Compare commits

..

18 Commits

Author SHA1 Message Date
unkinben cf854a2ace chore: derive version from git tags via hatch-vcs
Replace hardcoded version in pyproject.toml with hatch-vcs so the
package version is read from git tags at build time. Dockerfile
accepts a VERSION build arg and passes it as HATCH_VCS_PRETEND_VERSION
for builds without a git checkout. Makefile _tag target now rebuilds
the container with the correct version automatically.
2026-04-25 16:53:37 +10:00
unkinben 4651183ed1 feat: add Docker registry proxy support with proper cache classification
- Add /v2/ endpoint implementing OCI Distribution API for native docker pull support
- Add docker_auth.py with Bearer token challenge handling and in-memory token cache
- Classify tag-based manifests (/manifests/<tag>) as index (short TTL, mutable)
- Classify digest-pinned manifests (/manifests/sha256:...) and blobs as file cache (indefinite, immutable)
- Deduplicate blob storage by keying on sha256 digest rather than image path
- Support username/password auth per docker remote in remotes.yaml
2026-04-25 16:35:27 +10:00
unkinben 5733d52e51 Merge pull request 'feature/cache-flush-api-enhancement' (#7) from feature/cache-flush-api-enhancement into master
Reviewed-on: #7
2026-01-25 11:34:43 +11:00
unkinben bf8a176dda Bump version: 2.0.4 → 2.0.5 2026-01-25 11:09:29 +11:00
unkinben 3e8e819ecf feat: enhance cache flush API for remote-specific S3 clearing
Add support for remote-specific S3 object deletion using hierarchical key prefixes.
When a remote parameter is provided, only objects under that remote's hierarchy
(e.g., 'fedora/*') are deleted, preserving cached files for other remotes.

Key improvements:
- Use S3 list_objects_v2 Prefix parameter for targeted deletion
- Enhanced logging to indicate scope of cache clearing operations
- Backward compatible - existing behavior preserved when no remote specified

Resolves artifactapi-u46
2026-01-25 11:08:46 +11:00
unkinben de04e4d2b2 Merge pull request 'benvin/path-based-storage' (#6) from benvin/path-based-storage into master
Reviewed-on: #6
2026-01-25 00:00:53 +11:00
unkinben 16c8bd60eb Bump version: 2.0.3 → 2.0.4 2026-01-24 23:59:48 +11:00
unkinben d39550c4e8 chore: migrate from bumpver to bump-my-version for better reliability 2026-01-24 23:59:05 +11:00
unkinben 424de5cc13 fix: sync package version with bumpver version (2.0.3) 2026-01-24 23:54:42 +11:00
unkinben 3f54874421 Bump version 2.0.2 → 2.0.3 2026-01-24 23:53:19 +11:00
unkinben 6e8912eed1 chore: ignore local config overrides (docker-compose.yml, ca-bundle.pem) 2026-01-24 23:53:08 +11:00
unkinben 5a0e8b4e0b 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
2026-01-24 23:51:03 +11:00
unkinben 1a71a2c9fa Merge pull request 'feat: add cache flush API and fix cache key consistency' (#5) from benvin/cache_flush into master
Reviewed-on: #5
2026-01-13 19:02:52 +11:00
unkinben d2ecc6b1a0 feat: add cache flush API and fix cache key consistency
- Add PUT /cache/flush endpoint for selective cache management
- Fix critical cache key mismatch in cleanup_expired_index()
- Update index file detection to use full path instead of filename
- Bump version to 2.0.2

The key changes include adding a comprehensive cache flush API that supports selective flushing by cache type (index, files, metrics) and fixing a critical bug where cache keys were
inconsistent between storage and cleanup operations, preventing proper cache invalidation.
2026-01-13 19:02:31 +11:00
unkinben e4013e6a2a Merge pull request 'feat: index caching' (#4) from benvin/index_caching into master
Reviewed-on: #4
2026-01-13 18:14:39 +11:00
unkinben 9defc78e21 feat: index caching
- improve index detection for rpms
- improve logging
2026-01-13 18:13:47 +11:00
unkinben f40675f3d2 Merge pull request 'feat: add fedora index files' (#3) from benvin/fedora_indexes into master
Reviewed-on: #3
2026-01-10 17:02:58 +11:00
unkinben b54e6c3e0c feat: add fedora index files
- ensure files matching xml.zck and xml.zst are marked as index files
2026-01-10 17:01:39 +11:00
9 changed files with 413 additions and 134 deletions
+4
View File
@@ -45,3 +45,7 @@ uv.lock
# Docker volumes
minio_data/
# Local configuration overrides
docker-compose.yml
ca-bundle.pem
+3 -5
View File
@@ -23,10 +23,6 @@ RUN wget -O /app/uv-x86_64-unknown-linux-musl.tar.gz https://github.com/astral-s
chmod +x /usr/local/bin/uv && \
uv --version
# Copy CA bundle from host
COPY ca-bundle.pem /app/ca-bundle.pem
RUN chmod 644 /app/ca-bundle.pem
# Create non-root user first
RUN adduser -D -s /bin/sh appuser && \
chown -R appuser:appuser /app
@@ -36,11 +32,13 @@ COPY --chown=appuser:appuser pyproject.toml uv.lock README.md ./
# Switch to appuser and install Python dependencies
USER appuser
RUN uv sync --frozen
ARG VERSION=dev
RUN HATCH_VCS_PRETEND_VERSION=${VERSION} uv sync --frozen
# Copy application source
COPY --chown=appuser:appuser src/ ./src/
COPY --chown=appuser:appuser remotes.yaml ./
COPY --chown=appuser:appuser ca-bundle.pem ./
# Expose port
EXPOSE 8000
+24 -2
View File
@@ -26,8 +26,6 @@ format:
uv run ruff format .
run:
uv venv --python 3.11 && \
source .venv/bin/activate && \
uv run python -m src.artifactapi.main
docker-up:
@@ -47,3 +45,27 @@ docker-clean:
docker system prune -f
docker-restart: docker-down docker-up
# Bump helpers — reads the latest semver tag and creates the next one.
# If no tag exists yet, starts from v0.0.0.
_LATEST := $(shell git tag --sort=-v:refname | grep -E '^v[0-9]+\.[0-9]+\.[0-9]+$$' | head -1)
_BASE := $(if $(_LATEST),$(_LATEST),v0.0.0)
_MAJ := $(shell echo $(_BASE) | sed 's/^v//' | cut -d. -f1)
_MIN := $(shell echo $(_BASE) | sed 's/^v//' | cut -d. -f2)
_PAT := $(shell echo $(_BASE) | sed 's/^v//' | cut -d. -f3)
patch:
@NEW=v$(_MAJ).$(_MIN).$(shell expr $(_PAT) + 1); \
git tag $$NEW && echo "Tagged $$NEW" && $(MAKE) _tag TAG=$$NEW
minor:
@NEW=v$(_MAJ).$(shell expr $(_MIN) + 1).0; \
git tag $$NEW && echo "Tagged $$NEW" && $(MAKE) _tag TAG=$$NEW
major:
@NEW=v$(shell expr $(_MAJ) + 1).0.0; \
git tag $$NEW && echo "Tagged $$NEW" && $(MAKE) _tag TAG=$$NEW
_tag:
git push origin $(TAG)
docker-compose build --no-cache --build-arg VERSION=$(TAG:v%=%)
-86
View File
@@ -1,86 +0,0 @@
version: '3.8'
services:
artifactapi:
build:
context: .
dockerfile: Dockerfile
no_cache: true
ports:
- "8000:8000"
environment:
- CONFIG_PATH=/app/remotes.yaml
- DBHOST=postgres
- DBPORT=5432
- DBUSER=artifacts
- DBPASS=artifacts123
- DBNAME=artifacts
- REDIS_URL=redis://redis:6379
- MINIO_ENDPOINT=minio:9000
- MINIO_ACCESS_KEY=minioadmin
- MINIO_SECRET_KEY=minioadmin
- MINIO_BUCKET=artifacts
- MINIO_SECURE=false
depends_on:
postgres:
condition: service_healthy
redis:
condition: service_healthy
minio:
condition: service_healthy
healthcheck:
test: ["CMD", "curl", "-f", "http://localhost:8000/health"]
interval: 30s
timeout: 10s
retries: 3
minio:
image: minio/minio:latest
ports:
- "9000:9000"
- "9001:9001"
environment:
MINIO_ROOT_USER: minioadmin
MINIO_ROOT_PASSWORD: minioadmin
command: server /data --console-address ":9001"
volumes:
- minio_data:/data
healthcheck:
test: ["CMD", "curl", "-f", "http://localhost:9000/minio/health/live"]
interval: 30s
timeout: 20s
retries: 3
redis:
image: redis:7-alpine
ports:
- "6379:6379"
volumes:
- redis_data:/data
command: redis-server --save 20 1
healthcheck:
test: ["CMD", "redis-cli", "ping"]
interval: 30s
timeout: 10s
retries: 3
postgres:
image: postgres:15-alpine
ports:
- "5432:5432"
environment:
POSTGRES_DB: artifacts
POSTGRES_USER: artifacts
POSTGRES_PASSWORD: artifacts123
volumes:
- postgres_data:/var/lib/postgresql/data
healthcheck:
test: ["CMD-SHELL", "pg_isready -U artifacts -d artifacts"]
interval: 30s
timeout: 10s
retries: 3
volumes:
minio_data:
redis_data:
postgres_data:
+5 -2
View File
@@ -1,6 +1,6 @@
[project]
name = "artifactapi"
version = "2.0.0"
dynamic = ["version"]
description = "Generic artifact caching system with support for various package managers"
dependencies = [
@@ -23,9 +23,12 @@ license = {text = "MIT"}
artifactapi = "artifactapi.main:main"
[build-system]
requires = ["hatchling"]
requires = ["hatchling", "hatch-vcs"]
build-backend = "hatchling.build"
[tool.hatch.version]
source = "vcs"
[tool.hatch.metadata]
allow-direct-references = true
+28 -6
View File
@@ -23,8 +23,20 @@ class RedisCache:
file_path.endswith("APKINDEX.tar.gz")
or file_path.endswith("Packages.gz")
or file_path.endswith("repomd.xml")
or "repodata/" in file_path
and file_path.endswith((".xml", ".xml.gz", ".xml.bz2", ".xml.xz"))
or ("repodata/" in file_path
and file_path.endswith((
".xml", ".xml.gz", ".xml.bz2", ".xml.xz", ".xml.zck", ".xml.zst",
".sqlite", ".sqlite.gz", ".sqlite.bz2", ".sqlite.xz", ".sqlite.zck", ".sqlite.zst",
".yaml.xz", ".yaml.gz", ".yaml.bz2", ".yaml.zst",
".asc", ".txt"
)))
# Docker tag-based manifests are mutable (index); digest-pinned are immutable (file)
or (
"/manifests/" in file_path
and not file_path.split("/manifests/", 1)[1].startswith("sha256:")
)
or "/tags/list" in file_path
or file_path.endswith("/tags/list")
)
def get_index_cache_key(self, remote_name: str, path: str) -> str:
@@ -61,9 +73,19 @@ 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:
# 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:
pass
+96
View File
@@ -0,0 +1,96 @@
import time
import logging
import re
from typing import Optional
import httpx
logger = logging.getLogger(__name__)
# In-memory token cache: key -> (token, expires_at)
_token_cache: dict[str, tuple[str, float]] = {}
_WWW_AUTH_RE = re.compile(
r'Bearer\s+realm="(?P<realm>[^"]+)"'
r'(?:,service="(?P<service>[^"]*)")?'
r'(?:,scope="(?P<scope>[^"]*)")?',
re.IGNORECASE,
)
def _cache_key(realm: str, service: str, scope: str, username: Optional[str]) -> str:
return f"{realm}|{service}|{scope}|{username or ''}"
def _get_cached_token(key: str) -> Optional[str]:
entry = _token_cache.get(key)
if entry and entry[1] > time.time():
return entry[0]
_token_cache.pop(key, None)
return None
def _store_token(key: str, token: str, expires_in: int) -> None:
# Expire 30s early to avoid using a token right as it expires
_token_cache[key] = (token, time.time() + max(expires_in - 30, 10))
async def fetch_token(
realm: str,
service: str,
scope: str,
username: Optional[str] = None,
password: Optional[str] = None,
) -> Optional[str]:
"""Fetch a Bearer token from a Docker registry auth server."""
key = _cache_key(realm, service, scope, username)
cached = _get_cached_token(key)
if cached:
return cached
params: dict[str, str] = {}
if service:
params["service"] = service
if scope:
params["scope"] = scope
auth = (username, password) if username and password else None
try:
async with httpx.AsyncClient(follow_redirects=True) as client:
response = await client.get(realm, params=params, auth=auth)
response.raise_for_status()
data = response.json()
except Exception as e:
logger.warning(f"Docker token fetch failed ({realm}): {e}")
return None
token = data.get("token") or data.get("access_token")
if not token:
logger.warning(f"Docker token response missing token field: {data}")
return None
expires_in = int(data.get("expires_in", 300))
_store_token(key, token, expires_in)
logger.debug(f"Docker token obtained (realm={realm}, service={service}, scope={scope}, expires_in={expires_in}s)")
return token
def parse_www_authenticate(header: str) -> Optional[tuple[str, str, str]]:
"""Parse WWW-Authenticate: Bearer header. Returns (realm, service, scope) or None."""
m = _WWW_AUTH_RE.search(header)
if not m:
return None
return m.group("realm"), m.group("service") or "", m.group("scope") or ""
async def get_docker_token_for_response(
www_authenticate: str,
username: Optional[str] = None,
password: Optional[str] = None,
) -> Optional[str]:
"""Given a WWW-Authenticate header value, fetch and return a Bearer token."""
parsed = parse_www_authenticate(www_authenticate)
if not parsed:
return None
realm, service, scope = parsed
return await fetch_token(realm, service, scope, username, password)
+233 -22
View File
@@ -1,18 +1,28 @@
import os
import re
import json
import hashlib
import logging
from typing import Dict, Any, Optional
import httpx
from fastapi import FastAPI, HTTPException, Response, Query, File, UploadFile
from fastapi import FastAPI, HTTPException, Response, Request, Query, File, UploadFile
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
from .cache import RedisCache
from .metrics import MetricsManager
from .docker_auth import get_docker_token_for_response
class ArtifactRequest(BaseModel):
@@ -20,7 +30,14 @@ class ArtifactRequest(BaseModel):
include_pattern: str
app = FastAPI(title="Artifact Storage API", version="2.0.0")
# Configure logging
logging.basicConfig(
level=logging.INFO,
format='%(asctime)s - %(name)s - %(levelname)s - %(message)s'
)
logger = logging.getLogger(__name__)
app = FastAPI(title="Artifact Storage API", version=__version__)
# Initialize components using config
config_path = os.environ.get("CONFIG_PATH")
@@ -45,7 +62,7 @@ def read_root():
config._check_reload()
return {
"message": "Artifact Storage API",
"version": "2.0.0",
"version": app.version,
"remotes": list(config.config.get("remotes", {}).keys()),
}
@@ -55,6 +72,87 @@ 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:
# Use prefix filtering for remote-specific deletion
list_params = {"Bucket": storage.bucket}
if remote:
list_params["Prefix"] = f"{remote}/"
response = storage.client.list_objects_v2(**list_params)
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:
scope = f" for remote '{remote}'" if remote else ""
result["flushed"]["operations"].append(f"Deleted {len(objects_to_delete)} S3 objects{scope}")
logger.info(f"Cache flush: Deleted {len(objects_to_delete)} S3 objects{scope}")
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:
@@ -68,6 +166,12 @@ async def construct_remote_url(remote_name: str, path: str) -> str:
status_code=500, detail=f"No base_url configured for remote '{remote_name}'"
)
# Handle Docker registry URLs
if remote_config.get("type") == "docker":
# Convert Docker paths to v2 API format
# e.g., library/nginx/manifests/latest -> v2/library/nginx/manifests/latest
return f"{base_url}/v2/{path}"
return f"{base_url}/{path}"
@@ -98,13 +202,11 @@ 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})")
return {
"url": url,
"cached_url": storage.get_url(key),
@@ -112,12 +214,41 @@ async def cache_single_artifact(url: str, remote_name: str, path: str) -> dict:
}
try:
remote_config = config.get_remote_config(remote_name) or {}
is_docker = remote_config.get("type") == "docker" or "/v2/" in url
# Prepare headers for Docker registry requests
headers = {}
if is_docker:
if "/manifests/" in url:
headers["Accept"] = (
"application/vnd.docker.distribution.manifest.v2+json,"
"application/vnd.oci.image.manifest.v1+json,"
"application/vnd.oci.image.index.v1+json,"
"application/vnd.docker.distribution.manifest.list.v2+json"
)
elif "/blobs/" in url:
headers["Accept"] = "application/octet-stream"
async with httpx.AsyncClient(follow_redirects=True) as client:
response = await client.get(url)
response = await client.get(url, headers=headers)
# Handle Docker Bearer token challenge
if response.status_code == 401 and is_docker:
www_auth = response.headers.get("WWW-Authenticate", "")
username = remote_config.get("username")
password = remote_config.get("password")
token = await get_docker_token_for_response(www_auth, username, password)
if token:
headers["Authorization"] = f"Bearer {token}"
response = await client.get(url, headers=headers)
response.raise_for_status()
storage_path = storage.upload(key, response.content)
logger.info(f"Cache ADD SUCCESS: {url} (size: {len(response.content)} bytes, key: {key})")
return {
"url": url,
"cached_url": storage.get_url(key),
@@ -173,6 +304,7 @@ async def get_artifact(remote_name: str, path: str):
# Check if artifact matches configured patterns
if not await check_artifact_patterns(remote_name, repo_path, file_path, path):
logger.info(f"PATTERN BLOCKED: {remote_name}/{path} - not matching include patterns")
raise HTTPException(
status_code=403, detail="Artifact not allowed by configuration patterns"
)
@@ -180,24 +312,20 @@ 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)
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
if not cache.is_index_valid(remote_name, path):
# Index has expired, remove it from S3
logger.info(f"Index EXPIRED: {remote_name}/{path} - removing from cache")
cache.cleanup_expired_index(storage, remote_name, path)
cached_key = None # Force re-download
@@ -207,6 +335,9 @@ async def get_artifact(remote_name: str, path: str):
artifact_data = storage.download_object(cached_key)
filename = os.path.basename(path)
# Log cache hit
logger.info(f"Cache HIT: {remote_name}/{path} (size: {len(artifact_data)} bytes, key: {cached_key})")
# Determine content type based on file extension
content_type = "application/octet-stream"
if filename.endswith(".tar.gz"):
@@ -245,9 +376,11 @@ async def get_artifact(remote_name: str, path: str):
)
# Artifact not cached, cache it first
logger.info(f"Cache MISS: {remote_name}/{path} - fetching from remote: {remote_url}")
result = await cache_single_artifact(remote_url, remote_name, path)
if result["status"] == "error":
logger.error(f"Cache ADD FAILED: {remote_name}/{path} - {result['error']}")
raise HTTPException(
status_code=502, detail=f"Failed to fetch artifact: {result['error']}"
)
@@ -258,10 +391,11 @@ async def get_artifact(remote_name: str, path: str):
cache_config = config.get_cache_config(remote_name)
index_ttl = cache_config.get("index_ttl", 300) # Default 5 minutes
cache.mark_index_cached(remote_name, path, index_ttl)
logger.info(f"Index file cached with TTL: {remote_name}/{path} (ttl: {index_ttl}s)")
# 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)
@@ -283,7 +417,7 @@ 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)
)
@@ -301,6 +435,79 @@ async def get_artifact(remote_name: str, path: str):
raise HTTPException(status_code=500, detail=f"Error serving artifact: {str(e)}")
@app.get("/v2/")
async def docker_v2_ping():
return Response(
content="{}",
media_type="application/json",
headers={"Docker-Distribution-Api-Version": "registry/2.0"},
)
@app.api_route("/v2/{remote_name}/{path:path}", methods=["GET", "HEAD"])
async def docker_v2_proxy(request: Request, remote_name: str, path: str):
remote_config = config.get_remote_config(remote_name)
if not remote_config:
raise HTTPException(status_code=404, detail=f"Remote '{remote_name}' not configured")
if remote_config.get("type") != "docker":
raise HTTPException(status_code=400, detail=f"Remote '{remote_name}' is not a docker remote")
remote_url = await construct_remote_url(remote_name, path)
cached_key = storage.get_object_key(remote_name, path)
if not storage.exists(cached_key):
cached_key = None
is_index = cache.is_index_file(path)
if cached_key and is_index:
if not cache.is_index_valid(remote_name, path):
logger.info(f"Index EXPIRED: {remote_name}/{path} - removing from cache")
cache.cleanup_expired_index(storage, remote_name, path)
cached_key = None
if not cached_key:
logger.info(f"Cache MISS: {remote_name}/{path} - fetching from remote: {remote_url}")
result = await cache_single_artifact(remote_url, remote_name, path)
if result["status"] == "error":
raise HTTPException(status_code=502, detail=f"Failed to fetch: {result['error']}")
if result["status"] == "cached" and is_index:
cache_config = config.get_cache_config(remote_name)
index_ttl = cache_config.get("index_ttl", 300)
cache.mark_index_cached(remote_name, path, index_ttl)
logger.info(f"Index file cached with TTL: {remote_name}/{path} (ttl: {index_ttl}s)")
artifact_data = storage.download_object(storage.get_object_key(remote_name, path))
is_blob = "/blobs/" in path
if is_blob:
content_type = "application/octet-stream"
else:
try:
manifest_json = json.loads(artifact_data)
content_type = manifest_json.get("mediaType")
if not content_type:
if "manifests" in manifest_json:
content_type = "application/vnd.oci.image.index.v1+json"
else:
content_type = "application/vnd.oci.image.manifest.v1+json"
except Exception:
content_type = "application/vnd.oci.image.manifest.v1+json"
digest = f"sha256:{hashlib.sha256(artifact_data).hexdigest()}"
headers = {
"Docker-Distribution-Api-Version": "registry/2.0",
"Docker-Content-Digest": digest,
"Content-Length": str(len(artifact_data)),
}
if request.method == "HEAD":
return Response(status_code=200, headers=headers, media_type=content_type)
metrics.record_cache_hit(remote_name, len(artifact_data))
return Response(content=artifact_data, media_type=content_type, headers=headers)
async def discover_artifacts(remote: str, include_pattern: str) -> list[str]:
if "github.com" in remote:
return await discover_github_releases(remote, include_pattern)
@@ -531,7 +738,11 @@ 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}
+20 -11
View File
@@ -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 +54,27 @@ 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}"
# Special handling for Docker registry blobs (use digest as key for deduplication)
if "/blobs/sha256:" in clean_path:
# Extract the SHA256 digest for Docker blobs
parts = clean_path.split("/blobs/sha256:")
if len(parts) == 2:
digest = parts[1]
return f"{remote_name}/blobs/sha256/{digest}"
# 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: