pkgs/applications/graphics/tesseract/fetch-language-hashes

   1 #!/usr/bin/env bash
   2
   3 # Usage:
   4 #   ./fetch-language-hashes <tessdataRev> [<language code>…]
   5 #
   6 #   Fetches all languages if no language codes are given.
   7 #
   8 # Example:
   9 #   ./fetch-language-hashes 4.0.0 eng spa
  10 #
  11 #   Output:
  12 #     eng = "0iy0...";
  13 #     spa = "15kw...";
  14
  15 set -e
  16
  17 (( $# >= 1 )) || exit 1
  18 tessdataRev=$1
  19 shift
  20
  21 if (( $# > 0 )); then
  22     langCodes="$@"
  23 else
  24     repoPage=$(curl -fs https://github.com/tesseract-ocr/tessdata/tree/$tessdataRev || {
  25                    >&2 echo "Invalid tessdataRev: $tessdataRev"
  26                    exit 1
  27                })
  28     langCodes=$(echo $(echo "$repoPage" | grep -ohP "(?<=/)[^/ ]+?(?=\.traineddata)" | sort -u))
  29 fi
  30
  31 for lang in $langCodes; do
  32     url=https://github.com/tesseract-ocr/tessdata/raw/$tessdataRev/$lang.traineddata
  33     hash=$(nix hash to-sri --type sha256 $(nix-prefetch-url $url 2>/dev/null))
  34     echo "$lang = \"$hash\";"
  35 done