Note: Triggers full index rebuild
This release brings a bunch of improvements that we've made in the aftermath of releasing the logic-v2 scoring algorithm. Some little tweaks to how things are scored as well as new configuration options (for example, one to exclude aliases from name matching). For names, in addition to the detailed explanation on how the name was matched, we now include the names of the query and the result entity that were used to make the match. We've also made some performance optimizations to logic-v2 and have made it so that requests can be scored concurrently in yente, keeping response times more even in high-load scenarios.
We've spent some time to make sure that yente works with Amazon OpenSearch Serverless (AOSS), an AWS product that provides a fully managed search index that serves as a backend for yente. Please refer to the documentation for how to configure AOSS as the search provider. That said, we don't have any real-life experience running this configuration in production -— if you have problem reports or success stories running yente with OpenSearch Serverless, please let us know!
This release contains two breaking changes for everyone running yente using the example docker-compose.yml that is shipped with yente:
- First, yente is now configured to use
delivery.opensanctions.orgas a data retrieval location by default, encouraging prospective commercial users to create a delivery token — see our documentation for more information. Setting this up only takes a minute and nothing about our licensing terms is changing. The idea here is not to create a wall around our data, but to steer our commercial customers towards a mechanism that allows us to better support them by collecting telemetry, and to provide preferred service to them in case things go south in a technical or legal way for the public data bucket. Non-commercial users and users exempted from commercial licensing can easily configureyenteto access the public data archive directly. - Further, the
appservice now depends on theindex, meaning thatdocker compose up -d appwill now also bring up an index by default (see c0f5b71).
As usual, this release contains updates to our whole stack, including an upgrade of the cryptography package from 46.0.2 to 46.0.4.
What's Changed
- build(deps): bump aiohttp[speedups] from 3.12.15 to 3.13.0 by @dependabot[bot] in #908
- match: log dataset, threshold, algorithm as passed in request by @leonhandreke in #911
- build(deps): bump fastapi from 0.118.0 to 0.119.0 by @dependabot[bot] in #913
- build(deps): bump aiofiles from 24.1.0 to 25.1.0 by @dependabot[bot] in #910
- build(deps): bump structlog from 25.4.0 to 25.5.0 by @dependabot[bot] in #931
- build(deps): bump uvicorn[standard] from 0.37.0 to 0.38.0 by @dependabot[bot] in #919
- build(deps): bump orjson from 3.11.3 to 3.11.4 by @dependabot[bot] in #928
- Allow concurrency between scoring calls by @pudo in #929
- build(deps): bump fastapi from 0.119.0 to 0.120.1 by @dependabot[bot] in #930
- build(deps): bump aiohttp[speedups] from 3.13.0 to 3.13.1 by @dependabot[bot] in #920
- build(deps): bump cryptography from 46.0.2 to 46.0.3 by @dependabot[bot] in #918
- build(deps): update types-aiofiles requirement from <25.0,>=24.0 to >=24.0,<26.0 by @dependabot[bot] in #914
- deps: Get rid of jellyfish, we don't depend on it (anymore?) by @leonhandreke in #915
- build(deps): bump elasticsearch[async] from 8.19.1 to 8.19.2 by @dependabot[bot] in #933
- build(deps): bump aiohttp[speedups] from 3.13.1 to 3.13.2 by @dependabot[bot] in #932
- build(deps): bump fastapi from 0.120.1 to 0.120.3 by @dependabot[bot] in #938
- Nomenklatura with
nm_name_property& new FtM by @pudo in #934 - build(deps): bump fastapi from 0.120.3 to 0.120.4 by @dependabot[bot] in #939
- build(deps): bump fastapi from 0.120.4 to 0.121.0 by @dependabot[bot] in #941
- Wait for index before starting app in docker compose by @jbothma in #945
- build(deps): bump fastapi from 0.121.0 to 0.121.1 by @dependabot[bot] in #946
- build(deps): bump pyicu from 2.15.3 to 2.16 by @dependabot[bot] in #944
- build(deps): bump rigour from 1.4.0 to 1.4.2 by @dependabot[bot] in #948
- build(deps): bump fastapi from 0.121.1 to 0.121.2 by @dependabot[bot] in #951
- build(deps): bump nomenklatura from 4.1.10 to 4.2.0 by @dependabot[bot] in #949
- build(deps): bump fastapi from 0.121.2 to 0.121.3 by @dependabot[bot] in #954
- opensearch: some fixes to make Amazon OpenSearch Serverless work by @leonhandreke in #937
- build(deps): bump opensearch-py[async] from 3.0.0 to 3.1.0 by @dependabot[bot] in #956
- match: Switch default and best algorithms to logic-v2 by @leonhandreke in #961
- build(deps): bump actions/checkout from 5 to 6 by @dependabot[bot] in #955
- build(deps-dev): bump pytest from 8.4.2 to 9.0.1 by @dependabot[bot] in #950
- Raise an exception if scope not found in defined catalog by @leonhandreke in #923
- doc(index.md):fix double slash in hosted API link by @mainibet in #995
- docs(settings.md): update YENTE_QUERY_CONCURRENCY default to 50 by @mainibet in #996
- build(deps): bump uvicorn[standard] from 0.38.0 to 0.40.0 by @dependabot[bot] in #981
- build(deps): bump anyio from 4.11.0 to 4.12.1 by @dependabot[bot] in #994
- build(deps): bump elasticsearch[async] from 8.19.2 to 8.19.3 by @dependabot[bot] in #983
- build(deps-dev): bump pytest from 9.0.1 to 9.0.2 by @dependabot[bot] in #971
- build(deps): bump aiohttp[speedups] from 3.13.2 to 3.13.3 by @dependabot[bot] in #990
- build(deps): bump python-multipart from 0.0.20 to 0.0.21 by @dependabot[bot] in #977
- FtM stack upgrade by @leonhandreke in #1000
- Modernize pyproject.toml a bit and advise use of uv by @leonhandreke in #1002
- build(deps): bump orjson from 3.11.4 to 3.11.5 by @dependabot[bot] in #970
- build(deps): bump fastapi from 0.121.3 to 0.128.0 by @dependabot[bot] in #986
- Configure yente with delivery.opensanctions.com as default by @leonhandreke in #924
- build(deps): bump python-multipart from 0.0.21 to 0.0.22 by @dependabot[bot] in #1003
- build(deps): bump cryptography from 46.0.3 to 46.0.4 by @dependabot[bot] in #1006
- build(deps): bump rigour from 1.6.1 to 1.6.2 by @dependabot[bot] in #1007
- docs: Add monitoring docs for /healthz, /readyz, /catalog by @leonhandreke in #1005
Full Changelog: v5.0.2...v5.2.0