feat: implement hierarchical S3 keys and automated version management

This commit introduces two major improvements:

1. **Hierarchical S3 Key Structure**:
   - Replace URL-based hashing with remote-name/hash(directory_path)/filename format
   - Enables remote-specific cache operations and intuitive S3 organization
   - Cache keys now independent of mirror URL changes
   - Example: fedora/886d215f6d1a0108/eccodes-2.44.0-1.fc42.x86_64.rpm

2. **Automated Version Management**:
   - Add bumpver for semantic version bumping
   - Single source of truth in pyproject.toml
   - FastAPI dynamically reads version from package metadata
   - Eliminates manual version synchronization between files

Changes:
- storage.py: New get_object_key(remote_name, path) method with directory hashing
- main.py: Dynamic version import and updated cache key generation calls
- cache.py: Updated to use new hierarchical key structure
- pyproject.toml: Added bumpver config and dev dependency

Breaking change: S3 key format changed, existing cache will need regeneration

pyproject.toml

@@ -40,4 +40,18 @@ dev = [
     "isort>=5.12.0",
     "mypy>=1.6.0",
     "ruff>=0.1.0",
+    "bumpver>=2025.1131",
 ]
+
+[tool.bumpver]
+current_version = "2.0.2"
+version_pattern = "MAJOR.MINOR.PATCH"
+commit_message = "Bump version {old_version} → {new_version}"
+commit = true
+tag = true
+push = false
+
+[[tool.bumpver.files]]
+filename = "pyproject.toml"
+search = 'version = "{current_version}"'
+replace = 'version = "{new_version}"'

src/artifactapi/cache.py

@@ -76,10 +76,8 @@ class RedisCache:
                 if remote_config:
                     base_url = remote_config.get("base_url")
                     if base_url:
-                        # Construct URL the same way as construct_remote_url
-                        remote_url = f"{base_url.rstrip('/')}/{path}"
-                        # Use URL-based key (same as cache_single_artifact)
-                        s3_key = storage.get_object_key(remote_url)
+                        # Use hierarchical path-based key (same as cache_single_artifact)
+                        s3_key = storage.get_object_key(remote_name, path)
                         if storage.exists(s3_key):
                             storage.client.delete_object(Bucket=storage.bucket, Key=s3_key)
         except Exception:

src/artifactapi/main.py

@@ -9,6 +9,13 @@ from fastapi.responses import PlainTextResponse, JSONResponse
 from pydantic import BaseModel
 from prometheus_client import generate_latest, CONTENT_TYPE_LATEST
 
+try:
+    from importlib.metadata import version
+    __version__ = version("artifactapi")
+except ImportError:
+    # Fallback for development when package isn't installed
+    __version__ = "dev"
+
 from .config import ConfigManager
 from .database import DatabaseManager
 from .storage import S3Storage
@@ -28,7 +35,7 @@ logging.basicConfig(
 )
 logger = logging.getLogger(__name__)
 
-app = FastAPI(title="Artifact Storage API", version="2.0.2")
+app = FastAPI(title="Artifact Storage API", version=__version__)
 
 # Initialize components using config
 config_path = os.environ.get("CONFIG_PATH")
@@ -181,11 +188,8 @@ async def check_artifact_patterns(
 
 
 async def cache_single_artifact(url: str, remote_name: str, path: str) -> dict:
-    # Check if using URL-based key or path-based key
-    if url.startswith("http"):
-        key = storage.get_object_key(url)
-    else:
-        key = storage.get_object_key_from_path(remote_name, path)
+    # Use hierarchical path-based key
+    key = storage.get_object_key(remote_name, path)
 
     if storage.exists(key):
         logger.info(f"Cache ALREADY EXISTS: {url} (key: {key})")
@@ -267,15 +271,10 @@ async def get_artifact(remote_name: str, path: str):
     # Construct the remote URL
     remote_url = await construct_remote_url(remote_name, path)
 
-    # Check if artifact is already cached (try both URL and path-based keys)
-    url_key = storage.get_object_key(remote_url)
-    path_key = storage.get_object_key_from_path(remote_name, path)
-
-    cached_key = None
-    if storage.exists(url_key):
-        cached_key = url_key
-    elif storage.exists(path_key):
-        cached_key = path_key
+    # Check if artifact is already cached
+    cached_key = storage.get_object_key(remote_name, path)
+    if not storage.exists(cached_key):
+        cached_key = None
 
     # For index files, check Redis TTL validity
     filename = os.path.basename(path)
@@ -355,7 +354,7 @@ async def get_artifact(remote_name: str, path: str):
 
     # Now return the cached artifact
     try:
-        cache_key = storage.get_object_key(remote_url)
+        cache_key = storage.get_object_key(remote_name, path)
         artifact_data = storage.download_object(cache_key)
         filename = os.path.basename(path)
 
@@ -377,7 +376,7 @@ async def get_artifact(remote_name: str, path: str):
         metrics.record_cache_miss(remote_name, len(artifact_data))
 
         # Record artifact mapping in database
-        cache_key = storage.get_object_key(remote_url)
+        cache_key = storage.get_object_key(remote_name, path)
         database.record_artifact_mapping(
             cache_key, remote_name, path, len(artifact_data)
         )
@@ -625,7 +624,11 @@ async def list_cached_artifacts(
 
         cached_artifacts = []
         for url in matching_urls:
-            key = storage.get_object_key(url)
+            # Extract path from URL for hierarchical key generation
+            from urllib.parse import urlparse
+            parsed = urlparse(url)
+            path = parsed.path
+            key = storage.get_object_key(remote, path)
             if storage.exists(key):
                 cached_artifacts.append(
                     {"url": url, "cached_url": storage.get_url(key), "key": key}

src/artifactapi/storage.py

@@ -1,6 +1,5 @@
 import os
 import hashlib
-from urllib.parse import urlparse
 import boto3
 from botocore.config import Config
 from botocore.exceptions import ClientError
@@ -55,17 +54,19 @@ class S3Storage:
         except ClientError:
             self.client.create_bucket(Bucket=self.bucket)
 
-    def get_object_key(self, url: str) -> str:
-        url_hash = hashlib.sha256(url.encode()).hexdigest()[:16]
-        parsed = urlparse(url)
-        filename = os.path.basename(parsed.path)
-        return f"{parsed.netloc}/{url_hash}/{filename}"
+    def get_object_key(self, remote_name: str, path: str) -> str:
+        # Extract directory path and filename
+        clean_path = path.lstrip('/')
+        filename = os.path.basename(clean_path)
+        directory_path = os.path.dirname(clean_path)
 
-    def get_object_key_from_path(self, remote_name: str, path: str) -> str:
-        # Create a key based on the API path for direct access
-        path_hash = hashlib.sha256(path.encode()).hexdigest()[:16]
-        filename = os.path.basename(path)
-        return f"{remote_name}/{path_hash}/{filename}"
+        # Hash the directory path to keep keys manageable while preserving remote structure
+        if directory_path:
+            path_hash = hashlib.sha256(directory_path.encode()).hexdigest()[:16]
+            return f"{remote_name}/{path_hash}/{filename}"
+        else:
+            # If no directory, just use remote and filename
+            return f"{remote_name}/{filename}"
 
     def exists(self, key: str) -> bool:
         try: