference(selected_sample):
+        # Auto-advance to next question if not the last one
+        if collector.current_index < len(collector.data) - 1:
+            collector.current_index += 1
+            return (
+                gr.update(value="✓ Selection recorded! Moving to next question..."),
+                gr.update(interactive=False),
+                *update_display()
+            )
+        else:
+            return (
+                gr.update(value="✓ Selection recorded! You've completed all questions."),
+                gr.update(interactive=False),
+                *update_display()
+            )
+    
+    return (
+        gr.update(value="Error recording selection"),
+        gr.update(interactive=False)
+    )
+
+def previous_question():
+    """Go to previous question"""
+    global collector
+    if collector and collector.current_index > 0:
+        collector.current_index -= 1
+        return update_display()
+    return update_display()
+
+def next_question():
+    """Go to next question"""
+    global collector
+    if collector and collector.current_index < len(collector.data) - 1:
+        collector.current_index += 1
+        return update_display()
+    return update_display()
+
+def export_results():
+    """Export the results"""
+    global collector
+    if collector:
+        summary = collector.save_results()
+        return f"""Results exported successfully!
+        
+Total questions: {summary['total_questions']}
+Answered: {summary['answered_questions']}
+
+Preferences:
+- Sample 0: {summary['sample_zero_count']} votes
+- Sample 1: {summary['sample_one_count']} votes
+- Sample 2: {summary['sample_two_count']} votes
+
+Files saved: preferences_results.json, preferences_summary.json"""
+    return "No data to export"
+
+# Create the Gradio interface
+with gr.Blocks(theme=gr.themes.Soft()) as demo:
+    gr.Markdown(
+        """
+        # 📊 Preference Collection Interface
+        
+        Upload a JSONL file and select your preferred sample for each question.
+        """
+    )
+    
+    with gr.Row():
+        file_input = gr.File(
+            label="Upload JSONL File",
+            file_types=[".jsonl"],
+            type="filepath"
+        )
+        load_btn = gr.Button("Load File", variant="primary")
+    
+    gr.Markdown("---")
+    
+    # Question display
+    with gr.Column():
+        intro_display = gr.Textbox(
+            label="📝 Question / Introductory Example",
+            lines=3,
+            interactive=False
+        )
+        
+        progress_label = gr.Textbox(
+            label="Progress",
+            value="0 / 0",
+            interactive=False,
+            max_lines=1
+        )
+        
+        status_label = gr.Textbox(
+            label="Status",
+            value="Not answered yet",
+            interactive=False,
+            max_lines=1
+        )
+    
+    gr.Markdown("### 🤔 Choose your preferred sample:")
+    
+    # Sample displays
+    with gr.Row():
+        with gr.Column():
+            sample_0_display = gr.Textbox(
+                label="Sample 0",
+                lines=6,
+                interactive=False
+            )
+            select_0_btn = gr.Button(
+                "Select Sample 0",
+                variant="secondary",
+                interactive=False
+            )
+        
+        with gr.Column():
+            sample_1_display = gr.Textbox(
+                label="Sample 1",
+                lines=6,
+                interactive=False
+            )
+            select_1_btn = gr.Button(
+                "Select Sample 1",
+                variant="secondary",
+                interactive=False
+            )
+        
+        with gr.Column():
+            sample_2_display = gr.Textbox(
+                label="Sample 2",
+                lines=6,
+                interactive=False
+            )
+            select_2_btn = gr.Button(
+                "Select Sample 2",
+                variant="secondary",
+                interactive=False
+            )
+    
+    # Action buttons
+    with gr.Row():
+        prev_btn = gr.Button("← Previous", interactive=False)
+        confirm_btn = gr.Button("✓ Confirm Selection", variant="primary", interactive=False)
+        next_btn = gr.Button("Next →", interactive=False)
+    
+    confirm_status = gr.Textbox(label="Action Status", interactive=False)
+    
+    gr.Markdown("---")
+    
+    # Export section
+    with gr.Row():
+        export_btn = gr.Button("📥 Export Results", variant="primary")
+        export_status = gr.Textbox(label="Export Status", interactive=False)
+    
+    # Event handlers
+    load_btn.click(
+        initialize_collector,
+        inputs=[file_input],
+        outputs=[
+            intro_display, sample_0_display, sample_1_display, sample_2_display,
+            status_label, select_0_btn, select_1_btn, select_2_btn,
+            prev_btn, next_btn, progress_label
+        ]
+    )
+    
+    select_0_btn.click(
+        lambda: select_sample("sample_zero"),
+        outputs=[select_0_btn, select_1_btn, select_2_btn, confirm_btn]
+    )
+    
+    select_1_btn.click(
+        lambda: select_sample("sample_one"),
+        outputs=[select_0_btn, select_1_btn, select_2_btn, confirm_btn]
+    )
+    
+    select_2_btn.click(
+        lambda: select_sample("sample_two"),
+        outputs=[select_0_btn, select_1_btn, select_2_btn, confirm_btn]
+    )
+    
+    confirm_btn.click(
+        confirm_selection,
+        outputs=[
+            confirm_status, confirm_btn,
+            intro_display, sample_0_display, sample_1_display, sample_2_display,
+            status_label, select_0_btn, select_1_btn, select_2_btn,
+            prev_btn, next_btn, progress_label
+        ]
+    )
+    
+    prev_btn.click(
+        previous_question,
+        outputs=[
+            intro_display, sample_0_display, sample_1_display, sample_2_display,
+            status_label, select_0_btn, select_1_btn, select_2_btn,
+            prev_btn, next_btn, progress_label
+        ]
+    )
+    
+    next_btn.click(
+        next_question,
+        outputs=[
+            intro_display, sample_0_display, sample_1_display, sample_2_display,
+            status_label, select_0_btn, select_1_btn, select_2_btn,
+            prev_btn, next_btn, progress_label
+        ]
+    )
+    
+    export_btn.click(
+        export_results,
+        outputs=[export_status]
+    )
+
+# Launch the app
+if __name__ == "__main__":
+    # Create a sample JSONL file for testing
+    sample_data = [
+        {
+            "introductory_example": "Which response best explains machine learning?",
+            "sample_zero": "Machine learning is a subset of AI that enables systems to learn and improve from experience without being explicitly programmed.",
+            "sample_one": "Machine learning involves training computers to recognize patterns in data and make decisions based on those patterns.",
+            "sample_two": "Machine learning is the science of getting computers to act without being explicitly programmed by using algorithms and statistical models."
+        },
+        {
+            "introductory_example": "How would you describe cloud computing?",
+            "sample_zero": "Cloud computing delivers computing services over the internet, including servers, storage, databases, and software.",
+            "sample_one": "Cloud computing is the on-demand availability of computer system resources, especially data storage and computing power.",
+            "sample_two": "Cloud computing provides shared processing resources and data to computers and other devices on demand via the internet."
+        },
+        {
+            "introductory_example": "What is the best explanation of blockchain?",
+            "sample_zero": "Blockchain is a distributed ledger technology that records transactions across many computers in a way that cannot be altered retroactively.",
+            "sample_one": "Blockchain is a system of recording information in a way that makes it difficult to change, hack, or cheat the system.",
+            "sample_two": "Blockchain is a decentralized, distributed database that maintains a continuously growing list of ordered records called blocks."
+        }
+    ]
+    
+    # Save sample data
+    with open("sample_questions.jsonl", "w") as f:
+        for item in sample_data:
+            f.write(json.dumps(item) + "\n")
+    
+    print("Sample file 'sample_questions.jsonl' created!")
+    
+    demo.launch(share=False)