update pr

pr_paper_central_tab.py

@@ -9,32 +9,31 @@ import requests
 
 # PR function remains the same
 def create_pr_in_hf_dataset(new_entry, oauth_token: gr.OAuthToken):
-    # Dataset and filename
+    # Dataset and directory
     REPO_ID = 'IAMJB/paper-central-pr'
-    FILENAME = 'data.json'
+    DATA_DIR = 'data'
 
     # Initialize HfApi
     api = HfApi()
     token = oauth_token.token
 
-    # Ensure the repository exists and has an initial empty data.json if not present
+    # Ensure the repository exists and the data directory is initialized
     try:
         # Create the repository if it doesn't exist
         api.create_repo(repo_id=REPO_ID, token=token, repo_type='dataset', exist_ok=True)
 
-        # Check if data.json exists; if not, create it with an empty list
+        # Check if the data directory exists; if not, create it
         files = api.list_repo_files(REPO_ID, repo_type='dataset', token=token)
-        if FILENAME not in files:
-            # Initialize with empty list
-            empty_data = []
-            temp_filename = 'temp_data.json'
+        if DATA_DIR not in files:
+            # Create an empty directory (you can add a .gitkeep file)
+            temp_filename = 'temp_gitkeep'
             with open(temp_filename, 'w') as f:
-                json.dump(empty_data, f)
-            commit = CommitOperationAdd(path_in_repo=FILENAME, path_or_fileobj=temp_filename)
+                pass  # Empty file
+            commit = CommitOperationAdd(path_in_repo=f"{DATA_DIR}/.gitkeep", path_or_fileobj=temp_filename)
             api.create_commit(
                 repo_id=REPO_ID,
                 operations=[commit],
-                commit_message="Initialize data.json",
+                commit_message="Initialize data directory",
                 repo_type="dataset",
                 token=token,
             )
@@ -42,26 +41,16 @@ def create_pr_in_hf_dataset(new_entry, oauth_token: gr.OAuthToken):
     except Exception as e:
         return f"Error creating or accessing repository: {e}"
 
-    # Download existing data from the dataset
-    try:
-        # Download the existing data.json file
-        local_filepath = hf_hub_download(repo_id=REPO_ID, filename=FILENAME, repo_type='dataset', token=token)
-        with open(local_filepath, 'r') as f:
-            data = json.load(f)
-    except Exception as e:
-        print(f"Error downloading existing data: {e}")
-        data = []
-
-    # Add the new entry
-    data.append(new_entry)
-
-    # Save to temporary file
-    temp_filename = 'temp_data.json'
+    # Save the new entry to a temporary file
+    temp_filename = f"{new_entry['arxiv_id']}.json"
     with open(temp_filename, 'w') as f:
-        json.dump(data, f, indent=2)
+        json.dump(new_entry, f, indent=2)
+
+    # Define the path in the repository
+    path_in_repo = f"{DATA_DIR}/{temp_filename}"
 
     # Create commit operation
-    commit = CommitOperationAdd(path_in_repo=FILENAME, path_or_fileobj=temp_filename)
+    commit = CommitOperationAdd(path_in_repo=path_in_repo, path_or_fileobj=temp_filename)
 
     # Create PR
     try: