Update app.py

app.py

@@ -9,7 +9,7 @@ df = pd.read_csv(DATA_FILE)
 # Normalize column names
 df.columns = df.columns.str.strip()
 
-# Header interface
+# Page header
 st.title("Russian Router Ranking (RRR) Leaderboard")
 st.markdown("""
 [GitHub Repository](https://github.com/EvilFreelancer/rrr-benchmark)  
@@ -18,38 +18,34 @@ The table shows the accuracy and performance of the models on the
 """)
 
 
+# Function to sort model sizes numerically (e.g., 7b < 13b < 32b, etc.)
 def model_size_sort_key(size: str):
-    """
-    Converts model size (e.g. '7b', '1m') to a numeric key for sorting.
-    'm' = mega (1e6), 'b' = billion (1e9)
-    """
     if not isinstance(size, str):
         return float('inf')
     match = re.match(r"(\d+(?:\.\d+)?)([mb])", size.lower())
     if not match:
-        return float('inf')  # unknown or malformed size
-
+        return float('inf')
     num, unit = match.groups()
     multiplier = 1e6 if unit == 'm' else 1e9
     return float(num) * multiplier
 
 
-# Sidebar filtering
+# Sidebar filters
 with st.sidebar:
     st.header("Filters")
 
-    # Name of model (case-insensitive sort)
+    # Model name filter (case-insensitive sort)
     model_name_options = sorted(df["model_name"].dropna().unique(), key=str.lower)
     model_name = st.multiselect("Select model:", options=model_name_options)
 
-    # Size of model (numeric-aware sort)
+    # Model size filter (numerical sort)
     model_size_options = sorted(df["model_size"].dropna().unique(), key=model_size_sort_key)
     model_size = st.multiselect("Select size:", options=model_size_options)
 
-    # Level of quantization
+    # Quantization level filter (default alphabetical sort)
     model_quant = st.multiselect("Select quantization:", options=sorted(df["model_quant"].dropna().unique()))
 
-# Apply filters
+# Apply filters to the dataset
 filtered_df = df.copy()
 if model_name:
     filtered_df = filtered_df[filtered_df["model_name"].isin(model_name)]
@@ -58,17 +54,45 @@ if model_size:
 if model_quant:
     filtered_df = filtered_df[filtered_df["model_quant"].isin(model_quant)]
 
-# Column formatting for display
+# Format specification for numerical columns
 format_dict = {
     "accuracy":          "{:.2%}".format,
     "avg_response_time": "{:.3f}".format,
     "avg_token_count":   "{:.1f}".format
 }
 
-# Remove model column
-display_df = filtered_df.drop(columns=["model"], errors="ignore")
 
-# Display the table sorted by accuracy in descending order
-st.dataframe(
-    display_df.sort_values(by="accuracy", ascending=False).reset_index(drop=True).style.format(format_dict)
+# Function to render model_name as a clickable link with a tooltip (title)
+def make_clickable_label(row):
+    model_field = row["model"]
+    name = row["model_name"]
+    if model_field.startswith("hf.co/"):
+        url = f"https://{model_field}"
+    else:
+        url = f"https://ollama.com/library/{model_field}"
+    return f'<a href="{url}" title="{model_field}" target="_blank">{name}</a>'
+
+
+# Create new column with HTML links for model_name
+display_df = filtered_df.copy()
+display_df["model_name"] = display_df.apply(make_clickable_label, axis=1)
+
+# Drop 'model' column from display (but keep it for link rendering)
+display_df = display_df.drop(columns=["model"], errors="ignore")
+
+# Apply sorting, formatting, and styling
+styled = (
+    display_df.sort_values(by="accuracy", ascending=False)
+    .reset_index(drop=True)
+    .style
+    .format(format_dict)
+    .set_sticky(axis="index")  # Keep first column visible on scroll
+    .hide(axis="index")  # Hide row index
+    .set_properties(subset=["model_name"], **{"text-align": "left"})  # Align left
+)
+
+# Display table with HTML support for links
+st.write(
+    styled.to_html(escape=False),
+    unsafe_allow_html=True
 )