hf_hub_download

model_comparison.py

@@ -5,7 +5,7 @@ import plotly.express as px
 from yaml import safe_load
 import user_evaluation_variables
 from pathlib import Path
-from huggingface_hub import snapshot_download, CommitScheduler, login
+from huggingface_hub import hf_hub_download, CommitScheduler, login
 from profanity_check import predict
 
 databaseDF = None
@@ -18,13 +18,13 @@ EVAL_DATABASE_DIR.mkdir(parents=True, exist_ok=True)
 GEN_EVAL_DATABASE_PATH = EVAL_DATABASE_DIR / f"general_eval_database.yaml"
 TASK_EVAL_DATABASE_PATH = EVAL_DATABASE_DIR / f"task_oriented_eval_database.yaml"
 
-EVAL_DATABASE_UPDATE_SCHEDULER = CommitScheduler(
-    repo_id="try-before-you-bias-data",
-    repo_type="dataset",
-    folder_path=EVAL_DATABASE_DIR,
-    path_in_repo="data",
-    every=2,
-)
+# EVAL_DATABASE_UPDATE_SCHEDULER = CommitScheduler(
+#     repo_id="try-before-you-bias-data",
+#     repo_type="dataset",
+#     folder_path=EVAL_DATABASE_DIR,
+#     path_in_repo="data",
+#     every=2,
+# )
 
 # def get_evaluation_id(evalType, debugging):
 #     global GEN_EVAL_DATABASE_PATH
@@ -75,22 +75,22 @@ def dataframe_with_selections(df):
     selected_rows = edited_df[edited_df.Select]
     return selected_rows.drop('Select', axis=1)
 def add_user_evalID_columns_to_df(df, evalDataPath):
-    with EVAL_DATABASE_UPDATE_SCHEDULER.lock:
-        with open(evalDataPath, 'r') as f:
-            yamlData = safe_load(f)
-            for user in yamlData['evaluations']['username']:
-                if df is None:
-                    df = pd.DataFrame(yamlData['evaluations']['username'][user]).T
-                    df.insert(0, "Eval. ID", list(yamlData['evaluations']['username'][user].keys()), True)
-                else:
-                    df = pd.concat([df, pd.DataFrame(yamlData['evaluations']['username'][user]).T],
-                                           ignore_index=True)
-                evalIDIterator = 0
-                for index, row in df.iterrows():
-                    if row['Eval. ID'] is np.nan:
-                        df.loc[index, 'Eval. ID'] = list(yamlData['evaluations']['username'][user].keys())[
-                            evalIDIterator]
-                        evalIDIterator += 1
+    # with open(evalDataPath, 'r') as f:
+    # yamlData = safe_load(f)
+    yamlData = hf_hub_download(repo_id="try-before-you-bias-data", filename=evalDataPath, repo_type="dataset")
+    for user in yamlData['evaluations']['username']:
+        if df is None:
+            df = pd.DataFrame(yamlData['evaluations']['username'][user]).T
+            df.insert(0, "Eval. ID", list(yamlData['evaluations']['username'][user].keys()), True)
+        else:
+            df = pd.concat([df, pd.DataFrame(yamlData['evaluations']['username'][user]).T],
+                                   ignore_index=True)
+        evalIDIterator = 0
+        for index, row in df.iterrows():
+            if row['Eval. ID'] is np.nan:
+                df.loc[index, 'Eval. ID'] = list(yamlData['evaluations']['username'][user].keys())[
+                    evalIDIterator]
+                evalIDIterator += 1
     return df
     
 def initialise_page(tab):