switching to docker sdk

Dockerfile

@@ -0,0 +1,20 @@
+FROM python:3.13.5-slim
+
+WORKDIR /app
+
+RUN apt-get update && apt-get install -y \
+    build-essential \
+    curl \
+    git \
+    && rm -rf /var/lib/apt/lists/*
+
+COPY requirements.txt ./
+COPY . .
+
+RUN pip3 install -r requirements.txt
+
+EXPOSE 8501
+
+HEALTHCHECK CMD curl --fail http://localhost:8501/_stcore/health
+
+ENTRYPOINT ["streamlit", "run", "app.py", "--server.port=8501", "--server.address=0.0.0.0"]

README.md

@@ -3,9 +3,10 @@ title: Video Challenge Leaderboard
 emoji: 🏢
 colorFrom: yellow
 colorTo: blue
-sdk: streamlit
-sdk_version: 1.48.1
-app_file: app.py
+sdk: docker
+app_port: 8501
+tags:
+  - streamlit
 pinned: false
 short_description: Leaderboard
 ---

metric.py

@@ -1,8 +1,25 @@
 from collections import defaultdict
+import traceback
 from huggingface_hub import hf_hub_download
 import pandas as pd
 from sklearn.metrics import roc_auc_score, roc_curve
 from typing import Any, Dict
+import numpy as np
+
+
+def check_if_score_ok(df):
+    if df["score"].min() >= .5 and df["score"].max() <= 1.:
+        print("assuming max prob is reported... recomputing")
+        ## assume in this case they are reporting max P(real), P(generated)
+        pred_generated = df["submission_pred"] == "generated"
+        pred_real = df["submission_pred"] == "real"
+
+        df.loc[pred_real, "score"] = 1. - df.loc[pred_real, "score"]
+        not_predicted = ~(pred_generated | pred_real)
+        if not_predicted.any():
+            df.loc[not_predicted,"score"] = np.random.rand(not_predicted.sum())
+
+    return df
 
 
 def compute_auc(df: pd.DataFrame) -> float:
@@ -20,10 +37,13 @@ def compute_auc(df: pd.DataFrame) -> float:
         ## Only one class
         if ytrue.all() or (~ytrue).all():
             return float("nan")
-
+        
+        df = check_if_score_ok(df)
+        
         return roc_auc_score(ytrue, df["score"])
     except Exception as e:
         print(f"AUC exception: {e}")
+        # traceback.print_exc()
         return float("nan")
 
 
@@ -36,9 +56,14 @@ def compute_roc_curve(df: pd.DataFrame, keep_every: int = 10) -> Dict[Any, Any]:
             return {"fpr": [], "tpr": [], "threshold": []}
 
         df = df.loc[~isna]
+
+        df = check_if_score_ok(df)
+
+
         fpr, tpr, threshold = roc_curve(df["pred"] == "generated", df["score"])
         if len(fpr) < keep_every:
             return {"fpr": fpr.tolist(), "tpr": tpr.tolist(), "threshold": threshold.tolist()}
+        
 
         # Sample every keep_every
         return {

requirements.txt

@@ -2,4 +2,7 @@ scikit-learn
 numpy
 streamlit
 huggingface_hub
-pytz
+pytz
+altair
+pandas
+hf_transfer