Skip to content

Latest commit

 

History

History
139 lines (103 loc) · 4.48 KB

CHANGELOG.rst

File metadata and controls

139 lines (103 loc) · 4.48 KB

Change Log

1.1.0

  • Support a third butler registry, and make the maximum number of registries easier to change.

1.0.0

  • Add seq_num column to the message table.

0.9.7

  • add_message: require the exposure to be in a registry (already be ingested), to prevent adding messages with an invalid (obs_id, instrument). Document the is_new argument as deprecated and ignored.
  • find_exposures: apply ordering before offset & limit.
  • Fix some outdated help.
  • Modernize dependencies.

0.9.6

  • Install butler with the [postgres] option to make it compatible with a Postgres back end.

0.9.5

  • pyproject.toml: add PGUSER and PGPASSWORD to passenv.
  • Update input and computed requirements.
  • Apply dependabot alerts to .github/workflows/ci.yaml.

0.9.4

  • find_exposures:

    • Fix broken order_by handling (and possibly broken limit handling) and improve validation of the order_by order parameter.
    • Allow timespan_begin and timespan_end to be invalid (in which case they are set to None). Exposure ought to always have a valid timespan, but we have some that do not.
  • find_messages: improve validation of the order_by query parameter.

  • Update to python 3.10.

  • Get daf_butler from pypi (as lsst-daf-butler) instead of github.

  • Modernize type annotations, applying changes that required Python 3.9 or 3.10. Use native types or collections.abc where possible. Replace typing.Union and typing.Optional with |. Remove from __future__ import annotations where possible.

0.9.3

  • Improve alembic migration to handle the case that the message table does not exist.
  • Add tests/test_alembic_migration.py.
  • LogMessageDatabase: add message_table constructor argument to improve encapsulation.
  • setup.cfg: specify asyncio_mode = auto.

0.9.2

  • Dockerfile was not copying the test repos.

0.9.1

  • Dockerfile: switch to a simpler base image, as per current SQuaRE recommendations.
  • Add scripts/start-api.sh to run schema evolution and start the service.

0.9.0

  • Add support for schema evolution using alembic.
  • Add "level" and "urls" columns to the message table.

0.8.0

  • Add a "tags" field to messages. Tags must be at least two letters long, contain only ASCII letters. digits, and _ (underscore), and start with a letter. Tags are transformed to lowercase.
  • find_exposures: add registry, offset, and order_by parameters. It is no longer possible to search both registries at the same time, because that does not work well with order_by and offset.
  • Add get_instruments to show which instruments each butler registry supports.
  • Work around a bug in the butler that made the service fail when run with two registries that contained data for different cameras.
  • Replace the old test registry with two registries, each with data for a different camera.
  • Include the raw images used to generate the registries (in a highly compressed form, with 0 for all pixel values) in order to simplify creating new versions of the registries.
  • Add a hack to work around test failures on github: open Butlers with writeable=True when running tests. Undo this hack using DM-33642 once it's safe to do so.

0.7.1

  • Fix a bug in tests/test_find_messages.py: it could search for "", which is not a valid search.
  • Improve random seeding in unit tests.
  • Use description instead of title for the metadata for returned objects. This improves the appearance of the interactive UI (especially redoc).

0.7.0

Modernize the code:

  • Update the specified and generated requirements.
  • Use the main branch of daf_butler, instead of master.
  • Fix some type annotations.
  • For endpoints with an optional final "/" provide two endpoints: with and without the final "/". This fixes a new source of unit test breakage: uvicorn unexpectedly redirects. Hide the version with a final "/" from the API docs. Expand the unit tests to test both versions of each of these endpoints.
  • Fix a new sqlachemy warning about empty and_ clauses.

0.6.0

  • Support unicode in message text and enhance the tests to check it.
  • delete messages/id: fix the declared status code. (It didn't cause problems, but it was confusing.)

0.5.1

  • Fix return value from deleting messages.
  • Update generated requirements.

0.5.0

  • Allow searching for exposures.

  • Enhance message search:

    • Make the "is_x" arguments tri-state, to expand the search options.
    • Document how to send array parameters.