Added new interim scoring message, removed extra print

.pre-commit-config.yaml

@@ -1,4 +1,4 @@
-exclude: '^.*\.(ipynb|json)$'
+exclude: '^.*\.(ipynb|json|pdb)$'
 repos:
   - repo: https://github.com/pre-commit/pre-commit-hooks
     rev: v5.0.0

about.py

@@ -16,7 +16,7 @@ WEBSITE_HEADER = f"""
     You can **predict any or all of the 5 properties**, and you can filter the main leaderboard by property.
     See more details in the "{ABOUT_TAB_NAME}" tab.
 
-    🗓️ There will be a test set scoring on **October 13th** (which will score all the latest test set submissions at that point). 
+    🗓️ There will be a test set scoring on **October 13th** (which will score all the latest test set submissions at that point).
     Use this to refine your models before the final submission deadline on **1 November 2025**.
 """
 
@@ -42,7 +42,7 @@ Here we invite the community to submit and develop better predictors, which will
 #### 🏆 Prizes
 
 For each of the 5 properties in the competition, there is a prize for the model with the highest performance for that property on the private test set.
-There is also an 'open-source' prize for the best reproducible model: one that is trained on the GDPa1 dataset (reporting cross-validation results) and assessed on the private test set where authors provide all training code and data. 
+There is also an 'open-source' prize for the best reproducible model: one that is trained on the GDPa1 dataset (reporting cross-validation results) and assessed on the private test set where authors provide all training code and data.
 This will be judged by a panel (i.e. by default the model with the highest average Spearman correlation across all properties will be selected, but a really good model on just one property may be better for the community).
 
 For each of these 6 prizes, participants have the choice between
@@ -192,7 +192,7 @@ You do **not** need to predict all 5 properties — each property has its own le
    - Include the `"hierarchical_cluster_IgG_isotype_stratified_fold"` column if submitting cross-validation predictions.
 3. You can resubmit as often as you like; only your latest submission will count for both the leaderboard and final test set scoring.
 
-The GDPa1 results should appear on the leaderboard within a minute, and can also be calculated manually using average Spearman rank correlation across the 5 folds. 
+The GDPa1 results should appear on the leaderboard within a minute, and can also be calculated manually using average Spearman rank correlation across the 5 folds.
 
 ## Cross-validation
 

app.py

@@ -11,6 +11,7 @@ load_dotenv()  # Load environment variables from .env file (before imports)
 from about import ABOUT_INTRO, ABOUT_TEXT, FAQS, SUBMIT_INSTRUCTIONS, WEBSITE_HEADER
 from constants import (
     ASSAY_RENAME,  # noqa: F401
+    FIRST_DEADLINE,
     SEQUENCES_FILE_DICT,
     LEADERBOARD_DISPLAY_COLUMNS,
     ABOUT_TAB_NAME,
@@ -38,8 +39,21 @@ def format_leaderboard_table(df_results: pd.DataFrame, assay: str | None = None)
     # Note: We can also just say the following as a text box at the bottom of the leaderboard: "Note: Results for the Heldout Test Set are only evaluated at competition close"
     # Convert spearman column to string to avoid dtype incompatibility when assigning text
     df["spearman"] = df["spearman"].astype(str)
+    # Cast submission_time to datetime
+    df["submission_time"] = pd.to_datetime(df["submission_time"], errors="coerce")
+    # Before the first deadline: Say we're busy evaluating
     df.loc[
-        (df["dataset"] == "Heldout Test Set") & (df["spearman"] == "nan"), "spearman"
+        (df["dataset"] == "Heldout Test Set")
+        & (df["spearman"] == "nan")
+        & (df["submission_time"] <= FIRST_DEADLINE),
+        "spearman",
+    ] = "Busy evaluating first deadline"
+    # After the first deadline: Evaluated at competition close
+    df.loc[
+        (df["dataset"] == "Heldout Test Set")
+        & (df["spearman"] == "nan")
+        & (df["submission_time"] > FIRST_DEADLINE),
+        "spearman",
     ] = "N/A, evaluated at competition close"
 
     # Finally, rename columns for readability
@@ -64,6 +78,7 @@ def get_leaderboard_object(assay: str | None = None):
         filter_columns=LEADERBOARD_COLUMNS_RENAME_LIST(filter_columns),
         every=15,
         render=True,
+        height=500,  # Set a fixed height to make it scrollable
     )
     return lb
 
@@ -87,9 +102,7 @@ with gr.Blocks(theme=gr.themes.Default(text_size=sizes.text_lg)) as demo:
 
     with gr.Row():
         with gr.Column(scale=6):  # bigger text area
-            gr.Markdown(
-                WEBSITE_HEADER
-            )
+            gr.Markdown(WEBSITE_HEADER)
         with gr.Column(scale=2):  # smaller side column for logo
             gr.Image(
                 value="./assets/competition_logo.jpg",
@@ -232,7 +245,6 @@ with gr.Blocks(theme=gr.themes.Default(text_size=sizes.text_lg)) as demo:
                     if isinstance(answer, list):
                         # Italicize each line
                         italicized_answer = "  \n".join(f"*{item}*" for item in answer)
-                        print(italicized_answer)
                         gr.Markdown(italicized_answer)
                     else:
                         gr.Markdown(f"*{answer}*")  # Italics for answers

constants.py

@@ -4,6 +4,7 @@ Constants for the Antibody Developability Benchmark
 
 import os
 from huggingface_hub import HfApi
+import pandas as pd
 
 ASSAY_LIST = ["AC-SINS_pH7.4", "PR_CHO", "HIC", "Tm2", "Titer"]
 ASSAY_RENAME = {
@@ -99,3 +100,6 @@ BASELINE_USERNAMES = ["loodvanniekerkginkgo"]
 
 def LEADERBOARD_COLUMNS_RENAME_LIST(columns: list[str]) -> list[str]:
     return list(map(lambda x: LEADERBOARD_COLUMNS_RENAME.get(x, x), columns))
+
+# First deadline: 2025-10-14 23:59:59 EST
+FIRST_DEADLINE = pd.to_datetime("2025-10-14 23:59:59").tz_localize("US/Eastern")

submit.py

@@ -41,7 +41,9 @@ def upload_submission(
     }
     with tempfile.NamedTemporaryFile(mode="w", suffix=".json", delete=False) as tmp:
         # json.dump(record, tmp, indent=2)
-        json.dump(record, tmp)  # Note: No indent because indents and large text contents cause this error: https://github.com/huggingface/datasets/issues/3227
+        json.dump(
+            record, tmp
+        )  # Note: No indent because indents and large text contents cause this error: https://github.com/huggingface/datasets/issues/3227
         tmp.flush()
         tmp_name = tmp.name