Skip to content

HR: thirteen ARKOD archives beside the rolling file - #241

Merged
ivorbosloper merged 5 commits into
mainfrom
split/hr
Sep 15, 2026
Merged

ivorbosloper merged 5 commits into
mainfrom
split/hr

Conversation

@ivorbosloper

@ivorbosloper ivorbosloper commented Sep 11, 2026

Copy link
Copy Markdown
Collaborator

HR: thirteen ARKOD archives beside the rolling file- #241

Editions 2011-2023 beside the rolling file, 18.7M field-years.

Three things had to be right before the archives could be published at all, and
none of them was visible in the output.

The converter declared both `sources` and `variants`. `sources` wins in the base
converter, so every `--variant` converted the current file: thirteen copies of
2024 would have been published as thirteen years.

`index_as_id` replaced ARKOD's own parcel id with a row number. The archives
carry that id, unique within an edition; only the rolling current file lacks
one, so the row index is the fallback rather than the rule.

The archives set the GeoPackage srs_id to 0 ("Undefined geographic SRS") while
holding EPSG:3765 metres, which put projected coordinates in the STAC extent
and built the Hilbert grid from a global geographic extent, leaving every parcel
in one cell of it.

`missing_schemas` also required six properties the archives do not all carry:
2011 leaves mines_status null for 909k of its 1.29M rows.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DVx9uQV2QPM8ecPAY3ZjXG
Comment thread fiboa_cli/datasets/hr.py Outdated
Comment thread fiboa_cli/datasets/hr.py
@ivorbosloper
ivorbosloper requested a review from m-mohr September 14, 2026 10:43
Comment thread fiboa_cli/datasets/hr.py Outdated
to_epsg() returns None for a CRS that has no EPSG code, not for an invalid one:
a custom projection is valid, projected, and answers None. Overriding it with
EPSG:3765 would reinterpret real coordinates as Croatian TM.

The case this guards against is an archive shipping no .prj, which is crs is
None. A CRS that is declared is trusted; pyproj reprojects from WKT without an
EPSG code just as well.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DVx9uQV2QPM8ecPAY3ZjXG
@ivorbosloper
ivorbosloper requested a review from m-mohr September 15, 2026 07:34
@ivorbosloper
ivorbosloper merged commit acab2f5 into main Sep 15, 2026
7 checks passed
@ivorbosloper
ivorbosloper deleted the split/hr branch September 15, 2026 12:05
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants