import gradio as gr
import json
import time
import traceback
from validation import validate_json, validate_croissant, validate_records, generate_validation_report
import requests
def process_file(file):
results = []
json_data = None
# Check 1: JSON validation
json_valid, json_message, json_data = validate_json(file.name)
results.append(("JSON Format Validation", json_valid, json_message))
if not json_valid:
return results, None
# Check 2: Croissant validation
croissant_valid, croissant_message = validate_croissant(json_data)
results.append(("Croissant Schema Validation", croissant_valid, croissant_message))
if not croissant_valid:
return results, None
# Check 3: Records validation
records_valid, records_message = validate_records(json_data)
results.append(("Records Generation Test", records_valid, records_message))
# Generate detailed report
report = generate_validation_report(file.name, json_data, results)
return results, report
def create_ui():
with gr.Blocks(theme=gr.themes.Soft()) as app:
gr.Markdown("# 🔎🥐 Croissant Validator for NeurIPS D&B")
gr.Markdown("""
Upload your Croissant JSON-LD file or enter a URL to validate if it meets the requirements for NeurIPS submission.
The validator will check:
1. If the file is valid JSON
2. If it passes Croissant schema validation
3. If records can be generated within a reasonable time
""")
# Track the active tab for conditional UI updates
active_tab = gr.State("upload") # Default to upload tab
# Create a container for the entire input section
with gr.Group():
# Input tabs
with gr.Tabs() as tabs:
with gr.TabItem("Upload File", id="upload_tab"):
file_input = gr.File(label="Upload Croissant JSON-LD File", file_types=[".json", ".jsonld"])
validate_btn = gr.Button("Validate Uploaded File", variant="primary")
with gr.TabItem("URL Input", id="url_tab"):
url_input = gr.Textbox(
label="Enter Croissant JSON-LD URL",
placeholder="e.g. https://huggingface.co/api/datasets/facebook/natural_reasoning/croissant"
)
fetch_btn = gr.Button("Fetch and Validate", variant="primary")
# Change initial message to match upload tab
upload_progress = gr.HTML(
"""Ready for upload
""",
visible=True)
# Now create the validation results section in a separate group
with gr.Group():
# Validation results
validation_results = gr.HTML(visible=False)
# Report button and options in an expandable group
with gr.Group(visible=False) as report_group:
gr.Markdown("### Validation Report")
with gr.Row():
copy_btn = gr.Button("📋 Copy to Clipboard")
download_btn = gr.Button("⬇️ Download as .md")
report_text = gr.Textbox(visible=False) # Hidden textbox to store report content
# Define CSS for the validation UI
gr.HTML("""
""")
# Update helper messages based on tab changes
def on_tab_change(evt: gr.SelectData):
tab_id = evt.value
if tab_id == "Upload File":
return "upload", """Ready for upload
""", gr.update(visible=False)
else:
return "url", """Enter a URL to fetch
""", gr.update(visible=False)
def on_copy_click(report):
return report
def on_download_click(report, file_name):
report_file = f"report_{file_name}.md"
with open(report_file, "w") as f:
f.write(report)
return report_file
def on_file_upload(file):
if file is None:
return [
"""Ready for upload
""",
gr.update(visible=False),
gr.update(visible=False),
None
]
return [
"""✅ File uploaded successfully
""",
gr.update(visible=False),
gr.update(visible=False),
None
]
def fetch_from_url(url):
if not url:
return """Please enter a URL
""", gr.update(visible=False)
try:
# Fetch JSON from URL
response = requests.get(url, timeout=10)
response.raise_for_status()
json_data = response.json()
# Show success message
progress_html = """✅ JSON fetched successfully from URL
"""
# Validate the fetched JSON
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))
if not croissant_valid:
return progress_html, build_results_html(results)
records_valid, records_message = validate_records(json_data)
results.append(("Records Generation Test", records_valid, records_message))
return progress_html, build_results_html(results)
except requests.exceptions.RequestException as e:
error_message = f"❌ Error fetching URL: {str(e)}"
return f"""{error_message}
""", gr.update(visible=False)
except json.JSONDecodeError as e:
error_message = f"❌ URL did not return valid JSON: {str(e)}"
return f"""{error_message}
""", gr.update(visible=False)
except Exception as e:
error_message = f"❌ Unexpected error: {str(e)}"
return f"""{error_message}
""", gr.update(visible=False)
def build_results_html(results):
# Build validation results HTML
html = ''
for i, (test_name, passed, message) in enumerate(results):
status_class = "status-success" if passed else "status-error"
status_icon = "✓" if passed else "✗"
html += f'''
'''
html += '