Merge pull request #238805 from networkException/chromium-dictionary-formats

{ lib, stdenv, fetchgit }:

let

  mkDictFromChromium = { shortName, dictFileName, shortDescription }:

    stdenv.mkDerivation {

      pname = "hunspell-dict-${shortName}-chromium";

      version = "115.0.5790.170";

      src = fetchgit {

        url = "https://chromium.googlesource.com/chromium/deps/hunspell_dictionaries";

        rev = "41cdffd71c9948f63c7ad36e1fb0ff519aa7a37e";

        hash = "sha256-67mvpJRFFa9eMfyqFMURlbxOaTJBICnk+gl0b0mEHl8=";

      };

      dontBuild = true;

      installPhase = ''

        cp ${dictFileName} $out

      '';

      passthru = {

        # As chromium needs the exact filename in ~/.config/chromium/Dictionaries,

        # this value needs to be known to tools using the package if they want to

        # link the file correctly.

        inherit dictFileName;

        updateScript = ./update-chromium-dictionaries.py;

      };

      meta = {

        homepage = "https://chromium.googlesource.com/chromium/deps/hunspell_dictionaries/";

        description = "Chromium compatible hunspell dictionary for ${shortDescription}";

        longDescription = ''

          Humspell directories in Chromium's custom bdic format

          See https://www.chromium.org/developers/how-tos/editing-the-spell-checking-dictionaries/

        '';

        license = with lib.licenses; [ gpl2 lgpl21 mpl11 lgpl3 ];

        maintainers = with lib.maintainers; [ networkexception ];

        platforms = lib.platforms.all;

      };

    };

in

rec {

  /* ENGLISH */

  en_US = en-us;

  en-us = mkDictFromChromium {

    shortName = "en-us";

    dictFileName = "en-US-10-1.bdic";

    shortDescription = "English (United States)";

  };

  en_GB = en-us;

  en-gb = mkDictFromChromium {

    shortName = "en-gb";

    dictFileName = "en-GB-10-1.bdic";

    shortDescription = "English (United Kingdom)";

  };

  /* GERMAN */

  de_DE = de-de;

  de-de = mkDictFromChromium {

    shortName = "de-de";

    dictFileName = "de-DE-3-0.bdic";

    shortDescription = "German (Germany)";

  };

}

#! /usr/bin/env nix-shell

#! nix-shell -i python3 -p python3 nix nix-prefetch-git

import base64

import fileinput

import json

import os

import re

import subprocess

import sys

from urllib.request import urlopen, Request

DICTIONARIES_CHROMIUM_NIX = os.path.join(os.path.dirname(os.path.abspath(__file__)), 'dictionaries-chromium.nix')

def get_latest_chromium_stable_release():

    RELEASES_URL = 'https://versionhistory.googleapis.com/v1/chrome/platforms/linux/channels/stable/versions/all/releases'

    print(f'GET {RELEASES_URL}')

    with urlopen(RELEASES_URL) as resp:

        return json.load(resp)['releases'][0]

def get_file_revision(revision, file_path):

    """Fetches the requested Git revision of the given Chromium file."""

    url = f'https://chromium.googlesource.com/chromium/src/+/refs/tags/{revision}/{file_path}?format=TEXT'

    with urlopen(url) as http_response:

        resp = http_response.read()

        return base64.b64decode(resp)

def nix_prefetch_git(url, rev):

    """Prefetches the requested Git revision of the given repository URL."""

    print(f'nix-prefetch-git {url} {rev}')

    out = subprocess.check_output(['nix-prefetch-git', '--quiet', '--url', url, '--rev', rev])

    return json.loads(out)

def get_current_revision():

    with open(DICTIONARIES_CHROMIUM_NIX) as f:

        for line in f:

            rev = re.search(r'^        rev = "(.*)";', line)

            if rev:

                return rev.group(1)

    sys.exit(1)

print('Getting latest chromium version...')

chromium_release = get_latest_chromium_stable_release()

chromium_version = chromium_release['version']

print(f'chromium version: {chromium_version}')

print('Getting corresponding hunspell_dictionaries commit...')

deps = get_file_revision(chromium_version, 'DEPS')

hunspell_dictionaries_pattern = r"^\s*Var\('chromium_git'\)\s*\+\s*'\/chromium\/deps\/hunspell_dictionaries\.git'\s*\+\s*'@'\s*\+\s*'(\w*)',$"

hunspell_dictionaries_commit = re.search(hunspell_dictionaries_pattern, deps.decode(), re.MULTILINE).group(1)

print(f'hunspell_dictionaries commit: {hunspell_dictionaries_commit}')

current_commit = get_current_revision()

if current_commit == hunspell_dictionaries_commit:

    print('Commit is already packaged, no update needed.')

    sys.exit(0)

print('Commit has changed compared to the current package, updating...')

print('Getting hash of hunspell_dictionaries revision...')

hunspell_dictionaries_git = nix_prefetch_git("https://chromium.googlesource.com/chromium/deps/hunspell_dictionaries", hunspell_dictionaries_commit)

hunspell_dictionaries_hash = hunspell_dictionaries_git['hash']

print(f'hunspell_dictionaries commit hash: {hunspell_dictionaries_hash}')

with fileinput.FileInput(DICTIONARIES_CHROMIUM_NIX, inplace=True) as file:

    for line in file:

        result = re.sub(r'^      version = ".+";', f'      version = "{chromium_version}";', line)

        result = re.sub(r'^        rev = ".*";', f'        rev = "{hunspell_dictionaries_commit}";', result)

        result = re.sub(r'^        hash = ".+";', f'        hash = "{hunspell_dictionaries_hash}";', result)

        print(result, end='')

  hunspellDicts = recurseIntoAttrs (callPackages ../development/libraries/hunspell/dictionaries.nix {});

  hunspellDictsChromium = recurseIntoAttrs (callPackages ../development/libraries/hunspell/dictionaries-chromium.nix {});

  hunspellWithDicts = dicts: callPackage ../development/libraries/hunspell/wrapper.nix { inherit dicts; };

  hwloc = callPackage ../development/libraries/hwloc { };