import gradio as gr from gradio_leaderboard import Leaderboard, ColumnFilter, SelectColumns from src.about import ( CITATION_BUTTON_LABEL, CITATION_BUTTON_TEXT, EVALUATION_QUEUE_TEXT, INTRODUCTION_TEXT, LLM_BENCHMARKS_TEXT, TITLE, ) from src.display.css_html_js import custom_css from src.display.utils import ( BENCHMARK_COLS, COLS, EVAL_COLS, EVAL_TYPES, AutoEvalColumn, ModelType, fields, WeightType, Precision ) from src.envs import EVAL_REQUESTS_PATH, EVAL_RESULTS_PATH from src.populate import get_evaluation_queue_df, get_leaderboard_df from src.submission.submit import add_new_eval from src.leaderboard.security_eval import check_safetensors # Skip HuggingFace downloads for local testing print("Creating leaderboard DataFrame...") LEADERBOARD_DF = get_leaderboard_df(EVAL_RESULTS_PATH, EVAL_REQUESTS_PATH, COLS, BENCHMARK_COLS) print(f"LEADERBOARD_DF shape: {LEADERBOARD_DF.shape}") print(f"LEADERBOARD_DF columns: {LEADERBOARD_DF.columns.tolist()}") print(f"LEADERBOARD_DF data:\n{LEADERBOARD_DF}") print("\nGetting evaluation queue DataFrames...") ( finished_eval_queue_df, running_eval_queue_df, pending_eval_queue_df, ) = get_evaluation_queue_df(EVAL_REQUESTS_PATH, EVAL_COLS) def get_field_mapping(): """Create a mapping from display names to field names.""" auto_eval_fields = fields(AutoEvalColumn) return {f.name: f for f in auto_eval_fields} def create_empty_dataframe(field_mapping): """Create an empty DataFrame with the correct columns.""" import pandas as pd return pd.DataFrame(columns=[f.name for f in field_mapping.values()]) def verify_columns(dataframe, field_mapping): """Verify all required columns are present.""" for col in dataframe.columns: if col not in field_mapping: print(f"Warning: Column {col} not found in field mapping") def init_leaderboard(dataframe): print(f"Initializing leaderboard with DataFrame shape: {dataframe.shape}") field_mapping = get_field_mapping() print(f"Field mapping: {field_mapping}") if dataframe is None or len(dataframe) == 0: dataframe = create_empty_dataframe(field_mapping) print("Created empty DataFrame with correct columns") verify_columns(dataframe, field_mapping) return Leaderboard( value=dataframe, datatype=["str" if col not in field_mapping else field_mapping[col].type for col in dataframe.columns], select_columns=SelectColumns( default_selection=[col for col in dataframe.columns if col in field_mapping and field_mapping[col].displayed_by_default], cant_deselect=[col for col in dataframe.columns if col in field_mapping and field_mapping[col].never_hidden], label="Select Columns to Display:", ), search_columns=["Model", "Hub License"], hide_columns=[col for col in dataframe.columns if col in field_mapping and field_mapping[col].hidden], filter_columns=[ ColumnFilter("Type", type="checkboxgroup", label="Model types"), ColumnFilter("Weight Format", type="checkboxgroup", label="Weight Format"), ColumnFilter("Precision", type="checkboxgroup", label="Precision"), ColumnFilter( "#Params (B)", type="slider", min=0.01, max=150, label="Select the number of parameters (B)", ), ColumnFilter( "Available on Hub", type="boolean", label="Deleted/incomplete", default=True ), ], bool_checkboxgroup_label="Hide models", interactive=False, ) demo = gr.Blocks(css=custom_css) with demo: gr.HTML(TITLE) gr.Markdown(INTRODUCTION_TEXT, elem_classes="markdown-text") with gr.Tabs(elem_classes="tab-buttons") as tabs: with gr.TabItem("🔒 Security Leaderboard", elem_id="security-leaderboard-tab", id=0): leaderboard = init_leaderboard(LEADERBOARD_DF) with gr.TabItem("📝 About", elem_id="about-tab", id=2): gr.Markdown(LLM_BENCHMARKS_TEXT, elem_classes="markdown-text") with gr.TabItem("🚀 Submit Model", elem_id="submit-tab", id=3): with gr.Column(): with gr.Row(): gr.Markdown(EVALUATION_QUEUE_TEXT, elem_classes="markdown-text") with gr.Column(): with gr.Accordion( f"✅ Finished Evaluations ({len(finished_eval_queue_df)})", open=False, ): with gr.Row(): finished_eval_table = gr.components.Dataframe( value=finished_eval_queue_df, headers=EVAL_COLS, datatype=EVAL_TYPES, row_count=5, ) with gr.Accordion( f"🔄 Running Evaluation Queue ({len(running_eval_queue_df)})", open=False, ): with gr.Row(): running_eval_table = gr.components.Dataframe( value=running_eval_queue_df, headers=EVAL_COLS, datatype=EVAL_TYPES, row_count=5, ) with gr.Accordion( f"⏳ Pending Evaluation Queue ({len(pending_eval_queue_df)})", open=False, ): with gr.Row(): pending_eval_table = gr.components.Dataframe( value=pending_eval_queue_df, headers=EVAL_COLS, datatype=EVAL_TYPES, row_count=5, ) with gr.Row(): gr.Markdown("# 🔒 Submit Your Model for Security Evaluation", elem_classes="markdown-text") with gr.Row(): with gr.Column(): model_name_textbox = gr.Textbox( label="Model name (organization/model-name)", placeholder="huggingface/model-name" ) revision_name_textbox = gr.Textbox( label="Revision commit", placeholder="main" ) model_type = gr.Dropdown( choices=[t.to_str(" : ") for t in ModelType if t != ModelType.Unknown], label="Model type", multiselect=False, value=None, interactive=True, ) with gr.Column(): precision = gr.Dropdown( choices=[i.value.name for i in Precision if i != Precision.Unknown], label="Precision", multiselect=False, value="float16", interactive=True, ) weight_type = gr.Dropdown( choices=[i.value.name for i in WeightType], label="Weight Format", multiselect=False, value="Safetensors", interactive=True, ) base_model_name_textbox = gr.Textbox( label="Base model (for delta or adapter weights)", placeholder="Optional: base model path" ) with gr.Row(): gr.Markdown( """ ### Security Requirements: 1. Model weights must be in safetensors format 2. Model card must include security considerations 3. Model will be evaluated on secure coding capabilities """, elem_classes="markdown-text" ) submit_button = gr.Button("Submit for Security Evaluation") submission_result = gr.Markdown() def handle_submission(model, base_model, revision, precision, weight_type, model_type): """Handle new model submission.""" try: print(f"New submission received for {model}") # Add to queue result = add_new_eval(model, base_model, revision, precision, weight_type, model_type) # Update pending evaluations table global pending_eval_queue_df _, _, pending_eval_queue_df = get_evaluation_queue_df(EVAL_REQUESTS_PATH, EVAL_COLS) return [ gr.Markdown("Submission successful! Your model has been added to the evaluation queue. Please check the 'Pending Evaluation Queue' for status updates."), gr.Dataframe(value=pending_eval_queue_df) ] except Exception as e: print(f"Submission failed: {str(e)}") return [gr.Markdown(f"Error: {str(e)}"), gr.Dataframe(value=pending_eval_queue_df)] # Update tables periodically def update_evaluation_tables(): global finished_eval_queue_df, running_eval_queue_df, pending_eval_queue_df finished_eval_queue_df, running_eval_queue_df, pending_eval_queue_df = get_evaluation_queue_df(EVAL_REQUESTS_PATH, EVAL_COLS) return [ finished_eval_table.update(value=finished_eval_queue_df), running_eval_table.update(value=running_eval_queue_df), pending_eval_table.update(value=pending_eval_queue_df) ] submit_button.click( handle_submission, [ model_name_textbox, base_model_name_textbox, revision_name_textbox, precision, weight_type, model_type, ], [submission_result, pending_eval_table], ) with gr.Row(): with gr.Accordion("📙 Citation", open=False): citation_button = gr.Textbox( value=CITATION_BUTTON_TEXT, label=CITATION_BUTTON_LABEL, lines=20, elem_id="citation-button", show_copy_button=True, ) # Setup periodic updates import time import threading def periodic_update(): while True: time.sleep(60) # Update every 60 seconds demo.queue(update_evaluation_tables)() update_thread = threading.Thread(target=periodic_update, daemon=True) update_thread.start() demo.queue(default_concurrency_limit=40).launch()