Change log

All notable changes to Safir will be documented in this file.

Versioning follows semver.

This project uses scriv to maintain the change log. Changes for the upcoming release can be found in changelog.d.

9.0.1 (2024-12-11)

Bug fixes

  • Fix dependencies on safir-logging and safir-arq from the safir PyPI module to allow the latest versions to be installed.

9.0.0 (2024-12-11)

Backwards-incompatible changes

  • Rewrite the Safir UWS support to use Pydantic models for job parameters. Services built on the Safir UWS library will need to change all job creation dependencies to return Pydantic models.

  • UWS clients must now pass an additional job_summary_type argument to UWSAppSettings.build_uws_config and implement to_xml_model in their implementation of ParametersModel, returning a subclass of the vo-models Parameters class.

  • Use the Wobbly service rather than a direct database connection to store UWS job information. Services built on the Safir UWS library must now configure a Wobbly URL and will switch to Wobbly’s storage instead of their own when updated to this release of Safir.

  • Case-insensitivity of form POST parameters to UWS routes is now handled by middleware, and the uws_post_params_dependency function has been removed. Input parameter dependencies for UWS applications can now assume that all parameter keys will be in lowercase.

  • Support an execution duration of 0 in the Safir UWS library, mapping it to no limit on the execution duration. Note that this will not be allowed by the default configuration and must be explicitly allowed by an execution duration validation hook.

  • Convert all models returned by the Safir UWS library to Pydantic. Services built on the Safir UWS library will have to change the types of validator functions for destruction time and execution duration.

  • Safir no longer provides the safir.uws.ErrorCode enum or the exception safir.uws.MultiValuedParameterError. These values were specific to a SODA service, and different IVOA UWS services use different error codes. The Safir UWS library now takes error code as a string, and each application should define its own set of error codes in accordance with the IVOA standard it is implementing.

New features

  • Add unit testing support for application metrics. Tests can define a mock application metrics event handler and inspect it after running application code to see what events would have been published.

Bug fixes

  • Append a colon after the error code when reporting UWS errors.

Other changes

  • Render all UWS XML output except for error VOTables using vo-models rather than hand-written XML templates.

8.0.0 (2024-11-26)

Backwards-incompatible changes

  • Add serializers to HumanTimedelta and SecondsTimedelta that serialize those Pydantic fields to a float number of seconds instead of ISO 8601 durations. This means those data types now can be round-tripped (serialized and then deserialized to the original value), whereas before they could not be.

  • parse_isodatetime and normalize_isodatetime now accept exactly the date formats accepted by the IVOA DALI standard. This means seconds are now required, the trailing Z is now optional (times are always interpreted as UTC regardless), and the time is optional and interpreted as 00:00:00 if missing.

New features

  • Add new safir.pydantic.UtcDatetime type that is equivalent to datetime but coerces all incoming times to timezone-aware UTC. This type should be used instead of using normalize_datetime as a validator.

  • Add new safir.pydantic.IvoaIsoDatetime type that accepts any ISO 8601 date and time that matches the IVOA DALI standard for timestamps. This follows the same rules as parse_isodatetime now follows. This type should be used instead of using normalize_isodatetime as a validator.

  • Add new safir.database.PaginatedLinkData model that parses the contents of an HTTP Link header and extracts pagination information.

  • Add safir.database.drop_database utility function to drop all database tables mentioned in a SQLAlchemy ORM schema and delete the Alembic version information if it is present. This is primarily useful for tests.

  • Publishing a metrics event no longer waits on confirmation of message delivery to Kafka. This makes publishing much more performant. All events will still be delivered as long as an app awaits EventManager.aclose in its lifecycle.

  • safir.kafka.PydanticSchemaManager takes an optional structlog BoundLogger. If not provided, the default logger is a BoundLogger, rather than a logging.Logger.

  • Pass the logger parameter to safir.metrics.KafkaMetricsConfiguration.make_manager to the PydanticSchemaManager instance that it creates.

  • Add CaseInsensitiveFormMiddleware to lower-case the keys of form POST parameters. This can be used to support the case-insensitive keys requirement of IVOA standards.

Bug fixes

  • Correctly validate stringified floating-point numbers of seconds as inputs to the SecondsTimedelta type instead of truncating it to an integer.

  • Allow timedelta as a member of a union field in a safir.metrics.EventPayload.

  • Add missing dependency on alembic to the safir[uws] extra.

Other changes

  • Document how to test an application’s database schema against its Alembic migrations to ensure that the schema hasn’t been modified without adding a corresponding migration.

7.0.0 (2024-11-08)

Backwards-incompatible changes

  • Remove EventDuration from safir.metrics. Instead, use Python’s built-in timedelta as the type for any field that previously used EventDuration. It will be serialized the same way, as an Avro double number of seconds.

Bug fixes

  • Support multi-valued parameters as input to UWS POST routes when the values are specified by repeating the parameter.

  • Declare that the safir.metrics package is typed by adding a py.typed file.

6.5.1 (2024-10-21)

Bug fixes

  • Allow and ignore extra attributes to MetricsConfiguration when metrics are disabled. This allows passing a partial metrics configuration even when they are disabled, which simplifies the structure of Phalanx configurations based on dumping the Helm values into a YAML configuraiton file.

6.5.0 (2024-10-18)

New features

  • Add new safir.kafka module to simplify configuring and constructing Kafka clients. Included in this module is safir.kafka.PydanticSchemaManager, which supports registering and evolving Avro schemas in the Confluent schema registry via Pydantic models. Based on @jonathansick’s kafkit work.

  • Add new safir.metrics module to enable publishing app metrics events.

Bug fixes

  • Fix integration issues with vo-models in the UWS support.

6.4.0 (2024-09-16)

New features

  • Add support functions for using Alembic to manage database schema migrations and to verify that the current database schema matches the expectations of the running application. The support in Safir is designed to use asyncpg for all database operations, which avoids having to add a separate dependency on a sync database client.

  • Add Alembic schema management support to the UWS library. This support is likely temporary and will probably be replaced in the future with a standalone UWS database service.

6.3.0 (2024-08-20)

New features

  • safir.logging is now available as a separate PyPI package, safir-logging, so that it can be installed in environments where the full Safir dependency may be too heavy-weight or conflict with other packages. Packages that do depend on safir can and should ignore this change and continue to assume depending on safir will be sufficient to provide safir.logging.

  • Allow the database URL passed to DatabaseSessionDependency.initialize to be a Pydantic Url. This simplifies logic for applications that use EnvAsyncPostgresDsn or other Pydantic URL types for their configuration.

  • Allow the hook URL argument to safir.testing.slack.mock_slack_webhook to be a Pydantic SecretStr.

Bug fixes

  • safir.logging previously exported a logger_name variable holding the logger name configured via safir.logging.configure_logging. This variable was never documented and was not intended for use outside of the library. It is no longer exported, has been renamed, and is now private to the library.

  • Fix construction of an Availability object reporting problems with the UWS database layer to use the correct field names and data type for the model.

6.2.0 (2024-08-02)

New features

  • Add new safir.uws and safir.arq.uws modules that provide the framework of an IVOA Universal Worker Service implementation.

  • Add new safir.testing.uws module that provides a mock UWS job runner for testing UWS applications.

  • safir.arq is now available as a separate PyPI package, safir-arq, so that it can be installed in environments where the full Safir dependency may be too heavy-weight or conflict with other packages. The safir[arq] dependency will continue to work as before (by installing safir-arq behind the scenes).

  • Add new abort_job method to instances of safir.arq.ArqQueue, which tells arq to abort a job that has been queued or in progress. To successfully abort a job that has already started, the arq worker must enable support for aborting jobs.

  • Add new utility function safir.arq.build_arq_redis_settings, which constructs the RedisSettings object used to create an arq Redis queue from a Pydantic Redis DSN.

  • Add new safir.arq.WorkerSettings class that models the acceptable parameters for an arq WorkerSettings object or class that Safir applications have needed.

  • Add new safir.pydantic.SecondsDatetime and safir.pydantic.HumanDatetime types for use in Pydantic models. These behave the same as datetime.timedelta fields but use custom validation. Both support a stringified number of seconds as input, and the latter also supports the interval strings parsed by safir.datetime.parse_timedelta.

  • Add new types safir.pydantic.EnvAsyncPostgresDsn and safir.pydantic.EnvRedisDsn, which validate PostgreSQL and Redis DSNs but rewrite them based on the environment variables set by tox-docker. Programs using these types for their configuration will therefore automatically honor tox-docker environment variables when running the test suite. EnvAsyncPostgresDsn also enforces that the scheme of the DSN is compatible with asyncpg and the Safir database support.

  • Add the decorator safir.database.retry_async_transaction, which retries a function or method a configurable number of times if it raises a SQLAlchemy exception from the underlying database API. This is primarily intended to retry database operations aborted due to transaction isolation.

  • safir.database.create_database_engine now accepts the database URL as a Pydantic Url as well as a str.

  • Allow the Slack webhook URL argument to SlackWebhookClient and SlackRouteErrorHandler to be given as a Pydantic SecretStr instead of a str. This simplifies code in applications that get that value from a secret.

Other changes

  • Safir is now built with nox instead of tox.

6.1.0 (2024-07-12)

New features

  • Add pull_requests to GitHubCheckSuiteModel to capture info about any pull requests associated with a GitHub check suite event.

6.0.0 (2024-06-10)

Backwards-incompatible changes

  • Drop safir.database.create_sync_session. This was only used by services that used Dramatiq for task management, since Dramatiq is sync-only. Services based on Safir should switch to arq and use only async database connections.

  • Drop DatabaseSessionDependency.override_engine. This was used for Gafaelfawr to share a database engine across all tests for speed, but this technique breaks with current versions of pytest-asyncio and is no longer used or safe to use.

New features

  • Allow the database password to be passed to create_database_engine and DatabaseSessionDependency.initialize as a Pydantic SecretStr.

  • Add new function safir.datetime.parse_timedelta, which parses a human-friendly syntax for specifying time durations into a Python datetime.timedelta.

  • Add support for gs URLs to safir.gcs.SignedURLService.

  • Support pickling of SlackException so that subclasses of it can be thrown by arq workers and unpickled correctly when retrieving results.

Bug fixes

  • Correctly honor the default_queue_name argument to RedisArqQueue.initialize.

5.2.2 (2024-03-15)

Bug fixes

  • Ensure that per-request database sessions provided by db_session_dependency are cleaned up even if the request aborts with an uncaught exception.

5.2.1 (2024-02-19)

Bug fixes

  • Fix the return type of safir.datetime.parse_isodatetime to not include None since the function never returns None.

5.2.0 (2024-01-19)

New features

  • Add a FastAPI dependency for retrieving a Gafaelfawr delegated token from the request headers: safir.dependencies.gafaelfawr.auth_delegated_token_dependency.

Bug fixes

  • Rewrite CaseInsensitiveQueryMiddleware and XForwardedMiddleware as pure ASGI middleware rather than using the Starlette BaseHTTPMiddleware class. The latter seems to be behind some poor error reporting of application exceptions, has caused problems in the past due to its complexity, and is not used internally by Starlette middleware.

5.1.0 (2023-12-07)

New features

  • Add support for label selectors in the list_node method of the Kubernetes mock.

5.0.0 (2023-12-05)

Backwards-incompatible changes

  • Safir now depends on Pydantic v2. Python code that uses any part of Safir related to Pydantic will also need to update to Pydantic v2, since the API is significantly different. See the Pydantic migration guide for more information.

  • safir.pydantic.validate_exactly_one_of is now a Pydantic model validator. It must be called with mode="after", since it operates in the model rather than on a raw dictionary.

  • Remove the GitHubAppClientFactory.create_app_client method, which does not work with the Gidgethub API. Instead, the documentation shows how to create a JWT with the GitHubAppClientFactory and pass it with requests.

  • safir.github.GitHubAppClientFactory now expects the application ID and installation ID (for create_installation_client) to be of type int, not str. This appears to match what GitHub’s API returns, but not what Gidgethub expects. The ID is converted to a string when passing it to Gidgethub.

New features

  • Allow the safir.logging.LogLevel enum to be created from strings of any case, which will allow the logging level to be specified with any case for Safir applications that use Pydantic to validate the field.

  • Add validated but ignored optional propagation_policy arguments to every delete method of the Kubernetes mock for better compatibility with the actual Kubernetes API. Previously, this argument was only accepted by delete_namespaced_job.

  • All mock Kubernetes methods now accept and ignore a _request_timeout error for better compatibility with the Kubernetes API.

  • Add delete, list, and watch support for persistent volume claims to the Kubernetes mock.

Bug fixes

  • safir.database.datetime_to_db, safir.datetime.format_datetime_for_logging, and safir.datetime.isodatetime now accept any datetime object with a time zone whose offset from UTC is 0, rather than only the datetime.UTC time zone object.

  • safir.pydantic.normalize_datetime now explicitly rejects input other than seconds since epoch or datetime objects with a validation error rather than attempting to treat the input as a datetime object and potentially throwing more obscure errors.

  • The _request_timeout parameters to mock Kubernetes methods now accept a float instead of an int to more correctly match the types of kubernetes_asyncio. The mock still does not accept a tuple of timeouts.

  • Avoid reusing the same metadata object when creating a Pod from a Job. Previous versions modified the spec part of the Job when adding additional metadata to the child Pod.

Other changes

  • Safir is now tested with Python 3.12 as well as Python 3.11.

4.5.0 (2023-09-12)

New features

  • Add list_namespaced_custom_object with watch support to the Kubernetes mock.

  • Add watch, field selector, and label selector support to list_namespace in the Kubernetes mock.

Bug fixes

  • The Kubernetes mock now correctly maintains the resource version of Ingress, Job, and Service objects, since they support watches which rely on resource versions.

  • When creating a Pod from a Job in the Kubernetes mock using generateName, randomize the Pod name like Kubernetes does rather than using a fixed name. This forces tests to scan correctly for pods associated with a job. If the Pod name or generateName was explicitly configured in the Job template, honor it.

  • read_namespace and list_namespace in the Kubernetes mock now only return namespace objects that have been explicitly created, not implicit namespaces created by creating another object without making a namespace first. This more closely matches the behavior of Kubernetes while still making it easy to use the mock in a test environment simulating a pre-existing namespace.

4.4.0 (2023-09-07)

New features

  • Add a safir.click.display_help helper function that implements a help command for Click-based command-line interfaces, with support for nested subcommands.

  • Add a new safir.asyncio.AsyncMultiQueue data structure, which is an asyncio multi-reader queue that delivers all messages to each reader independently.

  • Add read_ methods for the Kubernetes object types for which the mock provided create_ methods (NetworkPolicy and PersistentVolumeClaim).

Bug fixes

  • Fix typing of the safir.asyncio.run_with_asyncio decorator so that it doesn’t mask the type of the underlying function.

  • Kubernetes objects included in events are now serialized properly using the Kubernetes camel-case field names instead of the Python snake-case names. In addition to matching Kubernetes behavior more closely, this allows a watch configured with the Kubernetes model type to deserialize the object in the object key of the event dictionary. The type must be passed explicitly to the Watch constructor, since kubernetes_asyncio’s type autodetection does not work with Safir’s mock.

  • safir.testing.kubernetes.patch_kubernetes no longer mocks the entire ApiClient class since it is required for deserialization of objects in Kubernetes events. It instead mocks the request method of that class for safety, to prevent any network requests to Kubernetes clusters when Kubernetes is mocked.

Other changes

  • Safir now uses the Ruff linter instead of flake8 and isort.

4.3.1 (2023-07-17)

Bug fixes

  • Safir now pins the major version of all of its non-development dependencies. The impetus for this change is to prevent upgrades to Pydantic 2.x until Safir’s Pydantic models are ready for that upgrade, but a similar principle applies to other dependencies. These pins will be selectively relaxed once Safir has been confirmed to work with a new major release.

4.3.0 (2023-05-23)

New features

  • All delete_* APIs in the mock Kubernetes API now support grace_period_seconds and a V1DeleteOptions body. Both are ignored.

  • delete_namespaced_job in the mock Kubernetes API now requires propagation_policy to be passed as a keyword argument if provided, and does not require it be set. It is validated against the values recognized by Kubernetes, and if set to Orphan, pods created by the job are not deleted.

Bug fixes

  • When reporting an error response to Slack using SlackWebException, put the response body in an attachment instead of a regular block, since it may be a full HTML error page. An attachment tells Slack to hide long content by default unless the viewer expands it.

4.2.2 (2023-05-17)

Bug fixes

  • Revert the documentation change in 4.2.1 to restore cross-references, since the docs-linkcheck failure appears to be a false positive.

Other changes

  • Safir now uses scriv to maintain its change log.

4.2.1 (2023-05-17)

Bug fixes

  • Fix syntax of literals in the MockKubernetesApi docstring.

4.2.0 (2023-05-17)

New features

  • Add create, delete, read, list (with watches), and (limited) patch support for Ingress objects to the mock Kubernetes API.

  • Add create, delete, read, and list (with watches) support for Job objects to the mock Kuberntes API, and mock the BatchV1Api.

  • Add delete, read, and list (with watches) support for Service objects to the mock Kubernetes API.

  • Add support for label selectors to list_namespaced_pod in the mock Kubernetes API.

  • Add create_namespaced_persistent_volume_claim to the mock Kubernetes API for testing.

  • Add support for deleting custom resources to the mock Kubernetes API.

Bug fixes

  • SlackWebException now extracts the method and URL of the request from more httpx exceptions, and avoids exceptions when the request information is not present.

4.1.0 (2023-05-08)

New features

  • Add read_* methods for ConfigMap and ResourceQuota to the mock Kubernetes API for testing.

  • Add patch_namespaced_pod_status to the mock Kubernetes API for testing. Application code is unlikely to call this, but it’s useful for test suites.

  • The mock list_namespaced_pod Kubernetes API now supports watches (but be aware that all changes must be made through the API).

Bug fixes

  • Fix concurrency locking when watching namespace events in the Kubernetes testing mock. The previous logic degenerated into busy-waiting rather than correctly waiting on a condition variable.

  • Watches for namespace events using the mock Kubernetes API now start with the next event, not the first stored event, if resource_version is not set, aligning behavior with the Kubernetes API.

4.0.0 (2023-04-19)

Backwards-incompatible changes

  • Safir now requires a minimum Python version of 3.11.

  • safir.pydantic.validate_exactly_one_of is now a Pydantic root validator instead of an individual field validator, which simplifies how it should be called.

  • Custom Kubernetes objects are no longer copied before being stored in the Kubernetes mock by create_namespaced_custom_object, and no longer automatically get a metadata.uid value. This brings handling of custom objects in line with the behavior of other mocked create_* and replace_* APIs.

  • All objects stored in the Kubernetes mock will be modified to set api_version, kind, and (where appropriate) namespace. If any of those values are set in the object, they are checked for correctness and rejected with AssertionError if they don’t match the API call and its parameters.

  • The mocked create_* Kubernetes APIs now use the name body for the parameter containing the API object, instead of using some other name specific to the kind of object. This fixes compatibility with the Python Kubernetes API that this class is mocking.

New features

  • The new safir.redis.PydanticRedisStorage class enables you to store Pydantic model objects in Redis. A safir.redis.EncryptedPydanticRedisStorage class also encrypts data in Redis. To use the safir.redis module, install Safir with the redis extra (i.e., pip install safir[redis]).

  • Add safir.fastapi.ClientRequestError base class for exceptions and the corresponding exception handler safir.fastapi.client_request_error_handler. These may be used together to raise structured exceptions and automatically transform them into structured HTTP errors with the correct HTTP status code and an error body compatible with safir.models.ErrorModel and FastAPI’s own internally-generated errors.

  • Add safir.slack.webhook.SlackWebException, which is a child class of safir.slack.webhook.SlackException that knows how to capture and report details from an underlying HTTPX exception.

  • Add a safir.testing.kubernetes.strip_none helper function that makes it easier to compare Kubernetes objects against expected test data.

  • Add a get_namespace_objects_for_test method to the Kubernetes mock to retrieve all objects (of any kind) in a namespace.

  • Add a mock for the list_nodes Kubernetes API, which returns data set by a call to the new set_nodes_for_test method.

  • Add optional rudimentary namespace tracking to the Kubernetes mock. Namespaces can be created, deleted (which deletes all objects in the namespace), read, and listed. Explicit namespace creation remains optional; if an object is created in a namespace and that namespace does not exist, one will be implicitly created by the mock.

  • Add support for namespaced events (the older core API, not the new events API) to the Kubernetes mock. Newly-created pods with the default initial status post an event by default; otherwise, events must be created by calling the mocked create_namespaced_event API. The list_namespaced_event API supports watches with timeouts.

  • Add rudimentary support for NetworkPolicy, ResourceQuota, and Service objects to the Kubernetes mock.

  • Add a safir.github module for writing GitHub integrations with GidgetHub and Pydantic modeling. safir.github.GitHubAppClientFactory helps create authenticated clients for GitHub Apps and app installations. safir.github.models contains Pydantic models for GitHub v3 REST API resources. safir.github.webhooks contains Pydantic models for GitHub webhook payloads.

Bug fixes

  • Stop adding the X-Auth-Request-User header to the OpenAPI schema for routes that use auth_dependency or auth_logger_dependency.

Other changes

  • The safir.testing.kubernetes.MockKubernetesApi mock now has rudimentary API documentation for the Kubernetes APIs that it supports.

  • The documentation for running commands with tox has been updated for the new command-line syntax in tox v4.

3.8.0 (2023-03-15)

New features

  • Add safir.slack.webhook.SlackWebhookClient and accompanying models to post a structured message to Slack via an incoming webhook. Add a safir.slack.blockkit.SlackException base class that can be used to create exceptions with supplemental metadata that can be sent to Slack as a formatted alert.

  • Add a FastAPI route class (safir.slack.webhook.SlackRouteErrorHandler) that reports all uncaught exceptions to Slack using an incoming webhook.

  • Add safir.datetime.format_datetime_for_logging to convert a datetime object into an easily human-readable representation.

  • Add safir.testing.slack.mock_slack_webhook and an associated mock class to mock a Slack webhook for testing.

  • Add microseconds=True parameter to safir.datetime.current_datetime to get a datetime object with microseconds filled in. By default, current_datetime suppresses the microseconds since databases often cannot store them, but there are some timestamp uses, such as error reporting, that benefit from microseconds and are never stored in databases.

3.7.0 (2023-03-06)

New features

  • Add a safir.testing.uvicorn.spawn_uvicorn helper function to spawn an ASGI app under an external Uvicorn process. Normally, ASGI apps should be tested by passing the app directly to an httpx.AsyncClient, but in some cases (such as Selenium testing) the app needs to listen to regular HTTP requests.

  • safir.database.initialize_database now creates a non-default schema if the underlying SQLAlchemy model declares one.

Bug fixes

  • After safir.logging.configure_uvicorn_logging was called, exception tracebacks were no longer expanded. Add the missing processor to the logging stack.

  • In safir.logging.configure_uvicorn_logging, send the access log to standard output instead of combining it with all other messages on standard error.

3.6.0 (2023-02-03)

New features

  • Add a safir.models.ErrorLocation enum holding valid values for the first element of the loc array in safir.models.ErrorModel. Due to limitations in Python typing, loc is not type-checked against this enum, but it may be useful to applications constructing FastAPI-compatible error messages.

  • Add safir.datetime.current_datetime to get a normalized datetime object representing the current date and time.

  • Add safir.datetime.isodatetime and safir.datetime.parse_isodatetime to convert to and from the most useful form of ISO 8601 dates, used by both Kubernetes and the IVOA UWS standard. Also add safir.pydantic.normalize_isodatetime to accept only that same format as input to a Pydantic model with a datetime field.

3.5.0 (2023-01-12)

New features

  • Add new helper class safir.gcs.SignedURLService to generate signed URLs to Google Cloud Storage objects using workload identity. To use this class, depend on safir[gcs].

  • Add the safir.testing.gcs module, which can be used to mock the Google Cloud Storage API for testing. To use this module, depend on safir[gcs].

  • Add new helper class safir.pydantic.CamelCaseModel, which is identical to pydantic.BaseModel except with configuration added to accept camel-case keys using the safir.pydantic.to_camel_case alias generator and overrides of dict and json to export in camel-case by default.

3.4.0 (2022-11-29)

New features

  • safir.logging.configure_logging and safir.logging.configure_uvicorn_logging now accept Profile and LogLevel enums in addition to strings for the profile and log_level parameters. The new enums are preferred. Support for enums was added in preparation for changing the FastAPI template to use pydantic.BaseSettings for its Configuration class and validate the values of the profile and log_level configuration parameters.

  • Add new function safir.pydantic.normalize_datetime, which can be used as a Pydantic validator for datetime fields. It also allows seconds since epoch as input, and ensures that the resulting datetime field in the model is timezone-aware and in UTC.

  • Add new function safir.pydantic.to_camel_case, which can be used as a Pydantic alias generator when a model needs to accept attributes with camel-case names.

  • Add new function safir.pydantic.validate_exactly_one_of, which can be used as a model validator to ensure that exactly one of a list of fields was set to a value other than None.

Bug fixes

  • In safir.testing.kubernetes.patch_kubernetes, correctly apply a spec to the mock of kubernetes_asyncio.client.ApiClient. Due to an ordering bug in previous versions, the spec was previously a mock object that didn’t apply any constraints.

3.3.0 (2022-09-15)

New features

  • Add new function safir.logging.configure_uvicorn_logging that routes Uvicorn logging through structlog for consistent formatting. It also adds context to Uvicorn logs in the format recognized by Google’s Cloud Logging.

  • Support the newer project metadata structure and URL naming used by pyproject.toml-only packages in safir.metadata.get_metadata.

3.2.0 (2022-05-13)

New features

  • New support for arq, the Redis-based asyncio distributed queue package. The safir.arq module provides an arq client and metadata/result data classes with a mock implementation for testing. The FastAPI dependency, safir.dependencies.arq.arq_dependency, provides a convenient way to use the arq client from HTTP handlers.

3.1.0 (2022-06-01)

New features

  • Add new FastAPI middleware CaseInsensitiveQueryMiddleware to aid in implementing the IVOA protocol requirement that the keys of query parameters be case-insensitive.

3.0.3 (2022-05-16)

Bug fixes

  • Correctly handle the possibility that request.client is None.

3.0.2 (2022-03-25)

Bug fixes

  • Fix handling of passwords containing special characters such as @ and / in safir.database.

3.0.1 (2022-02-24)

Bug fixes

  • safir.database retains the port in the database URL, if provided.

3.0.0 (2022-02-23)

Backward-incompatible changes

  • XForwardedMiddleware no longer sets forwarded_proto in the request state and instead directly updates the request scope so that subsequent handlers and middleware believe the request scheme is that given by an X-Forwarded-Proto header. This fixes the scheme returned by other request methods and attributes such as url_for in cases where the service is behind an ingress that terminates TLS.

New features

  • Add new FastAPI dependencies auth_dependency and auth_logger_dependency from the safir.dependencies.gafaelfawr module. auth_dependency returns the username of the user authenticated via Gafaelfawr (pulled from the X-Auth-Request-User header. auth_logger_dependency returns the same logger as logger_dependency but with the user parameter bound to the username from auth_dependency.

  • Add utility functions to initialize a database and create a sync or async session. The session creation functions optionally support a health check to ensure the database schema has been initialized.

  • Add new FastAPI dependency db_session_dependency that creates a task-local async SQLAlchemy session.

  • Add utility functions datetime_from_db and datetime_to_db to convert between timezone-naive UTC datetimes stored in a database and timezone-aware UTC datetimes used elsewhere in a program.

  • Add a run_with_async decorator that runs the decorated async function synchronously. This is primarily useful for decorating Click command functions (for a command-line interface) that need to make async calls.

2.4.2 (2022-01-24)

New features

  • Add a very basic and limited implementation of list_namespaced_pod to safir.testing.kubernetes.

2.4.1 (2022-01-14)

Bug fixes

  • In the Kubernetes mock API, change the expected body for patch_namespaced_custom_object_status to reflect the need to send a JSON patch for the entirety of the /status path in order to work around a kubernetes_asyncio bug.

2.4.0 (2022-01-13)

New features

  • Add an initialize_kubernetes helper function to load Kubernetes configuration. Add the safir.testing.kubernetes module, which can be used to mock the Kubernetes API for testing. To use the new Kubernetes support, depend on safir[kubernetes].

2.3.0 (2021-12-13)

New features

  • When logging in JSON format, use severity instead of level for the log level for compatibility with Google Log Explorer.

  • In the FastAPI logger_dependency, add log information about the incoming requst in a format compatible with Google Log Explorer.

2.2.0 (2021-11-09)

Bug fixes

  • Restore previous http_client_dependency behavior by enabling following redirects by default. This adjusts for the change of defaults in HTTPX 0.20.0.

2.1.1 (2021-10-29)

Other changes

  • Require structlog 21.2.0 and adjust logger configuration of exception handling for the expectations of that version.

2.1.0 (2021-08-09)

New features

  • Add safir.models.ErrorModel, which is a model of the error message format preferred by FastAPI. Using the model is not necessary but it’s helpful to reference it in API documentation to generate more complete information about the error messages.

Bug fixes

  • Mark all FastAPI dependencies as async so that FastAPI doesn’t run them in an external thread pool.

2.0.1 (2021-06-24)

New features

  • Defer creation of httpx.AsyncClient until the first time it is requested. Allow re-creation after aclose(), which makes it easier to write test suites.

2.0.0 (2021-06-21)

Backward-incompatible changes

As of this release, Safir is a helper library for FastAPI applications instead of aiohttp applications. Much of the library has changed. Authors of software that uses Safir should read the documentation again as part of the upgrade to FastAPI.

Included in this release is:

  • A FastAPI dependency to provide a structlog logger as configured by the safir.logging package, replacing the aiohttp middleware.

  • A FastAPI dependency to provide a global httpx.AsyncClient, replacing the middleware that provided an aiohttp client.

  • Starlette (FastAPI) middleware to parse X-Forwarded-* headers and update request information accordingly.

  • safir.metadata.get_metadata now returns a Pydantic model.

As of this release, Safir only supports Python 3.8 or later.

1.0.0 (2021-06-18)

Safir v1 will be the major release supporting aiohttp. Starting with Safir v2, Safir will become a support library for FastAPI applications.

This release has no significant difference from 0.2.0. The version bump is to indicate that the aiohttp support is stable.

0.2.0 (2021-06-10)

New features

  • configure_logging now supports an optional add_timestamp parameter (false by default) that adds a timestamp to each log message.

0.1.1 (2021-01-14)

Bug fixes

  • Fix duplicated log output when logging is configured multiple times.

0.1.0 (2020-02-26)

New features

  • The first release of Safir featuring:

    • safir.http for adding an aiohttp.ClientSession to your application.

    • safir.logging for configuring structlog loggers.

    • safir.metadata helps your gather and structure metadata about your application for publishing on metadata endpoints.

    • safir.middleware includes a logging middleware that adds a logger with bound context about the request to your Request object.

    • Documentation about these features and a tutorial for starting a new application with the roundtable_aiohttp_bot template.