v0.3.19

[v0.3.19] - 2025-10-07

Changed

• pittgoogle/schema.py
  • update _clean_for_json() to return None if math.isinf(value)
  • _name_in_bucket for the LsstSchema class now returns:
    {alert.schema.version}/kafkaPublishTimestamp={_date}/{objectid_key}={alert.objectid}/{sourceid_key}={alert.sourceid}.
    avro
• tests/test_alert.py
  • update test_name_in_bucket to account for the change in pittgoogle/schema.py/_name_in_bucket

v0.3.18

[v0.3.18] - 2025-09-16

Changed

• pittgoogle/alert.py
  • Address a bug in the get and get_key functions that raises an AttributeError when a field's value is None

v0.3.17

[v0.3.17] - 2025-09-01

Added

• schemas/lsst/8/0/
  • lsst.v8_0.alert.avsc
  • lsst.v8_0.diaForcedSources.avsc
  • lsst.v8_0.diaSource.avsc
  • lsst.v8_0.diaObject.avsc
  • lsst.v8_0.ssSource.avsc
  • lsst.v8_0.MPCORB.avsc
• schemas/lsst/9/0/
  • lsst.v9_0.alert.avsc
  • lsst.v9_0.diaForcedSources.avsc
  • lsst.v9_0.diaSource.avsc
  • lsst.v9_0.diaObject.avsc
  • lsst.v9_0.ssSource.avsc
  • lsst.v9_0.MPCORB.avsc
• sample_alerts/lsst/
  • lsst.v8_0.avro
  • lsst.v9_0.avro

Changed

• schemas/maps/lsst.yml
  • updated field name: ssObject -> ssSource
  • the mapping for the field objectid is now a dict[list] type
• pittgoogle/schema.py
  • added v8_0 and v9_0 to the list of supported schema versions for LSST alerts
  • the _name_in_bucket function uses the ssObjectId instead of the diaObjectId when the value of
    diaObjectId is null (each diaSource will be associated with either a diaObject OR ssObject)
  • _clean_for_json converts datetime.datetime objects to timestamp
• tests/test_alert.py
  • unit test for name_in_bucket has been updated to reflect the new pathname
• tests/test_schema.py
  • updated the JSON serialization test in test_sample_alertsto handle type mismatches
• tests/load_data.py
  • updated field names to accomodate the latest LSST alert schema version (v9_0)
• pittgoogle/alert.py
  • The get and get_key functions support fields that are list[dict] types
  • healpix9, healpix19, and healpix29 explicitly return int types

Removed

• alertid and ssobjectid from pittgoogle/alert.py, tests/test_alert.py, and all schema maps

v0.3.16

[v0.3.16] - 2025-08-18

Added

• Data listings for LSST
• LvkSchema properties _from_yaml and _name_in_bucket.
• "objectid" added to LVK schema map. Both "objectid" and "sourceid" now point to the
  LVK field "superevent_id" since there is no distinction between an LVK "object" and "source".
• Added a unit test for LvkSchema._name_in_bucket. Added a "schema_version" field to the test data in
  tests/data/sample_alerts/lvk/ to mimic our broker's ps_to_storage module (required for _name_in_bucket).

v0.3.15

[v0.3.15] - 2025-05-12

Changed

• Alert:
  • Updated the separator "." -> "_" used to define attribute key names.

v0.3.14

[v0.3.14] - 2025-04-28

Added

• Alert:
  • Added the following fields to Alert.attributes: n_previous_sources and ssobjectid.
• Unit tests:
  • Tests take into account new Alert.attributes keys.

Changed

• Rename Alert method _add_id_attributes() -> _add_attributes() to reflect that more properties are added.

Fixed

• Topic.publish(): Message attributes were improperly defined

v0.3.13

[v0.3.13] - 2025-04-04

Added

• Alert:
  • Properties: ra, dec, healpix9, healpix19, healpix29, name_in_bucket.
  • Add healpix properties to Alert.attributes.
• Schema:
  • Properties: _name_in_bucket, serializer, and deserializer. The last two can be used to
    switch between JSON and Avro on the fly.
  • Methods to support the LSST schema consolidation.
  • Schema map for the default schema.
• Topic.publish(): new keyword arg drop_cutouts.
• Dedicated classes for all schemas, including DefaultSchema, ElasticcSchema, LsstSchema,
  LvkSchema, and ZtfSchema. These are subclasses of pittgoogle.Schema.
• schema.Serializers class to hold all serializers and deserializers used by the schemas.
• Unit tests:
  • Tests for new Alert properties.
  • Tests for Schema and Serializers.
  • Randomly generated data for schema "lsst.v7_4.alert".
• Schema maps
  • LSST support for ssObjectId

Changed

• Alert:
  • from_* methods now call Schema._init_from_msg so the schema version can be retrieved from the message.
  • Schema.version is now added to the attributes automatically.
• Schema:
  • Consolidate LSST schemas into one. Name is "lsst".
  • Change default schema name "default_schema" -> "default".
  • serialize and deserialize methods moved to schema.Serializers.
  • JSON serializer now accepts dicts containing bytes values.
• Topic.publish() now handles the duties of Alert._prep_for_publish() (which is removed).
• Documentation updated with new instructions for developers to implement support for a new schema.
• Unit tests:
  • Move conftest.SampleAlert -> load_data.TestAlert
  • Split TestAlert -> TestAlertFrom, TestAlertProperties, and TestAlertMethods

Fixed

• Fix issue #76. Make Alert.dataframe succeed even when there are no previous sources.

Removed

• Alert._prep_for_publish(). Equivalent functionality was added to Topic.publish().
• SchemaHelpers class removed. These have been promoted to full Schema classes.
• Schema._init_from_msg. This functionality was moved to the Schema._from_yaml class methods.

v0.3.12

Added

• Data listings and references for LSST
• Support for LSST schema versions 7.2, 7.3, and 7.4
• README.md describing where the new LSST alert schemas were obtained.
• Add 'package_path' as a package-level variable.
• Add test data for LSST, LVK, and ZTF.
• Add unit tests for:
  • registry.Schemas
  • alert.Alert
  • utils.Cast
• Add Alert properties and methods:
  • skymap
  • drop_cutouts
  • _prep_for_publish
  • _str_to_datetime
• Add class alert.MockInput with support for Cloud Functions
• Add dependencies hpgeom and google-cloud-functions.

Changed

• Update data listings for ZTF and LVK
• schema.py now specifies all schema versions that are available for LSST
• Schema mappings for new LSST alert versions incorporated into schemas.yml
• Implement schema._ConfluentWireAvroSchema.serialize.
• Add IDs to Alert.attributes the first time it is accessed.
• Dependencies: Lazy load astropy.

Removed

• Remove utils.Cast._strip_cutouts_ztf (moved to Alert)

v0.3.11

[v0.3.11] - 2024-07-22

Added

• Add 'Surveys' documentation section with pages for ZTF and LVK.
• Add LVK to Data Listings page.

Changed

• Reorganize and update data listings.
• Add FAQ content.
• Clean up docs. Remove 'TODO's. Add autosummary to module pages.
• Update Google pricing on FAQ Costs.

Prep v0.3.10 release (#61)

[v0.3.10] - 2024-07-22

Added

• Add bigquery.Client class.
• Add class attributes Table.query, Table.schema.
• Add dependency on 'db-dtypes' to support BigQuery -> Pandas.

Changed

• Remove Table.auth and simplify Table.client. This functionality is now managed by
  bigquery.Client.
• In Table and Topic, the project ID is no longer changed away from what the user provided.
  It was more confusing and dangerous than it was helpful.

Removed

• Table.auth