Update app.py

app.py

@@ -53,9 +53,6 @@ except Exception:
     restart_space()
 
 LEADERBOARD_DF = get_leaderboard_df(EVAL_RESULTS_PATH, EVAL_REQUESTS_PATH, COLS, BENCHMARK_COLS)
-print("Initial LEADERBOARD_DF:")
-print(LEADERBOARD_DF.head())
-print(f"LEADERBOARD_DF shape: {LEADERBOARD_DF.shape}")
 original_df = LEADERBOARD_DF
 leaderboard_df = original_df.copy()
 (
@@ -66,11 +63,42 @@ leaderboard_df = original_df.copy()
 ) = get_evaluation_queue_df(EVAL_REQUESTS_PATH, EVAL_COLS)
 
 # Searching and filtering
+# def update_table(
+#     hidden_df: pd.DataFrame,
+#     columns: list,
+#     type_query: list,
+#     precision_query: str,
+#     size_query: list,
+#     add_special_tokens_query: list,
+#     num_few_shots_query: list,
+#     show_deleted: bool,
+#     show_merges: bool,
+#     show_flagged: bool,
+#     query: str,
+# ):
+#     print(f"Update table called with: type_query={type_query}, precision_query={precision_query}, size_query={size_query}")
+#     print(f"hidden_df shape before filtering: {hidden_df.shape}")
+    
+#     filtered_df = filter_models(hidden_df, type_query, size_query, precision_query, add_special_tokens_query, num_few_shots_query, show_deleted, show_merges, show_flagged)
+#     print(f"filtered_df shape after filter_models: {filtered_df.shape}")
+    
+#     filtered_df = filter_queries(query, filtered_df)
+#     print(f"filtered_df shape after filter_queries: {filtered_df.shape}")
+    
+#     print(f"Filter applied: query={query}, columns={columns}, type_query={type_query}, precision_query={precision_query}")
+#     print("Filtered dataframe head:")
+#     print(filtered_df.head())
+
+#     df = select_columns(filtered_df, columns)
+#     print(f"Final df shape: {df.shape}")
+#     print("Final dataframe head:")
+#     print(df.head())
+#     return df
 def update_table(
     hidden_df: pd.DataFrame,
     columns: list,
     type_query: list,
-    precision_query: str,
+    precision_query: list,
     size_query: list,
     add_special_tokens_query: list,
     num_few_shots_query: list,
@@ -78,24 +106,17 @@ def update_table(
     show_merges: bool,
     show_flagged: bool,
     query: str,
+    architecture_query: list,
+    license_query: list
 ):
-    print(f"Update table called with: type_query={type_query}, precision_query={precision_query}, size_query={size_query}")
-    print(f"hidden_df shape before filtering: {hidden_df.shape}")
-    
-    filtered_df = filter_models(hidden_df, type_query, size_query, precision_query, add_special_tokens_query, num_few_shots_query, show_deleted, show_merges, show_flagged)
-    print(f"filtered_df shape after filter_models: {filtered_df.shape}")
-    
+    filtered_df = filter_models(
+        hidden_df, type_query, size_query, precision_query,
+        add_special_tokens_query, num_few_shots_query,
+        show_deleted, show_merges, show_flagged,
+        architecture_query, license_query
+    )
     filtered_df = filter_queries(query, filtered_df)
-    print(f"filtered_df shape after filter_queries: {filtered_df.shape}")
-    
-    print(f"Filter applied: query={query}, columns={columns}, type_query={type_query}, precision_query={precision_query}")
-    print("Filtered dataframe head:")
-    print(filtered_df.head())
-
     df = select_columns(filtered_df, columns)
-    print(f"Final df shape: {df.shape}")
-    print("Final dataframe head:")
-    print(df.head())
     return df
 
 
@@ -140,17 +161,58 @@ def filter_queries(query: str, filtered_df: pd.DataFrame):
     return filtered_df
 
 
+# def filter_models(
+#     df: pd.DataFrame, type_query: list, size_query: list, precision_query: list, add_special_tokens_query: list, num_few_shots_query: list, show_deleted: bool, show_merges: bool, show_flagged: bool
+# ) -> pd.DataFrame:
+#     print(f"Initial df shape: {df.shape}")
+#     print(f"Initial df content:\n{df}")
+
+#     filtered_df = df
+
+#     # Model Type フィルタリング
+#     type_emoji = [t.split()[0] for t in type_query]
+#     filtered_df = filtered_df[filtered_df['T'].isin(type_emoji)]
+#     print(f"After type filter: {filtered_df.shape}")
+
+#     # Precision フィルタリング
+#     filtered_df = filtered_df[filtered_df['Precision'].isin(precision_query + ['Unknown', '?'])]
+#     print(f"After precision filter: {filtered_df.shape}")
+
+#     # Model Size フィルタリング
+#     if 'Unknown' in size_query:
+#         size_mask = filtered_df['#Params (B)'].isna() | (filtered_df['#Params (B)'] == 0)
+#     else:
+#         size_mask = filtered_df['#Params (B)'].apply(lambda x: any(x in NUMERIC_INTERVALS[s] for s in size_query if s != 'Unknown'))
+#     filtered_df = filtered_df[size_mask]
+#     print(f"After size filter: {filtered_df.shape}")
+
+#     # Add Special Tokens フィルタリング
+#     filtered_df = filtered_df[filtered_df['Add Special Tokens'].isin(add_special_tokens_query + ['Unknown', '?'])]
+#     print(f"After add_special_tokens filter: {filtered_df.shape}")
+
+#     # Num Few Shots フィルタリング
+#     filtered_df = filtered_df[filtered_df['Few-shot'].astype(str).isin([str(x) for x in num_few_shots_query] + ['Unknown', '?'])]
+#     print(f"After num_few_shots filter: {filtered_df.shape}")
+
+#     # Show deleted models フィルタリング
+#     if not show_deleted:
+#         filtered_df = filtered_df[filtered_df['Available on the hub'] == True]
+#     print(f"After show_deleted filter: {filtered_df.shape}")
+
+#     print("Filtered dataframe head:")
+#     print(filtered_df.head())
+#     return filtered_df
 def filter_models(
-    df: pd.DataFrame, type_query: list, size_query: list, precision_query: list, add_special_tokens_query: list, num_few_shots_query: list, show_deleted: bool, show_merges: bool, show_flagged: bool
+    df: pd.DataFrame, type_query: list, size_query: list, precision_query: list,
+    add_special_tokens_query: list, num_few_shots_query: list,
+    show_deleted: bool, show_merges: bool, show_flagged: bool,
+    architecture_query: list, license_query: list
 ) -> pd.DataFrame:
     print(f"Initial df shape: {df.shape}")
-    print(f"Initial df content:\n{df}")
-
-    filtered_df = df
 
     # Model Type フィルタリング
     type_emoji = [t.split()[0] for t in type_query]
-    filtered_df = filtered_df[filtered_df['T'].isin(type_emoji)]
+    filtered_df = df[df['T'].isin(type_emoji)]
     print(f"After type filter: {filtered_df.shape}")
 
     # Precision フィルタリング
@@ -161,7 +223,7 @@ def filter_models(
     if 'Unknown' in size_query:
         size_mask = filtered_df['#Params (B)'].isna() | (filtered_df['#Params (B)'] == 0)
     else:
-        size_mask = filtered_df['#Params (B)'].apply(lambda x: any(x in NUMERIC_INTERVALS[s] for s in size_query if s != 'Unknown'))
+        size_mask = filtered_df['#Params (B)'].apply(lambda x: any(pd.Interval(NUMERIC_INTERVALS[s].left, NUMERIC_INTERVALS[s].right).contains(x) for s in size_query if s != 'Unknown'))
     filtered_df = filtered_df[size_mask]
     print(f"After size filter: {filtered_df.shape}")
 
@@ -173,6 +235,16 @@ def filter_models(
     filtered_df = filtered_df[filtered_df['Few-shot'].astype(str).isin([str(x) for x in num_few_shots_query] + ['Unknown', '?'])]
     print(f"After num_few_shots filter: {filtered_df.shape}")
 
+    # Architecture フィルタリング
+    if architecture_query:
+        filtered_df = filtered_df[filtered_df['Architecture'].isin(architecture_query)]
+    print(f"After architecture filter: {filtered_df.shape}")
+
+    # License フィルタリング
+    if license_query:
+        filtered_df = filtered_df[filtered_df['Hub License'].isin(license_query)]
+    print(f"After license filter: {filtered_df.shape}")
+
     # Show deleted models フィルタリング
     if not show_deleted:
         filtered_df = filtered_df[filtered_df['Available on the hub'] == True]