refine

app.py

@@ -24,12 +24,10 @@ load_dotenv()
 # CONFIGURATION
 # =============================================================================
 
-AGENTS_REPO = "SWE-Arena/bot_metadata"  # HuggingFace dataset for assistant metadata
-AGENTS_REPO_LOCAL_PATH = os.path.expanduser("~/bot_metadata")  # Local git clone path
+AGENTS_REPO = "SWE-Arena/bot_data"  # HuggingFace dataset for assistant metadata
 LEADERBOARD_FILENAME = f"{os.getenv('COMPOSE_PROJECT_NAME')}.json"
 LEADERBOARD_REPO = "SWE-Arena/leaderboard_data"  # HuggingFace dataset for leaderboard data
 LONGSTANDING_GAP_DAYS = 30  # Minimum days for an issue to be considered long-standing
-GIT_SYNC_TIMEOUT = 300  # 5 minutes timeout for git pull
 MAX_RETRIES = 5
 
 LEADERBOARD_COLUMNS = [
@@ -104,113 +102,52 @@ def validate_github_username(identifier):
 # HUGGINGFACE DATASET OPERATIONS
 # =============================================================================
 
-def sync_agents_repo():
-    """
-    Sync local bot_metadata repository with remote using git pull.
-    This is MANDATORY to ensure we have the latest bot data.
-    Raises exception if sync fails.
-    """
-    if not os.path.exists(AGENTS_REPO_LOCAL_PATH):
-        error_msg = f"Local repository not found at {AGENTS_REPO_LOCAL_PATH}"
-        print(f"   Error {error_msg}")
-        print(f"   Please clone it first: git clone https://huggingface.co/datasets/{AGENTS_REPO}")
-        raise FileNotFoundError(error_msg)
-
-    if not os.path.exists(os.path.join(AGENTS_REPO_LOCAL_PATH, '.git')):
-        error_msg = f"{AGENTS_REPO_LOCAL_PATH} exists but is not a git repository"
-        print(f"   Error {error_msg}")
-        raise ValueError(error_msg)
-
-    try:
-        # Run git pull with extended timeout due to large repository
-        result = subprocess.run(
-            ['git', 'pull'],
-            cwd=AGENTS_REPO_LOCAL_PATH,
-            capture_output=True,
-            text=True,
-            timeout=GIT_SYNC_TIMEOUT
-        )
-
-        if result.returncode == 0:
-            output = result.stdout.strip()
-            if "Already up to date" in output or "Already up-to-date" in output:
-                print(f"   Success Repository is up to date")
-            else:
-                print(f"   Success Repository synced successfully")
-                if output:
-                    # Print first few lines of output
-                    lines = output.split('\n')[:5]
-                    for line in lines:
-                        print(f"     {line}")
-            return True
-        else:
-            error_msg = f"Git pull failed: {result.stderr.strip()}"
-            print(f"   Error {error_msg}")
-            raise RuntimeError(error_msg)
-
-    except subprocess.TimeoutExpired:
-        error_msg = f"Git pull timed out after {GIT_SYNC_TIMEOUT} seconds"
-        print(f"   Error {error_msg}")
-        raise TimeoutError(error_msg)
-    except (FileNotFoundError, ValueError, RuntimeError, TimeoutError):
-        raise  # Re-raise expected exceptions
-    except Exception as e:
-        error_msg = f"Error syncing repository: {str(e)}"
-        print(f"   Error {error_msg}")
-        raise RuntimeError(error_msg) from e
-
-
 def load_agents_from_hf():
-    """
-    Load all assistant metadata JSON files from local git repository.
-    ALWAYS syncs with remote first to ensure we have the latest bot data.
-    """
-    # MANDATORY: Sync with remote first to get latest bot data
-    print(f"   Syncing bot_metadata repository to get latest assistants...")
-    sync_agents_repo()  # Will raise exception if sync fails
-
-    assistants = []
-
-    # Scan local directory for JSON files
-    if not os.path.exists(AGENTS_REPO_LOCAL_PATH):
-        raise FileNotFoundError(f"Local repository not found at {AGENTS_REPO_LOCAL_PATH}")
-
-    # Walk through the directory to find all JSON files
-    files_processed = 0
-    print(f"   Loading assistant metadata from {AGENTS_REPO_LOCAL_PATH}...")
-
-    for root, dirs, files in os.walk(AGENTS_REPO_LOCAL_PATH):
-        # Skip .git directory
-        if '.git' in root:
-            continue
+    """Load all assistant metadata JSON files from HuggingFace dataset."""
+    try:
+        api = HfApi()
+        assistants = []
 
-        for filename in files:
-            if not filename.endswith('.json'):
-                continue
+        # List all files in the repository
+        files = list_repo_files_with_backoff(api=api, repo_id=AGENTS_REPO, repo_type="dataset")
 
-            files_processed += 1
-            file_path = os.path.join(root, filename)
+        # Filter for JSON files only
+        json_files = [f for f in files if f.endswith('.json')]
 
+        # Download and parse each JSON file
+        for json_file in json_files:
             try:
-                with open(file_path, 'r', encoding='utf-8') as f:
+                file_path = hf_hub_download_with_backoff(
+                    repo_id=AGENTS_REPO,
+                    filename=json_file,
+                    repo_type="dataset"
+                )
+
+                with open(file_path, 'r') as f:
                     agent_data = json.load(f)
 
-                # Only include active assistants
-                if agent_data.get('status') != 'active':
-                    continue
+                    # Only process assistants with status == "active"
+                    if agent_data.get('status') != 'active':
+                        continue
+
+                    # Extract github_identifier from filename (e.g., "assistant[bot].json" -> "assistant[bot]")
+                    filename_identifier = json_file.replace('.json', '')
 
-                # Extract github_identifier from filename
-                github_identifier = filename.replace('.json', '')
-                agent_data['github_identifier'] = github_identifier
+                    # Add or override github_identifier to match filename
+                    agent_data['github_identifier'] = filename_identifier
 
-                assistants.append(agent_data)
+                    assistants.append(agent_data)
 
             except Exception as e:
-                print(f"   Warning Error loading {filename}: {str(e)}")
+                print(f"Warning: Could not load {json_file}: {str(e)}")
                 continue
 
-    print(f"   Success Loaded {len(assistants)} active assistants (from {files_processed} total files)")
-    return assistants
+        print(f"Loaded {len(assistants)} assistants from HuggingFace")
+        return assistants
+
+    except Exception as e:
+        print(f"Could not load assistants from HuggingFace: {str(e)}")
+        return None
 
 
 def get_hf_token():

msr.py

@@ -29,8 +29,8 @@ load_dotenv()
 SCRIPT_DIR = os.path.dirname(os.path.abspath(__file__))
 BASE_DIR = os.path.dirname(SCRIPT_DIR)  # Parent directory
 
-AGENTS_REPO = "SWE-Arena/bot_metadata"
-AGENTS_REPO_LOCAL_PATH = os.path.join(BASE_DIR, "bot_metadata")  # Local git clone path
+AGENTS_REPO = "SWE-Arena/bot_data"
+AGENTS_REPO_LOCAL_PATH = os.path.join(BASE_DIR, "bot_data")  # Local git clone path
 DUCKDB_CACHE_FILE = os.path.join(SCRIPT_DIR, "cache.duckdb")
 GHARCHIVE_DATA_LOCAL_PATH = os.path.join(BASE_DIR, "gharchive/data")
 LEADERBOARD_FILENAME = f"{os.getenv('COMPOSE_PROJECT_NAME')}.json"
@@ -881,7 +881,7 @@ def fetch_all_metadata_streaming(conn, identifiers, start_date, end_date):
 
 def sync_agents_repo():
     """
-    Sync local bot_metadata repository with remote using git pull.
+    Sync local bot_data repository with remote using git pull.
     This is MANDATORY to ensure we have the latest bot data.
     Raises exception if sync fails.
     """
@@ -941,7 +941,7 @@ def load_agents_from_hf():
     ALWAYS syncs with remote first to ensure we have the latest bot data.
     """
     # MANDATORY: Sync with remote first to get latest bot data
-    print(f"   Syncing bot_metadata repository to get latest assistants...")
+    print(f"   Syncing bot_data repository to get latest assistants...")
     sync_agents_repo()  # Will raise exception if sync fails
 
     assistants = []
@@ -1180,8 +1180,8 @@ def construct_leaderboard_from_metadata(all_metadata_dict, assistants, wanted_re
         identifier = assistant.get('github_identifier')
         agent_name = assistant.get('name', 'Unknown')
 
-        bot_metadata = all_metadata_dict.get(identifier, [])
-        stats = calculate_issue_stats_from_metadata(bot_metadata)
+        bot_data = all_metadata_dict.get(identifier, [])
+        stats = calculate_issue_stats_from_metadata(bot_data)
 
         # Add wanted issues count
         resolved_wanted = len(wanted_resolved_dict.get(identifier, []))