feat: add cache flush API and fix cache key consistency #5

Merged
unkinben merged 1 commits from benvin/cache_flush into master 2026-01-13 19:02:53 +11:00
3 changed files with 94 additions and 7 deletions

View File

@ -1,6 +1,6 @@
[project]
name = "artifactapi"
version = "2.0.1"
version = "2.0.2"
description = "Generic artifact caching system with support for various package managers"
dependencies = [

View File

@ -66,9 +66,21 @@ class RedisCache:
return
try:
# Get the S3 key and remove it
s3_key = storage.get_object_key_from_path(remote_name, path)
if storage.exists(s3_key):
storage.client.delete_object(Bucket=storage.bucket, Key=s3_key)
# Construct the URL the same way as in the main flow
from .config import ConfigManager
import os
config_path = os.environ.get("CONFIG_PATH")
if config_path:
config = ConfigManager(config_path)
remote_config = config.get_remote_config(remote_name)
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)
if storage.exists(s3_key):
storage.client.delete_object(Bucket=storage.bucket, Key=s3_key)
except Exception:
pass

View File

@ -28,7 +28,7 @@ logging.basicConfig(
)
logger = logging.getLogger(__name__)
app = FastAPI(title="Artifact Storage API", version="2.0.1")
app = FastAPI(title="Artifact Storage API", version="2.0.2")
# Initialize components using config
config_path = os.environ.get("CONFIG_PATH")
@ -63,6 +63,81 @@ def health_check():
return {"status": "healthy"}
@app.put("/cache/flush")
def flush_cache(
remote: str = Query(default=None, description="Specific remote to flush (optional)"),
cache_type: str = Query(default="all", description="Type to flush: 'all', 'index', 'files', 'metrics'")
):
"""Flush cache entries for specified remote or all remotes"""
try:
result = {
"remote": remote,
"cache_type": cache_type,
"flushed": {
"redis_keys": 0,
"s3_objects": 0,
"operations": []
}
}
# Flush Redis entries based on cache_type
if cache_type in ["all", "index", "metrics"] and cache.available and cache.client:
patterns = []
if cache_type in ["all", "index"]:
if remote:
patterns.append(f"index:{remote}:*")
else:
patterns.append("index:*")
if cache_type in ["all", "metrics"]:
if remote:
patterns.append(f"metrics:*:{remote}")
else:
patterns.append("metrics:*")
for pattern in patterns:
keys = cache.client.keys(pattern)
if keys:
cache.client.delete(*keys)
result["flushed"]["redis_keys"] += len(keys)
logger.info(f"Cache flush: Deleted {len(keys)} Redis keys matching '{pattern}'")
if result["flushed"]["redis_keys"] > 0:
result["flushed"]["operations"].append(f"Deleted {result['flushed']['redis_keys']} Redis keys")
# Flush S3 objects if requested
if cache_type in ["all", "files"]:
try:
response = storage.client.list_objects_v2(Bucket=storage.bucket)
if 'Contents' in response:
objects_to_delete = [obj['Key'] for obj in response['Contents']]
for key in objects_to_delete:
try:
storage.client.delete_object(Bucket=storage.bucket, Key=key)
result["flushed"]["s3_objects"] += 1
except Exception as e:
logger.warning(f"Failed to delete S3 object {key}: {e}")
if objects_to_delete:
result["flushed"]["operations"].append(f"Deleted {len(objects_to_delete)} S3 objects")
logger.info(f"Cache flush: Deleted {len(objects_to_delete)} S3 objects")
except Exception as e:
result["flushed"]["operations"].append(f"S3 flush failed: {str(e)}")
logger.error(f"Cache flush S3 error: {e}")
if not result["flushed"]["operations"]:
result["flushed"]["operations"].append("No cache entries found to flush")
return result
except Exception as e:
logger.error(f"Cache flush error: {e}")
raise HTTPException(status_code=500, detail=f"Cache flush failed: {str(e)}")
async def construct_remote_url(remote_name: str, path: str) -> str:
remote_config = config.get_remote_config(remote_name)
if not remote_config:
@ -204,7 +279,7 @@ async def get_artifact(remote_name: str, path: str):
# For index files, check Redis TTL validity
filename = os.path.basename(path)
is_index = cache.is_index_file(filename)
is_index = cache.is_index_file(path) # Check full path, not just filename
if cached_key and is_index:
# Index file exists, but check if it's still valid