Skip to content

Releases: Joe-Wait/onspy

onspy v0.2.3

05 Apr 23:13

Choose a tag to compare

Changed

  • Parquet sync now skips noncompliant datasets (instead of failing full runs) when source artifacts are missing or unusable.
  • Sync outputs now include skipped_details with per-dataset reason and skip type (resume or noncompliant).

Added

  • Test coverage for noncompliant skip behavior in parquet sync.

Why this release

This keeps large full-sync workflows stable while still surfacing exactly which upstream datasets are noncompliant.

See CHANGELOG.md for full details.

onspy v0.2.2

05 Apr 22:46

Choose a tag to compare

Fixed

  • Resolve intermittent .tmp -> .parquet rename failures in streaming sync by using unique temp files.
  • Fix mixed-type Arrow conversion failures on large ASHE tables (e.g. soc column) by normalizing chunk dtype during stream writes.
  • Remove repeated dtype warnings in streamed large-dataset sync path.

Added

  • Output-directory sync lock (.onspy_parquet_sync.lock) to prevent concurrent sync jobs from clobbering the same target directory.
  • Test coverage for lock conflict and stale lock recovery.

See CHANGELOG.md for full details.

onspy v0.2.1

05 Apr 22:31

Choose a tag to compare

Fixed

  • Prevent OOM kills during parquet sync on very large datasets by streaming large CSV-backed tables directly to parquet.
  • Preserve true HTTP streaming by avoiding eager response body reads when stream=True.

Changed

  • Harden CSV ingestion/retry behavior for transient parse and HTTP failures.
  • Add regression and parquet sync tests for the large-dataset path.

See CHANGELOG.md for full details.

v0.2.0

02 Apr 23:39

Choose a tag to compare

Highlights

  • Refactored onspy around a unified core.py API used consistently by MCP, CLI, and Python usage.
  • Added MCP-first parquet workflow with download_all_parquet and download_datasets_parquet (repeatable --dataset-id in CLI).
  • Added deterministic observation behavior: wildcard * works for table-backed datasets, while API-only datasets require explicit filters.
  • Added get_dimension_options_detailed and curated boundary helpers (list_boundaries, download_boundary) for map-ready analysis.
  • Updated docs, SKILL guidance, CI/release checks, and test suite alignment for 0.2.0 publish readiness.

v0.1.3

16 Feb 17:48

Choose a tag to compare

[0.1.3] - 2026-02-16

Fixed

  • change ping URL to ons.gov.uk to avoid whitelisting issues in secure environments

Full Changelog: v0.1.2...v0.1.3

v0.1.2

10 Apr 21:47

Choose a tag to compare

[0.1.2] - 2025-04-10

Fixed

  • fix ons_get_latest function name

Full Changelog: v0.1.1...v0.1.2

v0.1.1

30 Mar 11:23

Choose a tag to compare

[0.1.1] - 2025-03-30

Added

  • ons_get_latest function
  • Wellbeing example
  • Update readme

Full Changelog: v0.1.0...v0.1.1

v0.1.0

26 Mar 00:33

Choose a tag to compare

[0.1.0] - 2025-03-25

Added

  • Initial release of onspy package
  • Functions to access and interact with ONS datasets
  • Functions to search and browse code lists
  • Support for downloading CSV data via the ONS API
  • Documentation and examples
  • Test suite with unit and integration tests