Fix dark theme view (#4)

- Fix dark theme styling (29bc66e0e2dd9fc094f71b4e079261e9caa88525)

app.py

@@ -5,9 +5,10 @@ import pandas as pd
 
 from src.about import INTRODUCTION_TEXT, TITLE
 from src.display.css_html_js import custom_css
-from src.leaderboard.columns import DisplayColumns, RequiredInputColumns
+from src.leaderboard.columns import DisplayColumns
 from src.leaderboard.input import load_csv_from_github
 from src.leaderboard.output import format_output_df
+from src.leaderboard.styling import apply_styling
 
 LEADERBOARD_GITHUB_URL = "https://github.com/upgini/mle-bench/blob/main/rankings/low/tabular/overall_ranks.csv"
 
@@ -40,41 +41,6 @@ def refresh_leaderboard():
     return df, status
 
 
-def apply_styling(df: pd.DataFrame):
-    """Apply styling to the leaderboard table."""
-
-    if df.empty:
-        return pd.DataFrame(columns=DisplayColumns.values())
-
-    display_df = df[DisplayColumns.values()]
-
-    style = (
-        display_df.style.background_gradient(
-            subset=[DisplayColumns.NORMALIZED_SCORE],
-            high=0.5,
-            low=0.0,
-            cmap="Greens",
-            gmap=df[RequiredInputColumns.MEAN_NORMALIZED_SCORE],
-        )
-        .background_gradient(
-            subset=[DisplayColumns.ANY_MEDAL_SCORE],
-            high=1.2,
-            low=0.0,
-            cmap="Oranges",
-            gmap=df[RequiredInputColumns.MEAN_MEDAL_PCT],
-        )
-        .format(
-            subset=(
-                df[RequiredInputColumns.MEAN_NORMALIZED_SCORE] == df[RequiredInputColumns.MEAN_NORMALIZED_SCORE].max(),
-                DisplayColumns.NORMALIZED_SCORE,
-            ),
-            formatter=lambda x: f"**{x}**",
-        )
-    )
-
-    return style
-
-
 def create_app():
     """Create and configure the Gradio app without launching it."""
     with gr.Blocks(title="Upgini MLE-Bench Leaderboard", css=custom_css) as demo:

requirements.txt

@@ -1,18 +1,15 @@
 APScheduler
 black
 datasets
-gradio
+gradio==5.49.1
 gradio[oauth]
-gradio_leaderboard==0.0.13
 gradio_client
-huggingface-hub>=0.18.0
+gradio_leaderboard==0.0.13
 matplotlib
 numpy
 pandas
+pre-commit
 pytest
-requests
 python-dateutil
+requests
 tqdm
-transformers
-tokenizers>=0.15.0
-sentencepiece

src/leaderboard/styling.py

@@ -0,0 +1,70 @@
+import gradio as gr
+import matplotlib.pyplot as plt
+import numpy as np
+import pandas as pd
+from matplotlib.colors import LinearSegmentedColormap
+
+from .columns import DisplayColumns, RequiredInputColumns
+
+
+def create_light_green_cmap():
+    cmap = plt.cm.get_cmap("Greens")
+    num_colors = 256
+    half_colors = np.linspace(0, 0.5, num_colors)
+    half_cmap = [cmap(val) for val in half_colors]
+    light_green_cmap = LinearSegmentedColormap.from_list("LightGreens", half_cmap, N=256)
+    return light_green_cmap
+
+
+def get_column_types(df: pd.DataFrame, markdown_columns: list[str]) -> list[str]:
+    types = []
+    for column_name in df.columns:
+        if column_name in markdown_columns:
+            types.append("markdown")
+        elif pd.api.types.is_numeric_dtype(df[column_name]):
+            types.append("number")
+        else:
+            types.append("str")
+    return types
+
+
+def apply_styling(df: pd.DataFrame):
+    """Apply styling to the leaderboard table."""
+
+    if df.empty:
+        return pd.DataFrame(columns=DisplayColumns.values())
+
+    display_df = df[DisplayColumns.values()]
+
+    style = (
+        display_df.style.background_gradient(
+            subset=[DisplayColumns.NORMALIZED_SCORE],
+            cmap=create_light_green_cmap(),
+            gmap=df[RequiredInputColumns.MEAN_NORMALIZED_SCORE] * 100,
+        )
+        .background_gradient(
+            subset=[DisplayColumns.ANY_MEDAL_SCORE],
+            cmap=create_light_green_cmap(),
+            gmap=df[RequiredInputColumns.MEAN_MEDAL_PCT],
+            high=1.5,
+        )
+        .highlight_max(subset=[DisplayColumns.NORMALIZED_SCORE], props="font-weight: bold")
+    )
+
+    markdown_columns = [
+        c
+        for c in DisplayColumns.values()
+        if c not in [DisplayColumns.NORMALIZED_SCORE, DisplayColumns.ANY_MEDAL_SCORE]
+    ]
+    column_types = get_column_types(display_df, markdown_columns)
+
+    return gr.DataFrame(
+        value=style,
+        wrap=True,
+        interactive=False,
+        type="pandas",
+        datatype=column_types,
+        label="Leaderboard",
+        elem_id="leaderboard-table",
+        show_search="search",
+    )