Update app.py

app.py

@@ -63,42 +63,11 @@ 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: list,
+    precision_query: str,
     size_query: list,
     add_special_tokens_query: list,
     num_few_shots_query: list,
@@ -106,17 +75,24 @@ def update_table(
     show_merges: bool,
     show_flagged: bool,
     query: str,
-    architecture_query: list,
-    license_query: list
 ):
-    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
-    )
+    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
 
 
@@ -129,23 +105,16 @@ def search_table(df: pd.DataFrame, query: str) -> pd.DataFrame:
     return df[(df[AutoEvalColumn.dummy.name].str.contains(query, case=False))]
 
 
-# def select_columns(df: pd.DataFrame, columns: list) -> pd.DataFrame:
-#     always_here_cols = [
-#         AutoEvalColumn.model_type_symbol.name,
-#         AutoEvalColumn.model.name,
-#     ]
-#     # We use COLS to maintain sorting
-#     filtered_df = df[
-#         always_here_cols + [c for c in COLS if c in df.columns and c in columns]# + [AutoEvalColumn.dummy.name]
-#     ]
-#     return filtered_df
 def select_columns(df: pd.DataFrame, columns: list) -> pd.DataFrame:
     always_here_cols = [
         AutoEvalColumn.model_type_symbol.name,
         AutoEvalColumn.model.name,
     ]
-    selected_cols = always_here_cols + [c for c in columns if c in df.columns]
-    return df[selected_cols]
+    # We use COLS to maintain sorting
+    filtered_df = df[
+        always_here_cols + [c for c in COLS if c in df.columns and c in columns]# + [AutoEvalColumn.dummy.name]
+    ]
+    return filtered_df
 
 
 def filter_queries(query: str, filtered_df: pd.DataFrame):
@@ -168,58 +137,17 @@ 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,
-    architecture_query: list, license_query: list
+    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 = df[df['T'].isin(type_emoji)]
+    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
@@ -230,7 +158,7 @@ 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(pd.Interval(NUMERIC_INTERVALS[s].left, NUMERIC_INTERVALS[s].right).contains(x) for s in size_query if s != 'Unknown'))
+        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}")
 
@@ -242,16 +170,6 @@ 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]