Requests-cache

Latest version: v1.2.0

Safety actively analyzes 631110 Python packages for vulnerabilities to keep your Python projects secure.

Scan your dependencies

Page 3 of 9

0.9.2

* Fix serialization in filesystem backend with binary content that is also valid UTF-8
* Fix some regression bugs introduced in 0.9.0:
* Add support for `params` as a positional argument to `CachedSession.request()`
* Add support for disabling expiration for a single request with `CachedSession.request(..., expire_after=-1)`

0.9.1

* Add support for python 3.10.2 and 3.9.10 (regarding resolving `ForwardRef` types during deserialization)
* Add support for key-only request parameters (regarding hashing request data for cache key creation)
* Reduce verbosity of log messages when encountering an invalid JSON request body

0.9.0

[See all issues and PRs for 0.9](https://github.com/requests-cache/requests-cache/milestone/4?closed=1)

🕗 **Expiration & headers:**
* Use `Cache-Control` **request** headers by default
* Add support for `Cache-Control: immutable`
* Add support for immediate expiration + revalidation with `Cache-Control: max-age=0` and `Expires: 0`
* Reset expiration for cached response when a `304 Not Modified` response is received
* Support `expire_after` param for `CachedSession.send()`

💾 **Backends:**
* **Filesystem:**
* Add better error message if parent path exists but isn't a directory
* **Redis:**
* Add optional integration with Redis TTL to improve performance for removing expired responses
* This is enabled by default, but may be disabled
* **SQLite:**
* Add better error message if parent path exists but isn't a directory

🚀 **Performance:**
* Fix duplicate read operation for checking whether to read from redirects cache
* Skip unnecessary contains check if a key is in the main responses cache
* Make per-request expiration thread-safe for both `CachedSession.request()` and `CachedSession.send()`
* Some micro-optimizations for request matching

đŸĒ˛ **Bugfixes:**
* Fix regression bug causing headers used for cache key to not guarantee sort order
* Handle some additional corner cases when normalizing request data
* Add support for `BaseCache` keyword arguments passed along with a backend instance
* Fix issue with cache headers not being used correctly if `cache_control=True` is used with an `expire_after` value
* Fix license metadata as shown on PyPI
* Fix `CachedResponse` serialization behavior when using stdlib `pickle` in a custom serializer

0.8.1

* Redact `ignored_parameters` from `CachedResponse.url` (if used for credentials or other sensitive info)
* Fix an incorrect debug log message about skipping cache write
* Add some additional aliases for `DbDict`, etc. so fully qualified imports don't break

0.8.0

[See all issues and PRs for 0.8](https://github.com/requests-cache/requests-cache/milestone/3?closed=1)

🕗 **Expiration & headers:**
* Add support for conditional requests and cache validation using:
* `ETag` + `If-None-Match` headers
* `Last-Modified` + `If-Modified-Since` headers
* `304 Not Modified` responses
* If a cached response is expired but contains a validator, a conditional request will by sent, and a new response will be cached and returned only if the remote content has not changed

💾 **Backends:**
* **Filesystem:**
* Add `FileCache.cache_dir` wrapper property
* Add `FileCache.paths()` method
* Add `use_cache_dir` option to use platform-specific user cache directory
* Return `pathlib.Path` objects for all file paths
* Use shorter hashes for file names
* **SQLite:**
* Add `SQLiteCache.db_path` wrapper property
* Add `use_memory` option and support for in-memory databases
* Add `use_cache_dir` option to use platform-specific user cache directory
* Return `pathlib.Path` objects for all file paths

🚀 **Performance:**
* Use `cattrs` by default for optimized serialization
* Slightly reduce size of serialized responses

↔ī¸ **Request matching:**
* Allow `create_key()` to optionally accept parameters for `requests.Request` instead of a request object
* Allow `match_headers` to optionally accept a list of specific headers to match
* Add support for custom cache key callbacks with `key_fn` parameter
* By default use blake2 instead of sha256 for generating cache keys

ℹī¸ **Cache convenience methods:**
* Add `BaseCache.update()` method as a shortcut for exporting to a different cache instance
* Allow `BaseCache.has_url()` and `delete_url()` to optionally take parameters for `requests.Request` instead of just a URL

đŸ“Ļ **Dependencies:**
* Add `appdirs` as a dependency for easier cross-platform usage of user cache directories
* Update `cattrs` from optional to required dependency
* Update `itsdangerous` from required to optional (but recommended) dependency
* Require `requests` 2.22+ and `urllib3` 1.25.5+

⚠ī¸ **Backwards-compatible API changes:**

The following changes are meant to make certain behaviors more obvious for new users, without breaking existing usage:
* For consistency with `Cache-Control: stale-if-error`, rename `old_data_on_error` to `stale_if_error`
* Going forward, any new options based on a standard HTTP caching feature will be named after that feature
* For clarity about matching behavior, rename `include_get_headers` to `match_headers`
* References in the docs to cache keys and related behavior are now described as 'request matching'
* For consistency with other backends, rename SQLite backend classes: `backends.sqlite.Db*` -> `SQLiteCache`, `SQLiteDict`, `SQLitePickleDict`
* Add aliases for all previous parameter/class names for backwards-compatibility

⚠ī¸ **Deprecations & removals:**
* Drop support for python 3.6
* Remove deprecated `core` module
* Remove deprecated `BaseCache.remove_old_entries()` method

-----

0.7.5

* Fix incorrect location of `redirects.sqlite` when using filesystem backend
* Fix issue in which `redirects.sqlite` would get included in response paths with filesystem backend
* Add aliases for forwards-compatibility with 0.8+
* Backport fixes from 0.8.1

Page 3 of 9

Š 2024 Safety CLI Cybersecurity Inc. All Rights Reserved.