backup

app.py

@@ -9,7 +9,7 @@ from huggingface_hub import CommitScheduler, HfApi
 api = HfApi(token=os.environ["HF_TOKEN"])
 
 scheduler = CommitScheduler(
-    repo_id="taesiri/EdgeQuest",
+    repo_id="taesiri/zb_dataset_storage",
     repo_type="dataset",
     folder_path="./data",
     path_in_repo="data",
@@ -18,27 +18,29 @@ scheduler = CommitScheduler(
 
 
 def generate_json_files(
-    system_message,
-    # New fields
     name,
     email_address,
     institution,
     openreview_profile,
+    authorship_interest,
     question_categories,
     subquestion_1_text,
     subquestion_1_answer,
     subquestion_2_text,
     subquestion_2_answer,
-    # Existing fields
+    subquestion_3_text,
+    subquestion_3_answer,
+    subquestion_4_text,
+    subquestion_4_answer,
+    subquestion_5_text,
+    subquestion_5_answer,
     question,
     final_answer,
     rationale_text,
-    # Question images
     image1,
     image2,
     image3,
     image4,
-    # Rationale images
     rationale_image1,
     rationale_image2,
 ):
@@ -47,7 +49,7 @@ def generate_json_files(
       1) Create a unique folder under ./data/
       2) Copy uploaded images (question + rationale) into that folder
       3) Produce two JSON files:
-         - request_urls.json   (local file paths in content)
+         - question.json   (local file paths in content)
          - request_base64.json (base64-encoded images in content)
       4) Return paths to both files for Gradio to provide as download links
     """
@@ -61,10 +63,6 @@ def generate_json_files(
     request_folder = os.path.join(parent_data_folder, request_id)
     os.makedirs(request_folder)
 
-    # Handle defaults
-    if not system_message:
-        system_message = "You are a helpful assistant"
-
     # Convert None strings
     def safe_str(val):
         return val if val is not None else ""
@@ -73,6 +71,7 @@ def generate_json_files(
     email_address = safe_str(email_address)
     institution = safe_str(institution)
     openreview_profile = safe_str(openreview_profile)
+    authorship_interest = safe_str(authorship_interest)
     # Convert question_categories to list
     question_categories = (
         [cat.strip() for cat in safe_str(question_categories).split(",")]
@@ -83,6 +82,12 @@ def generate_json_files(
     subquestion_1_answer = safe_str(subquestion_1_answer)
     subquestion_2_text = safe_str(subquestion_2_text)
     subquestion_2_answer = safe_str(subquestion_2_answer)
+    subquestion_3_text = safe_str(subquestion_3_text)
+    subquestion_3_answer = safe_str(subquestion_3_answer)
+    subquestion_4_text = safe_str(subquestion_4_text)
+    subquestion_4_answer = safe_str(subquestion_4_answer)
+    subquestion_5_text = safe_str(subquestion_5_text)
+    subquestion_5_answer = safe_str(subquestion_5_answer)
     question = safe_str(question)
     final_answer = safe_str(final_answer)
     rationale_text = safe_str(rationale_text)
@@ -117,6 +122,7 @@ def generate_json_files(
         {"type": "field", "label": "email_address", "value": email_address},
         {"type": "field", "label": "institution", "value": institution},
         {"type": "field", "label": "openreview_profile", "value": openreview_profile},
+        {"type": "field", "label": "authorship_interest", "value": authorship_interest},
         {"type": "field", "label": "question_categories", "value": question_categories},
         {"type": "field", "label": "subquestion_1_text", "value": subquestion_1_text},
         {
@@ -130,27 +136,23 @@ def generate_json_files(
             "label": "subquestion_2_answer",
             "value": subquestion_2_answer,
         },
-        {"type": "field", "label": "question", "value": question},
-        {"type": "field", "label": "final_answer", "value": final_answer},
-        {"type": "field", "label": "rationale_text", "value": rationale_text},
-    ]
-    content_list_base64 = [
-        {"type": "field", "label": "name", "value": name},
-        {"type": "field", "label": "email_address", "value": email_address},
-        {"type": "field", "label": "institution", "value": institution},
-        {"type": "field", "label": "openreview_profile", "value": openreview_profile},
-        {"type": "field", "label": "question_categories", "value": question_categories},
-        {"type": "field", "label": "subquestion_1_text", "value": subquestion_1_text},
+        {"type": "field", "label": "subquestion_3_text", "value": subquestion_3_text},
         {
             "type": "field",
-            "label": "subquestion_1_answer",
-            "value": subquestion_1_answer,
+            "label": "subquestion_3_answer",
+            "value": subquestion_3_answer,
         },
-        {"type": "field", "label": "subquestion_2_text", "value": subquestion_2_text},
+        {"type": "field", "label": "subquestion_4_text", "value": subquestion_4_text},
         {
             "type": "field",
-            "label": "subquestion_2_answer",
-            "value": subquestion_2_answer,
+            "label": "subquestion_4_answer",
+            "value": subquestion_4_answer,
+        },
+        {"type": "field", "label": "subquestion_5_text", "value": subquestion_5_text},
+        {
+            "type": "field",
+            "label": "subquestion_5_answer",
+            "value": subquestion_5_answer,
         },
         {"type": "field", "label": "question", "value": question},
         {"type": "field", "label": "final_answer", "value": final_answer},
@@ -169,132 +171,56 @@ def generate_json_files(
             }
         )
 
-        # 2) Base64 version
-        with open(file_path, "rb") as f:
-            file_bytes = f.read()
-        img_b64_str = base64.b64encode(file_bytes).decode("utf-8")
-        content_list_base64.append(
-            {
-                "type": "image_url",
-                "label": img_label,
-                "image_url": {"url": {"data:image/png;base64": img_b64_str}},
-            }
-        )
-
     # Build the final JSON structures for each approach
-    i = 1
-
-    assistant_content = [
-        {"type": "text", "text": rationale_text},
-        {"type": "text", "text": final_answer},
-    ]
-
     # A) URLs JSON
     item_urls = {
-        "custom_id": f"request______{i}",
+        "custom_id": f"question___{request_id}",
         # Metadata at top level
-        "name": name,
-        "email_address": email_address,
-        "institution": institution,
-        "openreview_profile": openreview_profile,
-        "question_categories": question_categories,
-        "question": {
-            "messages": [
-                {"role": "system", "content": system_message},
-                {
-                    "role": "user",
-                    "content": [
-                        {"type": "text", "label": "question", "value": question}
-                    ]
-                    + [
-                        item
-                        for item in content_list_urls
-                        if item.get("type") == "image_url"
-                        and "question_image" in item.get("label", "")
-                    ],
-                },
-            ],
-        },
-        "subquestions": [
-            {"text": subquestion_1_text, "answer": subquestion_1_answer},
-            {"text": subquestion_2_text, "answer": subquestion_2_answer},
-        ],
-        "answer": {
-            "final_answer": final_answer,
-            "rationale_text": rationale_text,
-            "rationale_images": [
-                item
-                for item in content_list_urls
-                if item.get("type") == "image_url"
-                and "rationale_image" in item.get("label", "")
-            ],
+        "author_info": {
+            "name": name,
+            "email_address": email_address,
+            "institution": institution,
+            "openreview_profile": openreview_profile,
+            "authorship_interest": authorship_interest,
         },
-    }
-
-    # B) Base64 JSON
-    item_base64 = {
-        "custom_id": f"request______{i}",
-        # Metadata at top level
-        "name": name,
-        "email_address": email_address,
-        "institution": institution,
-        "openreview_profile": openreview_profile,
-        # Question-related fields at top level
         "question_categories": question_categories,
-        "subquestions": [
-            {"text": subquestion_1_text, "answer": subquestion_1_answer},
-            {"text": subquestion_2_text, "answer": subquestion_2_answer},
+        "question": question,
+        "question_images": [
+            item["image_url"]["url"]["data:image/png;path"]
+            for item in content_list_urls
+            if item.get("type") == "image_url"
+            and "question_image" in item.get("label", "")
         ],
         "final_answer": final_answer,
         "rationale_text": rationale_text,
-        "body": {
-            "model": "MODEL_NAME",
-            "messages": [
-                {"role": "system", "content": system_message},
-                {
-                    "role": "user",
-                    "content": [
-                        {"type": "field", "label": "question", "value": question}
-                    ]
-                    + [
-                        item
-                        for item in content_list_base64
-                        if item.get("type") == "image_url"
-                        and "question_image" in item.get("label", "")
-                    ],
-                },
-                {
-                    "role": "assistant",
-                    "content": [
-                        {"type": "text", "text": rationale_text},
-                        {"type": "text", "text": final_answer},
-                        *[
-                            item
-                            for item in content_list_base64
-                            if item.get("type") == "image_url"
-                            and "rationale_image" in item.get("label", "")
-                        ],
-                    ],
-                },
-            ],
-        },
+        "rationale_images": [
+            item["image_url"]["url"]["data:image/png;path"]
+            for item in content_list_urls
+            if item.get("type") == "image_url"
+            and "rationale_image" in item.get("label", "")
+        ],
+        "subquestions_1_text": subquestion_1_text,
+        "subquestions_1_answer": subquestion_1_answer,
+        "subquestions_2_text": subquestion_2_text,
+        "subquestions_2_answer": subquestion_2_answer,
+        "subquestions_3_text": subquestion_3_text,
+        "subquestions_3_answer": subquestion_3_answer,
+        "subquestions_4_text": subquestion_4_text,
+        "subquestions_4_answer": subquestion_4_answer,
+        "subquestions_5_text": subquestion_5_text,
+        "subquestions_5_answer": subquestion_5_answer,
     }
 
     # Convert each to JSON line format
     urls_json_line = json.dumps(item_urls, ensure_ascii=False)
-    base64_json_line = json.dumps(item_base64, ensure_ascii=False)
 
     # 3) Write out two JSON files in request_folder
-    urls_jsonl_path = os.path.join(request_folder, "request_urls.json")
-    base64_jsonl_path = os.path.join(request_folder, "request_base64.json")
+    urls_jsonl_path = os.path.join(request_folder, "question.json")
 
     with open(urls_jsonl_path, "w", encoding="utf-8") as f:
         f.write(urls_json_line + "\n")
-    with open(base64_jsonl_path, "w", encoding="utf-8") as f:
-        f.write(base64_json_line + "\n")
 
-    # Return the two file paths so Gradio can offer them as downloads
-    return urls_jsonl_path, base64_jsonl_path
+    return urls_jsonl_path
 
 
 # Build the Gradio app
@@ -322,19 +248,29 @@ with gr.Blocks() as demo:
                             <li>At least <b>one question image</b></li>
                             <li>The <b>question text</b></li>
                             <li>The <b>final answer</b></li>
+                            <li><b>Sub-questions</b> with their answers (write 'N/A' if breaking into steps is not reasonable - please use sparingly)</li>
                         </ul>
                     </td>
                     <td style="vertical-align: top; padding: 8px; border: 1px solid #ddd;">
                         <ul style="margin: 0;">
-                            <li>Up to four question images</li>
+                            <li>Up to three additional question images</li>
                             <li>Supporting images for your answer</li>
                             <li><b>Rationale text</b> to explain your reasoning</li>
-                            <li><b>Sub-questions</b> with their answers</li>
                         </ul>
                     </td>
                 </tr>
             </table>
 
+            <h3>Question Criteria:</h3>
+            <ul>
+                <li>Make questions as challenging as possible. At a minimum, obtaining the correct answer needs to be beyond the capabilities of state-of-the-art large multimodal models.</li>
+                <li>Structure your questions to require multiple steps/sub-questions to reach the final answer (e.g., identifying/counting specific objects in the image or requiring a particular piece of knowledge) — this will likely enable better differentiation of model performance.</li>
+                <li>Include images/questions that are not copyright-restricted.</li>
+            </ul>
+
+            <h3>Authorship Opportunity:</h3>
+            <p>Would you like to be included as an author on our paper? Authorship is offered to anyone submitting 5 or more difficult questions!</p>
+
             <p>While not all fields are mandatory, providing additional context through optional fields will help create a more comprehensive dataset. After submitting a question, you can clear up the form to submit another one.</p>
             """
         )
@@ -353,6 +289,12 @@ with gr.Blocks() as demo:
             placeholder="Your OpenReview username or profile name",
         )
 
+    # Add authorship checkbox
+    authorship_input = gr.Checkbox(
+        label="Would you like to be considered for authorship? (Requires submitting 5+ difficult questions)",
+        value=False,
+    )
+
     gr.Markdown("## Question Information")
 
     # Question Images - Individual Tabs
@@ -402,30 +344,73 @@ with gr.Blocks() as demo:
     gr.Markdown("## Subquestions")
     with gr.Row():
         subquestion_1_text_input = gr.Textbox(
-            label="Subquestion 1 Text", lines=2, placeholder="First sub-question..."
+            label="Subquestion 1 Text",
+            lines=2,
+            placeholder="First sub-question...",
+            value="N/A",
         )
         subquestion_1_answer_input = gr.Textbox(
             label="Subquestion 1 Answer",
             lines=2,
             placeholder="Answer to sub-question 1...",
+            value="N/A",
         )
 
     with gr.Row():
         subquestion_2_text_input = gr.Textbox(
-            label="Subquestion 2 Text", lines=2, placeholder="Second sub-question..."
+            label="Subquestion 2 Text",
+            lines=2,
+            placeholder="Second sub-question...",
+            value="N/A",
         )
         subquestion_2_answer_input = gr.Textbox(
             label="Subquestion 2 Answer",
             lines=2,
             placeholder="Answer to sub-question 2...",
+            value="N/A",
         )
 
-    system_message_input = gr.Textbox(
-        label="System Message",
-        value="You are a helpful assistant",
-        lines=2,
-        placeholder="Enter the system message that defines the AI assistant's role and behavior...",
-    )
+    with gr.Row():
+        subquestion_3_text_input = gr.Textbox(
+            label="Subquestion 3 Text",
+            lines=2,
+            placeholder="Third sub-question...",
+            value="N/A",
+        )
+        subquestion_3_answer_input = gr.Textbox(
+            label="Subquestion 3 Answer",
+            lines=2,
+            placeholder="Answer to sub-question 3...",
+            value="N/A",
+        )
+
+    with gr.Row():
+        subquestion_4_text_input = gr.Textbox(
+            label="Subquestion 4 Text",
+            lines=2,
+            placeholder="Fourth sub-question...",
+            value="N/A",
+        )
+        subquestion_4_answer_input = gr.Textbox(
+            label="Subquestion 4 Answer",
+            lines=2,
+            placeholder="Answer to sub-question 4...",
+            value="N/A",
+        )
+
+    with gr.Row():
+        subquestion_5_text_input = gr.Textbox(
+            label="Subquestion 5 Text",
+            lines=2,
+            placeholder="Fifth sub-question...",
+            value="N/A",
+        )
+        subquestion_5_answer_input = gr.Textbox(
+            label="Subquestion 5 Answer",
+            lines=2,
+            placeholder="Answer to sub-question 5...",
+            value="N/A",
+        )
 
     with gr.Row():
         submit_button = gr.Button("Submit")
@@ -441,16 +426,22 @@ with gr.Blocks() as demo:
 
     # On Submit, we call generate_json_files with all relevant fields
     def validate_and_generate(
-        sys_msg,
         nm,
         em,
         inst,
         orp,
+        auth,
         qcats,
         sq1t,
         sq1a,
         sq2t,
         sq2a,
+        sq3t,
+        sq3a,
+        sq4t,
+        sq4a,
+        sq5t,
+        sq5a,
         q,
         fa,
         rt,
@@ -475,6 +466,16 @@ with gr.Blocks() as demo:
             missing_fields.append("Final Answer")
         if not i1:
             missing_fields.append("First Question Image")
+        if not sq1t or not sq1t.strip() or not sq1a or not sq1a.strip():
+            missing_fields.append("First Sub-question and Answer")
+        if not sq2t or not sq2t.strip() or not sq2a or not sq2a.strip():
+            missing_fields.append("Second Sub-question and Answer")
+        if not sq3t or not sq3t.strip() or not sq3a or not sq3a.strip():
+            missing_fields.append("Third Sub-question and Answer")
+        if not sq4t or not sq4t.strip() or not sq4a or not sq4a.strip():
+            missing_fields.append("Fourth Sub-question and Answer")
+        if not sq5t or not sq5t.strip() or not sq5a or not sq5a.strip():
+            missing_fields.append("Fifth Sub-question and Answer")
 
         # If any required fields are missing, return a warning and keep all fields as is
         if missing_fields:
@@ -485,16 +486,22 @@ with gr.Blocks() as demo:
 
         # Only after successful validation, generate files but keep all fields
         results = generate_json_files(
-            sys_msg,
             nm,
             em,
             inst,
             orp,
+            auth,
             qcats,
             sq1t,
             sq1a,
             sq2t,
             sq2a,
+            sq3t,
+            sq3a,
+            sq4t,
+            sq4a,
+            sq5t,
+            sq5a,
             q,
             fa,
             rt,
@@ -515,16 +522,22 @@ with gr.Blocks() as demo:
     submit_button.click(
         fn=validate_and_generate,
         inputs=[
-            system_message_input,
             name_input,
             email_address_input,
             institution_input,
             openreview_profile_input,
+            authorship_input,
             question_categories_input,
             subquestion_1_text_input,
             subquestion_1_answer_input,
             subquestion_2_text_input,
             subquestion_2_answer_input,
+            subquestion_3_text_input,
+            subquestion_3_answer_input,
+            subquestion_4_text_input,
+            subquestion_4_answer_input,
+            subquestion_5_text_input,
+            subquestion_5_answer_input,
             question_input,
             final_answer_input,
             rationale_text_input,
@@ -539,22 +552,29 @@ with gr.Blocks() as demo:
     )
 
     # Clear button functionality
-    def clear_form_fields(sys_msg, name, email, inst, openreview, *args):
-        # Preserve personal info fields
+    def clear_form_fields(name, email, inst, openreview, authorship, *args):
+        # Preserve personal info fields and re-enable submit button
+        gr.Info("Form cleared! Ready for new submission 🔄")
         return [
-            "You are a helpful assistant",  # Reset system message to default
             name,  # Preserve name
             email,  # Preserve email
             inst,  # Preserve institution
-            openreview,  # Preserve OpenReview profile
-            None,  # Clear question categories
-            None,  # Clear subquestion 1 text
-            None,  # Clear subquestion 1 answer
-            None,  # Clear subquestion 2 text
-            None,  # Clear subquestion 2 answer
-            None,  # Clear question
-            None,  # Clear final answer
-            None,  # Clear rationale text
+            openreview,  # Preserve openreview
+            authorship,  # Preserve authorship interest
+            gr.update(value=""),  # Clear question categories
+            gr.update(value="N/A"),  # Reset subquestion 1 text to N/A
+            gr.update(value="N/A"),  # Reset subquestion 1 answer to N/A
+            gr.update(value="N/A"),  # Reset subquestion 2 text to N/A
+            gr.update(value="N/A"),  # Reset subquestion 2 answer to N/A
+            gr.update(value="N/A"),  # Reset subquestion 3 text to N/A
+            gr.update(value="N/A"),  # Reset subquestion 3 answer to N/A
+            gr.update(value="N/A"),  # Reset subquestion 4 text to N/A
+            gr.update(value="N/A"),  # Reset subquestion 4 answer to N/A
+            gr.update(value="N/A"),  # Reset subquestion 5 text to N/A
+            gr.update(value="N/A"),  # Reset subquestion 5 answer to N/A
+            gr.update(value=""),  # Clear question
+            gr.update(value=""),  # Clear final answer
+            gr.update(value=""),  # Clear rationale text
             None,  # Clear image1
             None,  # Clear image2
             None,  # Clear image3
@@ -562,30 +582,35 @@ with gr.Blocks() as demo:
             None,  # Clear rationale image1
             None,  # Clear rationale image2
             None,  # Clear output file urls
-            None,  # Clear output file base64
-            gr.update(interactive=True),  # Re-enable submit button
+            gr.Button(interactive=True),  # Re-enable submit button
         ]
 
     clear_button.click(
         fn=clear_form_fields,
         inputs=[
-            system_message_input,
             name_input,
             email_address_input,
             institution_input,
             openreview_profile_input,
+            authorship_input,
         ],
         outputs=[
-            system_message_input,
             name_input,
             email_address_input,
             institution_input,
             openreview_profile_input,
+            authorship_input,
             question_categories_input,
             subquestion_1_text_input,
             subquestion_1_answer_input,
             subquestion_2_text_input,
             subquestion_2_answer_input,
+            subquestion_3_text_input,
+            subquestion_3_answer_input,
+            subquestion_4_text_input,
+            subquestion_4_answer_input,
+            subquestion_5_text_input,
+            subquestion_5_answer_input,
             question_input,
             final_answer_input,
             rationale_text_input,
@@ -596,7 +621,6 @@ with gr.Blocks() as demo:
             rationale_image1,
             rationale_image2,
             output_file_urls,
-            output_file_base64,
             submit_button,
         ],
     )