wip: rewrite repository

An LRU cache makes 0 sense here

Author: MHajoha

questionpy_server/repository/__init__.py

@@ -3,45 +3,46 @@
 #  (c) Technische Universität Berlin, innoCampus <info@isis.tu-berlin.de>
 
 import logging
+import re
 from asyncio import to_thread
 from gzip import decompress
-from urllib.parse import urljoin
+from urllib.parse import urljoin, urlparse
 
 from questionpy_server.cache import CacheItemTooLargeError, FileCache
 from questionpy_server.repository.helper import download
 from questionpy_server.repository.models import RepoMeta, RepoPackage, RepoPackageIndex
 from questionpy_server.utils.logger import URLAdapter
 
+_SCHEME_AND_AUTH_PATTERN = re.compile(r"^https?://(?:[^/]+@)?")
+_FILENAME_SPECIAL_CHARACTERS_PATTERN = re.compile(r"[/\\?%*:|\"<>,;=\s]+")
+
+def _url_to_safe_path_part(url: str) -> str:
+    return _FILENAME_SPECIAL_CHARACTERS_PATTERN.sub("-", _SCHEME_AND_AUTH_PATTERN.sub("", url))
 
 class Repository:
     def __init__(self, url: str, cache: FileCache):
         self._url_base = url
         self._url_index = urljoin(self._url_base, "PACKAGES.json.gz")
         self._url_meta = urljoin(self._url_base, "META.json")
 
+        self._cache_key = _url_to_safe_path_part(url)
+
         self._cache = cache
 
         logger = logging.getLogger("questionpy-server:repository")
         self._log = URLAdapter(logger, {"url": self._url_base})
 
     async def get_meta(self) -> RepoMeta:
-        """Downloads and verifies metadata.
-
-        Returns:
-            RepoMeta: Metadata
-        """
+        """Downloads and verifies metadata."""
         meta = await download(self._url_meta)
         # TODO: verify downloaded data
         return RepoMeta.model_validate_json(meta)
 
     async def get_packages(self, meta: RepoMeta) -> dict[str, RepoPackage]:
         """Downloads and verifies package index.
 
-        Args:
-            meta (RepoMeta): Metadata
-
         Returns:
-            dict[str, RepoPackage]: package index, where keys are package hashes
+            Package index, where keys are package hashes.
         """
         try:
             # Try to get the index from cache.
@@ -71,11 +72,8 @@ async def get_packages(self, meta: RepoMeta) -> dict[str, RepoPackage]:
     async def get_package(self, package: RepoPackage) -> bytes:
         """Download a specific package from the repository.
 
-        Args:
-            package (RepoPackage): repository package
-
         Returns:
-            bytes: raw package bytes
+            Raw package bytes.
         """
         url = urljoin(self._url_base, package.path)
         return await download(url, size=package.size, expected_hash=package.sha256)