Update init.py

init.py

@@ -1,4 +1,4 @@
-import os 
+import os
 from constants import EVAL_REQUESTS_PATH
 from pathlib import Path
 from huggingface_hub import HfApi, Repository
@@ -8,67 +8,58 @@ QUEUE_REPO = os.environ.get("QUEUE_REPO")
 QUEUE_PATH = os.environ.get("QUEUE_PATH")
 
 hf_api = HfApi(
-    endpoint="https://huggingface.co", 
-    token=TOKEN_HUB, 
+    endpoint="https://huggingface.co",
+    token=TOKEN_HUB,
 )
-print(TOKEN_HUB)
+
 # Language code for Persian
 PERSIAN_LANGUAGE_CODE = "fa"
 
 def load_all_info_from_dataset_hub():
     eval_queue_repo = None
-    requested_models = None
+    requested_models = []
 
-    passed = True
     if TOKEN_HUB is None:
-        passed = False
-    else:
-        print("Pulling evaluation requests and results.")
+        raise ValueError("No Hugging Face token provided. Skipping evaluation requests and results.")
 
-        eval_queue_repo = Repository(
-            local_dir=QUEUE_PATH,
-            clone_from=QUEUE_REPO,
-            use_auth_token=TOKEN_HUB,
-            repo_type="dataset",
-        )
-        eval_queue_repo.git_pull()
-        
-        # Local directory where dataset repo is cloned + folder with eval requests
-        directory = QUEUE_PATH / EVAL_REQUESTS_PATH
-        requested_models = get_all_requested_models(directory)
-        requested_models = [p.stem for p in requested_models]
+    print("Pulling evaluation requests and results.")
 
-        # Filter models to only include those supporting Persian language
-        requested_models = filter_persian_models(requested_models)
-        
-        # Local directory where dataset repo is cloned
-        csv_results = get_csv_with_results(QUEUE_PATH)
-        if csv_results is None:
-            passed = False
-    if not passed:
-        raise ValueError("No Hugging Face token provided. Skipping evaluation requests and results.")
+    eval_queue_repo = Repository(
+        local_dir=QUEUE_PATH,
+        clone_from=QUEUE_REPO,
+        use_auth_token=TOKEN_HUB,
+        repo_type="dataset",
+    )
+    eval_queue_repo.git_pull()
 
-    return eval_queue_repo, requested_models, csv_results
+    # Local directory where dataset repo is cloned + folder with eval requests
+    directory = QUEUE_PATH / EVAL_REQUESTS_PATH
+    requested_models = get_all_requested_models(directory)
+    requested_models = [p.stem for p in requested_models]
+
+    # Local directory where dataset repo is cloned
+    csv_results = get_csv_with_results(QUEUE_PATH)
+    if csv_results is None:
+        raise ValueError("CSV results file not found.")
 
+    return eval_queue_repo, requested_models, csv_results
 
 def upload_file(requested_model_name, path_or_fileobj):
     dest_repo_file = Path(EVAL_REQUESTS_PATH) / path_or_fileobj.name
     dest_repo_file = str(dest_repo_file)
     hf_api.upload_file(
-            path_or_fileobj=path_or_fileobj,
-            path_in_repo=str(dest_repo_file),
-            repo_id=QUEUE_REPO,
-            token=TOKEN_HUB,
-            repo_type="dataset",
-            commit_message=f"Add {requested_model_name} to eval queue")
-
+        path_or_fileobj=path_or_fileobj,
+        path_in_repo=str(dest_repo_file),
+        repo_id=QUEUE_REPO,
+        token=TOKEN_HUB,
+        repo_type="dataset",
+        commit_message=f"Add {requested_model_name} to eval queue")
 
 def get_all_requested_models(directory):
     directory = Path(directory)
     all_requested_models = list(directory.glob("*.txt"))
     return all_requested_models
 
-
 def get_csv_with_results(directory):
     directory = Path(directory)
     all_csv_files = list(directory.glob("*.csv"))
@@ -77,10 +68,9 @@ def get_csv_with_results(directory):
         return None
     return latest[0]
 
-
-def is_model_on_hub(model_name, revision="main") -> bool:
+def is_model_on_hub(model_name, revision="main") -> (bool, str):
     try:
-        model_name = model_name.replace(" ","")
+        model_name = model_name.replace(" ", "")
         author = model_name.split("/")[0]
         model_id = model_name.split("/")[1]
         if len(author) == 0 or len(model_id) == 0:
@@ -90,7 +80,7 @@ def is_model_on_hub(model_name, revision="main") -> bool:
 
     try:
         models = list(hf_api.list_models(author=author, search=model_id))
-        matched = [model_name for m in models if m.modelId == model_name]
+        matched = [m.modelId for m in models if m.modelId == model_name]
         if len(matched) != 1:
             return False, "was not found on the hub!"
         else:
@@ -98,32 +88,3 @@ def is_model_on_hub(model_name, revision="main") -> bool:
     except Exception as e:
         print(f"Could not get the model from the hub.: {e}")
         return False, "was not found on hub!"
-
-
-def filter_persian_models(model_list):
-    """
-    Filters the provided list of models to include only those that support Persian (fa).
-    
-    Args:
-        model_list (list): List of model names to filter.
-    
-    Returns:
-        list: List of models that support Persian.
-    """
-    persian_models = []
-    for model_name in model_list:
-        try:
-            # Get model information from Hugging Face Hub
-            model_info = hf_api.model_info(model_name)
-            languages = model_info.cardData.get("languages", [])
-            
-            # Check if Persian ('fa') is listed in the model's languages
-            if PERSIAN_LANGUAGE_CODE in languages:
-                persian_models.append(model_name)
-                print(f"{model_name} supports Persian language.")
-            else:
-                print(f"{model_name} does not support Persian language. Skipping.")
-        except Exception as e:
-            print(f"Error fetching model info for {model_name}: {str(e)}")
-    
-    return persian_models