requirements fix and licence warnings

app.py

@@ -21,9 +21,9 @@ def process_file(file):
         return results, None
 
     # Check 2: Croissant validation
-    croissant_valid, croissant_message = validate_croissant(json_data)
+    croissant_valid, croissant_message, croissant_status = validate_croissant(json_data)
     croissant_message = croissant_message.replace("\n✓\n", "\n")
-    results.append(("Croissant Schema Validation", croissant_valid, croissant_message, "pass" if croissant_valid else "error"))
+    results.append(("Croissant Schema Validation", croissant_valid, croissant_message, croissant_status))
 
     if not croissant_valid:
         return results, None
@@ -43,7 +43,7 @@ def process_file(file):
     return results, report
 
 def create_ui():
-    with gr.Blocks(theme=gr.themes.Soft()) as app:
+    with gr.Blocks() as app:
         gr.HTML("<p align='center'><img src='https://upload.wikimedia.org/wikipedia/en/0/08/Logo_for_Conference_on_Neural_Information_Processing_Systems.svg' alt='NeurIPS Logo' width='400'/></p>")
         gr.Markdown("# 🥐 Croissant Validator for NeurIPS")
         gr.Markdown("""
@@ -85,7 +85,7 @@ def create_ui():
         # Now create the validation results section in a separate group
         with gr.Group():
             # Validation results
-            validation_results = gr.HTML(visible=False)            
+            validation_results = gr.HTML(value="", visible=True)            
             validation_progress = gr.HTML(visible=False)
 
             # Collapsible report section
@@ -99,7 +99,6 @@ def create_ui():
                     report_text = gr.Textbox(
                         label="Report Content", 
                         visible=True,
-                        show_copy_button=True,
                         lines=10,
                         elem_id="report-text-box"
                     )
@@ -107,10 +106,39 @@ def create_ui():
         # Define CSS for the validation UI
         gr.HTML("""
         <style>
+        /* Import Google Sans from Google Fonts */
+        @import url('https://fonts.googleapis.com/css2?family=Google+Sans:wght@400;500;600&display=swap');
+
         /* Set max width and center the app */
         .gradio-container {
             max-width: 750px !important;
             margin: 0 auto !important;
+            font-family: 'Google Sans', -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, sans-serif !important;
+            font-size: 16px !important;
+        }
+
+        .gradio-container * {
+            font-family: 'Google Sans', -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, sans-serif !important;
+        }
+
+        /* Breathing room inside tabs and groups */
+        .gradio-container .tab-nav {
+            gap: 6px !important;
+        }
+        .gradio-container .tabs {
+            padding-top: 14px !important;
+        }
+        .gradio-container .tab-nav button {
+            font-size: 15px !important;
+            padding: 12px 22px !important;
+        }
+        .gradio-container .tabitem button[class*="primary"] {
+            margin-top: 12px !important;
+            margin-bottom: 4px !important;
+        }
+        .gradio-container label, .gradio-container .label-wrap {
+            font-size: 15px !important;
+            margin-bottom: 6px !important;
         }
 
         /* Make basic containers transparent */
@@ -297,7 +325,7 @@ def create_ui():
                 return [
                     "upload",
                     """<div class="progress-status">Ready for upload</div>""",
-                    gr.update(visible=False),
+                    gr.update(value=""),  # Clear validation results
                     gr.update(visible=False),  # Hide report group
                     None,  # Clear report text
                     None,  # Clear report file
@@ -308,7 +336,7 @@ def create_ui():
                 return [
                     "url",
                     """<div class="progress-status">Enter a URL to fetch</div>""",
-                    gr.update(visible=False),
+                    gr.update(value=""),  # Clear validation results
                     gr.update(visible=False),  # Hide report group
                     None,  # Clear report text
                     None,  # Clear report file
@@ -329,7 +357,7 @@ def create_ui():
             if file is None:
                 return [
                     """<div class="progress-status">Ready for upload</div>""",
-                    gr.update(visible=False),
+                    gr.update(value=""),  # Clear validation results
                     gr.update(visible=False),  # Hide report group
                     None,  # Clear report text
                     None   # Clear report file
@@ -337,7 +365,7 @@ def create_ui():
 
             return [
                 """<div class="progress-status">✅ File uploaded successfully</div>""",
-                gr.update(visible=False),
+                gr.update(value=""),  # Clear validation results
                 gr.update(visible=False),  # Hide report group
                 None,  # Clear report text
                 None   # Clear report file
@@ -347,7 +375,7 @@ def create_ui():
             if not url:
                 return [
                     """<div class="progress-status">Please enter a URL</div>""",
-                    gr.update(visible=False),
+                    gr.update(value=""),
                     gr.update(visible=False),
                     None,
                     None
@@ -365,8 +393,8 @@ def create_ui():
                 results = []
                 results.append(("JSON Format Validation", True, "The URL returned valid JSON."))
                 
-                croissant_valid, croissant_message = validate_croissant(json_data)
-                results.append(("Croissant Schema Validation", croissant_valid, croissant_message))
+                croissant_valid, croissant_message, croissant_status = validate_croissant(json_data)
+                results.append(("Croissant Schema Validation", croissant_valid, croissant_message, croissant_status))
                 
                 if not croissant_valid:
                     return [
@@ -404,7 +432,7 @@ def create_ui():
                 error_message = f"Error fetching URL: {str(e)}"
                 return [
                     f"""<div class="progress-status">{error_message}</div>""",
-                    gr.update(visible=False),
+                    gr.update(value=""),
                     gr.update(visible=False),
                     None,
                     None
@@ -413,7 +441,7 @@ def create_ui():
                 error_message = f"URL did not return valid JSON: {str(e)}"
                 return [
                     f"""<div class="progress-status">{error_message}</div>""",
-                    gr.update(visible=False),
+                    gr.update(value=""),
                     gr.update(visible=False),
                     None,
                     None
@@ -422,7 +450,7 @@ def create_ui():
                 error_message = f"Unexpected error: {str(e)}"
                 return [
                     f"""<div class="progress-status">{error_message}</div>""",
-                    gr.update(visible=False),
+                    gr.update(value=""),
                     gr.update(visible=False),
                     None,
                     None
@@ -442,11 +470,14 @@ def create_ui():
                     status_class = "status-success"
                     status_icon = "✓"
                     message_with_emoji = "✅ " + message
-                elif status == "warning" and "Records" in test_name:
+                elif status == "warning":
                     status_class = "status-warning"
                     status_icon = "?"
-                    message_with_emoji = "⚠️ Could not automatically generate records. This is oftentimes not an issue (e.g. datasets could be too large or too complex), and it's not required to pass this test to submit to NeurIPS.\n\n" + message
-                else:  # error or non-records warning
+                    if "Records" in test_name:
+                        message_with_emoji = "⚠️ Could not automatically generate records. This is oftentimes not an issue (e.g. datasets could be too large or too complex), and it's not required to pass this test to submit to NeurIPS.\n\n" + message
+                    else:
+                        message_with_emoji = "⚠️ " + message
+                else:  # error
                     status_class = "status-error"
                     status_icon = "✗"
                     message_with_emoji = "❌ " + message
@@ -482,13 +513,29 @@ def create_ui():
 
         def on_validate(file):
             if file is None:
-                return [
-                    gr.update(visible=False),  # validation_results
+                yield [
+                    gr.update(value=""),  # validation_results
                     gr.update(visible=False),  # validation_progress
                     gr.update(visible=False),  # report_group
                     None,  # report_text
                     None   # report_md
                 ]
+                return
+
+            # Show progress spinner
+            progress_html = """
+            <div class="validation-progress">
+                <div class="loading-spinner"></div>
+                <span>Validating file...</span>
+            </div>
+            """
+            yield [
+                gr.update(value=""),  # validation_results
+                gr.update(visible=True, value=progress_html),  # validation_progress
+                gr.update(visible=False),  # report_group
+                None,  # report_text
+                None   # report_md
+            ]
 
             # Process the file and get results
             results, report = process_file(file)
@@ -508,7 +555,7 @@ def create_ui():
                     f.write(report)
 
             # Return final state
-            return [
+            yield [
                 build_results_html(results),  # validation_results
                 gr.update(visible=False),  # validation_progress
                 gr.update(visible=True) if report else gr.update(visible=False),  # report_group
@@ -545,11 +592,6 @@ def create_ui():
             ]
 
         validate_btn.click(
-            fn=show_progress,
-            inputs=None,
-            outputs=[validation_results, validation_progress, report_group, report_text, report_md],
-            queue=False
-        ).then(
             fn=on_validate,
             inputs=file_input,
             outputs=[validation_results, validation_progress, report_group, report_text, report_md]
@@ -564,7 +606,7 @@ def create_ui():
         # Footer
         gr.HTML("""
         <div style="text-align: center; margin-top: 20px;">
-            <p>Learn more about <a href="https://github.com/mlcommons/croissant" target="_blank">Croissant</a>.</p>
+            <p>Learn more about  🥐<a href="https://github.com/mlcommons/croissant" target="_blank">Croissant</a>.</p>
         </div>
         """)
 
@@ -584,4 +626,4 @@ def create_ui():
 
 if __name__ == "__main__":
     app = create_ui()
-    app.launch()
+    app.launch(theme=gr.themes.Soft())

validation.py

@@ -51,19 +51,30 @@ def validate_json(file_path):
         error_message = f"Error reading file: {str(e)}"
         return False, error_message, None
 
+REQUIRED_SCHEMA_FIELDS = ["license"]
+
 def validate_croissant(json_data):
     """Validate that the JSON follows Croissant schema."""
     try:
         dataset = mlc.Dataset(jsonld=json_data)
-        return True, "The dataset passes Croissant validation."
+        missing = [f for f in REQUIRED_SCHEMA_FIELDS if not json_data.get(f)]
+        if missing:
+            return True, (
+                "The dataset passes Croissant schema validation, but the `license` field is missing. "
+                "Please add a `license` field to your Croissant file. "
+                "Croissant recommends using the URL of a known license, e.g. one of the licenses listed at "
+                "<a href='https://spdx.org/licenses/' target='_blank'>https://spdx.org/licenses/</a>. "
+                "Note that if your paper is accepted at NeurIPS, a permissive license will be required."
+            ), "warning"
+        return True, "The dataset passes Croissant validation.", "pass"
     except mlc.ValidationError as e:
         error_details = traceback.format_exc()
         error_message = f"Validation failed: {str(e)}\n\n{error_details}"
-        return False, error_message
+        return False, error_message, "error"
     except Exception as e:
         error_details = traceback.format_exc()
         error_message = f"Unexpected error during validation: {str(e)}\n\n{error_details}"
-        return False, error_message
+        return False, error_message, "error"
     
 def try_generate_record(record_collection):
     try: