opendatabs/data.bs.ch_llms.txt
Watch
2
mirror of https://github.com/opendatabs/data.bs.ch_llms.txt.git synced 2026-08-26 12:24:24 +00:00
Code Releases Activity
Read-only mirror of https://github.com/opendatabs/data.bs.ch_llms.txt — Kanton Basel-Stadt. Issues & pull requests at the source.
3 commits 3 branches 0 tags 118 KiB
  • Python 100%
Find a file
Exact
Repository files (latest commit first)
Filename Latest commit message Latest commit date
Orhan Saeedi 16bc1b4b43 Update dependencies in pyproject.toml, revise README for API usage, and regenerate dataset documentation with updated timestamps and record counts.
2026-03-12 16:26:49 +01:00
.github/workflows Intial commit 2026-03-12 15:16:13 +01:00
llms Update dependencies in pyproject.toml, revise README for API usage, and regenerate dataset documentation with updated timestamps and record counts. 2026-03-12 16:26:49 +01:00
scripts Update dependencies in pyproject.toml, revise README for API usage, and regenerate dataset documentation with updated timestamps and record counts. 2026-03-12 16:26:49 +01:00
.python-version Intial commit 2026-03-12 15:16:13 +01:00
LICENSE Create LICENSE 2026-03-12 15:17:13 +01:00
llms.txt Intial commit 2026-03-12 15:16:13 +01:00
pyproject.toml Update dependencies in pyproject.toml, revise README for API usage, and regenerate dataset documentation with updated timestamps and record counts. 2026-03-12 16:26:49 +01:00
README.md Update dependencies in pyproject.toml, revise README for API usage, and regenerate dataset documentation with updated timestamps and record counts. 2026-03-12 16:26:49 +01:00
uv.lock Update dependencies in pyproject.toml, revise README for API usage, and regenerate dataset documentation with updated timestamps and record counts. 2026-03-12 16:26:49 +01:00

README.md

data.bs.ch llms.txt toolkit

Maintainer toolkit for generating and validating LLM-facing docs for data.bs.ch.

What this repository does

  • Publishes a concise root llms.txt router.
  • Generates dataset indexes for LLM navigation:
    • llms/datasets/index.md
    • llms/datasets/by-theme/index.md
    • llms/datasets/by-theme/*.md
  • Validates required structure and links before publishing.

Architecture

  • llms.txt: root routing manifest for LLMs.
  • llms/*.md: hand-authored guides.
  • scripts/generate_dataset_docs.py: builds dataset index pages.
  • scripts/validate_llms_docs.py: checks required files/sections and link integrity.
  • .github/workflows/validate-llms-docs.yml: CI validation workflow.
  • .github/workflows/refresh-llms-datasets.yml: scheduled generation + PR workflow.

Data source strategy

scripts/generate_dataset_docs.py uses the public Explore API only:

  • https://data.bs.ch/api/explore/v2.1/catalog/datasets

Only public datasets are indexed.

Prerequisites

  • uv installed
  • Python 3.12+ (see .python-version)

Setup (uv)

uv sync

Common commands (uv)

Generate dataset docs:

uv run python scripts/generate_dataset_docs.py

Validate docs:

uv run python scripts/validate_llms_docs.py

CI behavior

  • Validation workflow (validate-llms-docs.yml):
    • uv sync --frozen
    • generate docs
    • validate docs and links
  • Refresh workflow (refresh-llms-datasets.yml):
    • scheduled daily
    • regenerates dataset docs
    • opens a PR with generated changes

Publish/deploy checklist

  1. Ensure root file is reachable at /llms.txt.
  2. Ensure linked docs are reachable at:
    • /llms/getting-started.md
    • /llms/odsql-cheatsheet.md
    • /llms/query-cookbook.md
    • /llms/datasets/index.md
  3. Run local generate + validate commands.
  4. Confirm CI passes.
  5. Spot-check a few generated theme pages and dataset rows.

Troubleshooting

  • Missing required root links in validator
    • Ensure llms.txt includes absolute URLs for all critical docs.
  • Broken local links
    • Run validator and fix target paths in markdown.
  • Dataset count changes unexpectedly
    • Verify portal availability and Explore API connectivity.