hishel 0.1.1__tar.gz → 0.1.3__tar.gz

{hishel-0.1.1 → hishel-0.1.3}/.gitignore

@@ -1,4 +1,5 @@
 venv/
 __pycache__/
 .coverage
-.cache/
+.cache/
+.idea/

{hishel-0.1.1 → hishel-0.1.3}/CHANGELOG.md

@@ -1,8 +1,23 @@
 # Changelog
 
+## 0.1.3 (1st July, 2025)
+
+- Remove `types-redis` from dev dependencies (#336)
+- Bump redis to 6.0.0 and address async `.close()` deprecation warning (#336)
+- Avoid race condition when unlinking files in `FileStorage`. (#334)  
+- Allow prodiving a `path_prefix` in `S3Storage` and `AsyncS3Storage`. (#342)
+
+## 0.1.2 (5th April, 2025)
+
+- Add check for fips compliant python. (#325)
+- Fix compatibility with httpx. (#291)
+- Use `SyncByteStream` instead of `ByteStream`. (#298)
+- Don't raise exceptions if date-containing headers are invalid. (#318)
+- Fix for S3 Storage missing metadata in API request. (#320)
+
 ## 0.1.1 (2nd Nov, 2024)
 
-- FIx typig extensions nor found. (#290)
+- Fix typing extensions not found. (#290)
 
 ## 0.1.0 (2nd Nov, 2024)
 

{hishel-0.1.1 → hishel-0.1.3}/PKG-INFO

@@ -1,6 +1,6 @@
-Metadata-Version: 2.3
+Metadata-Version: 2.4
 Name: hishel
-Version: 0.1.1
+Version: 0.1.3
 Summary: Persistent cache implementation for httpx and httpcore
 Project-URL: Homepage, https://hishel.com
 Project-URL: Source, https://github.com/karpetrosyan/hishel
@@ -23,16 +23,16 @@ Classifier: Programming Language :: Python :: 3.12
 Classifier: Programming Language :: Python :: 3.13
 Classifier: Topic :: Internet :: WWW/HTTP
 Requires-Python: >=3.9
-Requires-Dist: httpx>=0.22.0
+Requires-Dist: httpx>=0.28.0
 Provides-Extra: redis
-Requires-Dist: redis==5.0.1; extra == 'redis'
+Requires-Dist: redis==6.2.0; extra == 'redis'
 Provides-Extra: s3
 Requires-Dist: boto3<=1.15.3,>=1.15.0; (python_version < '3.12') and extra == 's3'
 Requires-Dist: boto3>=1.15.3; (python_version >= '3.12') and extra == 's3'
 Provides-Extra: sqlite
 Requires-Dist: anysqlite>=0.0.5; extra == 'sqlite'
 Provides-Extra: yaml
-Requires-Dist: pyyaml==6.0.1; extra == 'yaml'
+Requires-Dist: pyyaml==6.0.2; extra == 'yaml'
 Description-Content-Type: text/markdown
 
 <p align="center" class="logo">
@@ -181,9 +181,24 @@ Help us grow and continue developing good software for you ❤️
 
 # Changelog
 
+## 0.1.3 (1st July, 2025)
+
+- Remove `types-redis` from dev dependencies (#336)
+- Bump redis to 6.0.0 and address async `.close()` deprecation warning (#336)
+- Avoid race condition when unlinking files in `FileStorage`. (#334)  
+- Allow prodiving a `path_prefix` in `S3Storage` and `AsyncS3Storage`. (#342)
+
+## 0.1.2 (5th April, 2025)
+
+- Add check for fips compliant python. (#325)
+- Fix compatibility with httpx. (#291)
+- Use `SyncByteStream` instead of `ByteStream`. (#298)
+- Don't raise exceptions if date-containing headers are invalid. (#318)
+- Fix for S3 Storage missing metadata in API request. (#320)
+
 ## 0.1.1 (2nd Nov, 2024)
 
-- FIx typig extensions nor found. (#290)
+- Fix typing extensions not found. (#290)
 
 ## 0.1.0 (2nd Nov, 2024)
 

{hishel-0.1.1 → hishel-0.1.3}/hishel/__init__.py

@@ -14,4 +14,4 @@ def install_cache() -> None:  # pragma: no cover
     httpx.Client = CacheClient  # type: ignore
 
 
-__version__ = "0.1.1"
+__version__ = "0.1.3"

{hishel-0.1.1 → hishel-0.1.3}/hishel/_async/_pool.py

@@ -47,7 +47,7 @@ class AsyncCacheConnectionPool(AsyncRequestInterface):
 
         self._storage = storage if storage is not None else AsyncFileStorage(serializer=JSONSerializer())
 
-        if not isinstance(self._storage, AsyncBaseStorage):
+        if not isinstance(self._storage, AsyncBaseStorage):  # pragma: no cover
             raise TypeError(f"Expected subclass of `AsyncBaseStorage` but got `{storage.__class__.__name__}`")
 
         self._controller = controller if controller is not None else Controller()

{hishel-0.1.1 → hishel-0.1.3}/hishel/_async/_storages.py

@@ -159,7 +159,7 @@ class AsyncFileStorage(AsyncBaseStorage):
 
         async with self._lock:
             if response_path.exists():
-                response_path.unlink()
+                response_path.unlink(missing_ok=True)
 
     async def update_metadata(self, key: str, response: Response, request: Request, metadata: Metadata) -> None:
         """
@@ -222,7 +222,7 @@ class AsyncFileStorage(AsyncBaseStorage):
             if response_path.is_file():
                 age = time.time() - response_path.stat().st_mtime
                 if age > self._ttl:
-                    response_path.unlink()
+                    response_path.unlink(missing_ok=True)
             return
 
         self._last_cleaned = time.monotonic()
@@ -507,7 +507,7 @@ class AsyncRedisStorage(AsyncBaseStorage):
         return self._serializer.loads(cached_response)
 
     async def aclose(self) -> None:  # pragma: no cover
-        await self._client.close()
+        await self._client.aclose()
 
 
 class AsyncInMemoryStorage(AsyncBaseStorage):
@@ -654,6 +654,8 @@ class AsyncS3Storage(AsyncBaseStorage):  # pragma: no cover
     :type check_ttl_every: tp.Union[int, float]
     :param client: A client for S3, defaults to None
     :type client: tp.Optional[tp.Any], optional
+    :param path_prefix: A path prefix to use for S3 object keys, defaults to "hishel-"
+    :type path_prefix: str, optional
     """
 
     def __init__(
@@ -663,6 +665,7 @@ class AsyncS3Storage(AsyncBaseStorage):  # pragma: no cover
         ttl: tp.Optional[tp.Union[int, float]] = None,
         check_ttl_every: tp.Union[int, float] = 60,
         client: tp.Optional[tp.Any] = None,
+        path_prefix: str = "hishel-",
     ) -> None:
         super().__init__(serializer, ttl)
 
@@ -680,6 +683,7 @@ class AsyncS3Storage(AsyncBaseStorage):  # pragma: no cover
             bucket_name=bucket_name,
             is_binary=self._serializer.is_binary,
             check_ttl_every=check_ttl_every,
+            path_prefix=path_prefix,
         )
         self._lock = AsyncLock()
 

{hishel-0.1.1 → hishel-0.1.3}/hishel/_async/_transports.py

@@ -64,7 +64,7 @@ class AsyncCacheTransport(httpx.AsyncBaseTransport):
 
         self._storage = storage if storage is not None else AsyncFileStorage(serializer=JSONSerializer())
 
-        if not isinstance(self._storage, AsyncBaseStorage):
+        if not isinstance(self._storage, AsyncBaseStorage):  # pragma: no cover
             raise TypeError(f"Expected subclass of `AsyncBaseStorage` but got `{storage.__class__.__name__}`")
 
         self._controller = controller if controller is not None else Controller()
@@ -152,7 +152,7 @@ class AsyncCacheTransport(httpx.AsyncBaseTransport):
                 # Controller has determined that the response needs to be re-validated.
                 assert isinstance(res.stream, tp.AsyncIterable)
                 revalidation_request = Request(
-                    method=res.method,
+                    method=res.method.decode(),
                     url=normalized_url(res.url),
                     headers=res.headers,
                     stream=AsyncCacheStream(res.stream),

{hishel-0.1.1 → hishel-0.1.3}/hishel/_controller.py

@@ -60,8 +60,12 @@ def get_freshness_lifetime(response: Response) -> tp.Optional[int]:
     if header_presents(response.headers, b"expires"):
         expires = extract_header_values_decoded(response.headers, b"expires", single=True)[0]
         expires_timestamp = parse_date(expires)
+        if expires_timestamp is None:
+            return None
         date = extract_header_values_decoded(response.headers, b"date", single=True)[0]
         date_timestamp = parse_date(date)
+        if date_timestamp is None:
+            return None
 
         return expires_timestamp - date_timestamp
     return None
@@ -72,11 +76,12 @@ def get_heuristic_freshness(response: Response, clock: "BaseClock") -> int:
 
     if last_modified:
         last_modified_timestamp = parse_date(last_modified[0])
-        now = clock.now()
+        if last_modified_timestamp is not None:
+            now = clock.now()
 
-        ONE_WEEK = 604_800
+            ONE_WEEK = 604_800
 
-        return min(ONE_WEEK, int((now - last_modified_timestamp) * 0.1))
+            return min(ONE_WEEK, int((now - last_modified_timestamp) * 0.1))
 
     ONE_DAY = 86_400
     return ONE_DAY
@@ -89,6 +94,8 @@ def get_age(response: Response, clock: "BaseClock") -> int:
         return float("inf")  # type: ignore
 
     date = parse_date(extract_header_values_decoded(response.headers, b"date")[0])
+    if date is None:
+        return float("inf")  # type: ignore
 
     now = clock.now()
 

{hishel-0.1.1 → hishel-0.1.3}/hishel/_headers.py

@@ -94,13 +94,13 @@ def parse_cache_control(cache_control_values: List[str]) -> "CacheControl":
                         for value_char in value:
                             if value_char not in tchar:
                                 raise ParseError(
-                                    f"The character '{value_char!r}' " "is not permitted for the unquoted values."
+                                    f"The character '{value_char!r}' is not permitted for the unquoted values."
                                 )
                     else:
                         for value_char in value[1:-1]:
                             if value_char not in qdtext:
                                 raise ParseError(
-                                    f"The character '{value_char!r}' " "is not permitted for the quoted values."
+                                    f"The character '{value_char!r}' is not permitted for the quoted values."
                                 )
                     break
 

{hishel-0.1.1 → hishel-0.1.3}/hishel/_s3.py

@@ -11,16 +11,22 @@ def get_timestamp_in_ms() -> float:
 
 class S3Manager:
     def __init__(
-        self, client: tp.Any, bucket_name: str, check_ttl_every: tp.Union[int, float], is_binary: bool = False
+        self,
+        client: tp.Any,
+        bucket_name: str,
+        check_ttl_every: tp.Union[int, float],
+        is_binary: bool = False,
+        path_prefix: str = "hishel-",
     ):
         self._client = client
         self._bucket_name = bucket_name
         self._is_binary = is_binary
         self._last_cleaned = time.monotonic()
         self._check_ttl_every = check_ttl_every
+        self._path_prefix = path_prefix
 
     def write_to(self, path: str, data: tp.Union[bytes, str], only_metadata: bool = False) -> None:
-        path = "hishel-" + path
+        path = self._path_prefix + path
         if isinstance(data, str):
             data = data.encode("utf-8")
 
@@ -43,7 +49,7 @@ class S3Manager:
         )
 
     def read_from(self, path: str) -> tp.Union[bytes, str]:
-        path = "hishel-" + path
+        path = self._path_prefix + path
         response = self._client.get_object(
             Bucket=self._bucket_name,
             Key=path,
@@ -57,7 +63,7 @@ class S3Manager:
         return tp.cast(str, content.decode("utf-8"))
 
     def remove_expired(self, ttl: int, key: str) -> None:
-        path = "hishel-" + key
+        path = self._path_prefix + key
 
         if time.monotonic() - self._last_cleaned < self._check_ttl_every:
             try:
@@ -72,22 +78,36 @@ class S3Manager:
 
         self._last_cleaned = time.monotonic()
         for obj in self._client.list_objects(Bucket=self._bucket_name).get("Contents", []):
-            if not obj["Key"].startswith("hishel-"):  # pragma: no cover
+            if not obj["Key"].startswith(self._path_prefix):  # pragma: no cover
                 continue
 
-            if get_timestamp_in_ms() - float(obj["Metadata"]["created_at"]) > ttl:
+            try:
+                metadata_obj = self._client.head_object(Bucket=self._bucket_name, Key=obj["Key"]).get("Metadata", {})
+            except ClientError as e:
+                if e.response["Error"]["Code"] == "404":
+                    continue
+
+            if not metadata_obj or "created_at" not in metadata_obj:
+                continue
+
+            if get_timestamp_in_ms() - float(metadata_obj["created_at"]) > ttl:
                 self._client.delete_object(Bucket=self._bucket_name, Key=obj["Key"])
 
     def remove_entry(self, key: str) -> None:
-        path = "hishel-" + key
+        path = self._path_prefix + key
         self._client.delete_object(Bucket=self._bucket_name, Key=path)
 
 
 class AsyncS3Manager:  # pragma: no cover
     def __init__(
-        self, client: tp.Any, bucket_name: str, check_ttl_every: tp.Union[int, float], is_binary: bool = False
+        self,
+        client: tp.Any,
+        bucket_name: str,
+        check_ttl_every: tp.Union[int, float],
+        is_binary: bool = False,
+        path_prefix: str = "hishel-",
     ):
-        self._sync_manager = S3Manager(client, bucket_name, check_ttl_every, is_binary)
+        self._sync_manager = S3Manager(client, bucket_name, check_ttl_every, is_binary, path_prefix)
 
     async def write_to(self, path: str, data: tp.Union[bytes, str], only_metadata: bool = False) -> None:
         return await to_thread.run_sync(self._sync_manager.write_to, path, data, only_metadata)

{hishel-0.1.1 → hishel-0.1.3}/hishel/_sync/_pool.py

@@ -47,7 +47,7 @@ class CacheConnectionPool(RequestInterface):
 
         self._storage = storage if storage is not None else FileStorage(serializer=JSONSerializer())
 
-        if not isinstance(self._storage, BaseStorage):
+        if not isinstance(self._storage, BaseStorage):  # pragma: no cover
             raise TypeError(f"Expected subclass of `BaseStorage` but got `{storage.__class__.__name__}`")
 
         self._controller = controller if controller is not None else Controller()

{hishel-0.1.1 → hishel-0.1.3}/hishel/_sync/_storages.py

@@ -159,7 +159,7 @@ class FileStorage(BaseStorage):
 
         with self._lock:
             if response_path.exists():
-                response_path.unlink()
+                response_path.unlink(missing_ok=True)
 
     def update_metadata(self, key: str, response: Response, request: Request, metadata: Metadata) -> None:
         """
@@ -222,7 +222,7 @@ class FileStorage(BaseStorage):
             if response_path.is_file():
                 age = time.time() - response_path.stat().st_mtime
                 if age > self._ttl:
-                    response_path.unlink()
+                    response_path.unlink(missing_ok=True)
             return
 
         self._last_cleaned = time.monotonic()
@@ -654,6 +654,8 @@ class S3Storage(BaseStorage):  # pragma: no cover
     :type check_ttl_every: tp.Union[int, float]
     :param client: A client for S3, defaults to None
     :type client: tp.Optional[tp.Any], optional
+    :param path_prefix: A path prefix to use for S3 object keys, defaults to "hishel-"
+    :type path_prefix: str, optional
     """
 
     def __init__(
@@ -663,6 +665,7 @@ class S3Storage(BaseStorage):  # pragma: no cover
         ttl: tp.Optional[tp.Union[int, float]] = None,
         check_ttl_every: tp.Union[int, float] = 60,
         client: tp.Optional[tp.Any] = None,
+        path_prefix: str = "hishel-",
     ) -> None:
         super().__init__(serializer, ttl)
 
@@ -680,6 +683,7 @@ class S3Storage(BaseStorage):  # pragma: no cover
             bucket_name=bucket_name,
             is_binary=self._serializer.is_binary,
             check_ttl_every=check_ttl_every,
+            path_prefix=path_prefix,
         )
         self._lock = Lock()
 

{hishel-0.1.1 → hishel-0.1.3}/hishel/_sync/_transports.py

@@ -5,7 +5,7 @@ import typing as tp
 
 import httpcore
 import httpx
-from httpx import ByteStream, Request, Response
+from httpx import SyncByteStream, Request, Response
 from httpx._exceptions import ConnectError
 
 from hishel._utils import extract_header_values_decoded, normalized_url
@@ -29,7 +29,7 @@ def generate_504() -> Response:
     return Response(status_code=504)
 
 
-class CacheStream(ByteStream):
+class CacheStream(SyncByteStream):
     def __init__(self, httpcore_stream: tp.Iterable[bytes]):
         self._httpcore_stream = httpcore_stream
 
@@ -64,7 +64,7 @@ class CacheTransport(httpx.BaseTransport):
 
         self._storage = storage if storage is not None else FileStorage(serializer=JSONSerializer())
 
-        if not isinstance(self._storage, BaseStorage):
+        if not isinstance(self._storage, BaseStorage):  # pragma: no cover
             raise TypeError(f"Expected subclass of `BaseStorage` but got `{storage.__class__.__name__}`")
 
         self._controller = controller if controller is not None else Controller()
@@ -152,7 +152,7 @@ class CacheTransport(httpx.BaseTransport):
                 # Controller has determined that the response needs to be re-validated.
                 assert isinstance(res.stream, tp.Iterable)
                 revalidation_request = Request(
-                    method=res.method,
+                    method=res.method.decode(),
                     url=normalized_url(res.url),
                     headers=res.headers,
                     stream=CacheStream(res.stream),

{hishel-0.1.1 → hishel-0.1.3}/hishel/_utils.py

@@ -1,8 +1,8 @@
 import calendar
+import hashlib
 import time
 import typing as tp
 from email.utils import parsedate_tz
-from hashlib import blake2b
 
 import anyio
 import httpcore
@@ -30,7 +30,7 @@ def normalized_url(url: tp.Union[httpcore.URL, str, bytes]) -> str:
 
     if isinstance(url, httpcore.URL):
         port = f":{url.port}" if url.port is not None else ""
-        return f'{url.scheme.decode("ascii")}://{url.host.decode("ascii")}{port}{url.target.decode("ascii")}'
+        return f"{url.scheme.decode('ascii')}://{url.host.decode('ascii')}{port}{url.target.decode('ascii')}"
     assert False, "Invalid type for `normalized_url`"  # pragma: no cover
 
 
@@ -49,10 +49,26 @@ def generate_key(request: httpcore.Request, body: bytes = b"") -> str:
 
     key_parts = [request.method, encoded_url, body]
 
-    key = blake2b(digest_size=16, usedforsecurity=False)
-    for part in key_parts:
-        key.update(part)
-    return key.hexdigest()
+    # FIPs mode disables blake2 algorithm, use sha256 instead when not found.
+    blake2b_hasher = None
+    sha256_hasher = hashlib.sha256(usedforsecurity=False)
+    try:
+        blake2b_hasher = hashlib.blake2b(digest_size=16, usedforsecurity=False)
+    except (ValueError, TypeError, AttributeError):
+        pass
+
+    hexdigest: str
+    if blake2b_hasher:
+        for part in key_parts:
+            blake2b_hasher.update(part)
+
+        hexdigest = blake2b_hasher.hexdigest()
+    else:
+        for part in key_parts:
+            sha256_hasher.update(part)
+
+        hexdigest = sha256_hasher.hexdigest()
+    return hexdigest
 
 
 def extract_header_values(
@@ -82,9 +98,11 @@ def header_presents(headers: tp.List[tp.Tuple[bytes, bytes]], header_key: bytes)
     return bool(extract_header_values(headers, header_key, single=True))
 
 
-def parse_date(date: str) -> int:
+def parse_date(date: str) -> tp.Optional[int]:
     expires = parsedate_tz(date)
-    timestamp = calendar.timegm(expires[:6])  # type: ignore
+    if expires is None:
+        return None
+    timestamp = calendar.timegm(expires[:6])
     return timestamp
 
 

{hishel-0.1.1 → hishel-0.1.3}/pyproject.toml

@@ -29,17 +29,17 @@ classifiers = [
     "Topic :: Internet :: WWW/HTTP",
 ]
 dependencies = [
-    "httpx>=0.22.0",
+    "httpx>=0.28.0",
 ]
 
 [project.optional-dependencies]
 
 yaml = [
-    "pyyaml==6.0.1",
+    "pyyaml==6.0.2",
 ]
 
 redis = [
-    "redis==5.0.1"
+    "redis==6.2.0"
 ]
 
 sqlite = [
@@ -92,7 +92,7 @@ filterwarnings = []
 
 [tool.coverage.run]
 omit = [
-    "venv/*", 
+    "venv/*",
     "hishel/_sync/*",
     "hishel/_s3.py"
 ]
@@ -122,3 +122,19 @@ select = [
 
 [tool.ruff.lint.isort]
 combine-as-imports = true
+
+[dependency-groups]
+dev = [
+    "anyio==4.7.0",
+    "coverage==7.6.10",
+    "hatch==1.9.3",
+    "mkdocs==1.6.1",
+    "mkdocs-material==9.5.1",
+    "mypy==1.14.1",
+    "pytest==8.3.4",
+    "ruff==0.11.0",
+    "trio==0.28.0",
+    "types-boto3==1.0.2",
+    "types-pyyaml==6.0.12.20240311",
+    "zipp>=3.19.1",
+]