v0.9.5
Fixed:
- aiohttp.ClientResponse.data should be awaited (#197) - thanks @cmbernard333

v0.9.4
- feat: support List Stores name filter (#181)
- feat: fix and improve retries and rate limit handling. (#176) - thanks @GMorris-professional
  The SDK now respects the rate limit headers (`Retry-After`) returned by the server and will retry the request after the specified time.
  If the header is not sent or on network errors, it will fall back to exponential backoff.
- feat: allow more user customizations for the token issuer (#186) - thanks @manuel-lang
- fix: ListRelations should not swallow errors (#183)
- fix: urllib3 compatibility < v2 (#187)

v0.9.3
- fix: urllib3 compatibility < v2 (#179)

v0.9.2
Fixed:
- fix(telemetry): fixes for telemetry attributes and metrics tracking
- fix: REST client should not close after `stream` request

v0.9.1
Added:
- feat: add `/streamed-list-objects` endpoint support (#163)
- feat: add `contextual_tuples` support for `/expand` endpoint requests (#164)

v0.9.0
Added:
- feat: add support for `start_time` parameter in `ReadChanges` endpoint (#156) - Note, this feature requires v1.8.0 of OpenFGA or newer
- feat!: add support for `BatchCheck` API (#154) - Note, this feature requires v1.8.2 of OpenFGA or newer
Changed:
- feat: remove client-side validation - thanks @GMorris-professional (#155)
- fix: change default max retry limit to 3 from 15 - thanks @ovindu-a (#155)
BREAKING CHANGE:
Usage of the existing batch_check should now use client_batch_check instead, additionally the existing
BatchCheckResponse has been renamed to ClientBatchCheckClientResponse.
Please see (#154)(#154) for more details on this change.

v0.8.1
Added:
- feat: allow specifying a request timeout (#151) - thanks @Oscmage!

v0.8.0
Added:
- feat: allow configuring the token endpoint (#137)
- feat: add per-HTTP request counter metric (#135)
Changed:
- refactor: remove SDK version for OpenTelemetry meter name (#134)
Fixed:
- fix: only send SDK method header from SDK wrapper methods (#142)
- fix: unable to pass `retry_params` (#144)
- fix: list users should send contextual tuples as a list (#147)
- fix: handle no models existing in `read_latest_authorization_model` (#147)

v0.7.2
Changed:
- Improve OpenTelemetry configuration (#127)
Fixed:
- Ensure max_parallel_requests is an int value in batch_check (#132)
- Inconsistency in 429 handling between sync/async client (#131)
- Ensure telemetry is reported when API exceptions are raised (#127)

v0.7.1
- fix: attribute values are now correctly exported as their intended types (previously, these were all sent as string values)
- fix: `http_client_request_duration` being reported in seconds rather than the intended milliseconds
- fix: sync client mistakenly passing the entire configuration (rather than just the OpenTelemetry configuration as intended) to `queryDuration()` and `requestDuration()`
- fix: some attributes may not have been exported as expected under some conditions
- fix: `queryDuration()` and `requestDuration()` may not have updated their histograms reliably when `attr_http_client_request_duration` or `attr_http_server_request_duration` (respectively) were not enabled (which is the default)