lychee.toml

################################################################################
# Config for the link checker lychee.
#
# Download & learn more at:
# https://github.com/lycheeverse/lychee
#
# Example config:
# https://github.com/lycheeverse/lychee/blob/master/lychee.example.toml
#
# Run `lychee . --dump` to list all found links that are being checked.
################################################################################

# Maximum number of concurrent link checks.
# Workaround for "too many open files" error on MacOS, see https://github.com/lycheeverse/lychee/issues/1248
max_concurrency = 32


# URLs to check (supports regex). Has preference over all excludes.
exclude_path = [
  # Unfortunately lychee doesn't yet read .gitignore https://github.com/lycheeverse/lychee/issues/1331
  # The following entries are there because of that:
  "__pycache__",
  "_deps/",
  ".pixi",
  "build",
  "docs/python/",
  "rerun_cpp/docs/html",
  "rerun_cpp/docs/xml",
  "rerun_py/site/",
  "target_ra",
  "target_wasm",
  "target",
  "rerun_cpp/tmp/",

  # Actually ignored files beyond .gitignore
  "scripts/screenshot_compare/assets/templates/",
]

# Exclude URLs and mail addresses from checking (supports regex).
exclude = [
  '.*?speculative-link',                                      # Skip speculative links
  '.github/workflows/.*.yml',                                 # GitHub action workflows cause issues on CI.
  '\$relpath\^',                                              # Relative paths as used by rerun_cpp's doc header.
  '%7B%7B.+%7D%7D',                                           # Ignore things that look like links because of `href` but are actually replacement strings. The URL escaping is for '{{.+}}' (this seems to be needed for html embedded urls since lychee assumes they use this encoding).
  'https://crates.io/crates/.*',                              # Avoid crates.io rate-limiting
  'https://cs.nyu.edu/~silberman/datasets/nyu_depth_v2.html', # Nyud links are down every now and then.
  'https://eigen.tuxfamily.org/',                             # Website down https://gitlab.com/libeigen/eigen/-/issues/2804
  'https://fonts.googleapis.com/',                            # Font API entrypoint, not a link.
  'https://fonts.gstatic.com/',                               # Font API entrypoint, not a link.
  'https://github.com/rerun-io/rerun/pull/\.*',               # Ignore links to our own pull requests (typically in changelog).
  'https://github.com/rerun-io/rerun/commit/\.*',             # Ignore links to our own commits (typically in changelog).
  'https://stackoverflow.com/.',                              # Stackoverflow links are no longer accessible from CI.
  'https://www.tensorflow.org/',                              # tensorflow.org apparently blocks CI.

  # Works but is really slow at times:
  'https://openaccess.thecvf.com/content/CVPR2023/html/Du_Learning_To_Render_Novel_Views_From_Wide-Baseline_Stereo_Pairs_CVPR_2023_paper.html',
  'https://anaconda.org/conda-forge/arrow-cpp',

  #'^file:///',                                  # Ignore local file links. They need to be tested, but it's useful for external links we have to ping.
]