[Fontations] Add fetch script for retrieving Google Fonts test data
Fetch an archive of per-language test strings, and a mapping of
language coverage to font names and files including the font files
themselves. This archive is produced by an internal tool [1]
and consists of only open-source SIL Open Font License licensed
fonts, currently only from the Noto set.
Name a revision in DEPS and add a fetch-fonts-testdata script
for local development usage. In testing on infra, other means for
fetching the testdata CIPD archive are expected to be used, such
as a bazel-driven download.
Preparation for an upcoming CL which uses the test data to iterate
over a set of languages, retrieves the language-specific test
strings and performs path equality and rasterisation tests.
[1] cl/559229536
Bug: skia:295807418
Change-Id: I4fc002deec60c03dfe1ea708cc2656df0e584878
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/752837
Commit-Queue: Dominik Röttsches <drott@google.com>
Reviewed-by: Kevin Lubick <kjlubick@google.com>
diff --git a/DEPS b/DEPS
index 0bb5945..df74073 100644
--- a/DEPS
+++ b/DEPS
@@ -9,6 +9,10 @@
# ninja CIPD package version.
# https://chrome-infra-packages.appspot.com/p/infra/3pp/tools/ninja
'ninja_version': 'version:2@1.8.2.chromium.3',
+
+ # googlefonts_testdata CIPD package version
+ # https://chrome-infra-packages.appspot.com/p/chromium/third_party/googlefonts_testdata/
+ 'googlefonts_testdata_version': 'version:20230409',
}
# If you modify this file, you will need to regenerate the Bazel version of this file (bazel/deps.bzl).
diff --git a/bin/fetch-fonts-testdata b/bin/fetch-fonts-testdata
new file mode 100755
index 0000000..7fc0f5f
--- /dev/null
+++ b/bin/fetch-fonts-testdata
@@ -0,0 +1,81 @@
+#!/usr/bin/env python3
+
+# Copyright 2023 Google Inc.
+#
+# Use of this source code is governed by a BSD-style license that can be
+# found in the LICENSE file.
+
+import hashlib
+import json
+import os
+import platform
+import re
+import stat
+import sys
+import tempfile
+import zipfile
+
+from urllib.request import urlopen
+
+TESTDATA_REL_DIR = 'third_party/externals/googlefonts_testdata'
+ZIP_ARCHIVE_NAME = 'googlefonts_testdata.zip'
+VERSION_FILE = 'googlefonts_testdata.version'
+
+def sha256sum(path):
+ try:
+ with open(path, 'rb') as f:
+ return hashlib.sha256(f.read()).hexdigest()
+ except OSError:
+ return ''
+
+
+os.chdir(os.path.join(os.path.dirname(__file__), os.pardir))
+
+with open('DEPS', 'rb') as f:
+ deps = f.read().decode()
+found = re.findall(r"'googlefonts_testdata_version':\s*'(\S+)'", deps)
+if len(found) != 1:
+ print('Unable to find desired revision in DEPS', file=sys.stderr)
+ exit(1)
+desired_version = found[0]
+
+# Determine which version (if any) we currently have.
+zip_archive_path = os.path.join(TESTDATA_REL_DIR, ZIP_ARCHIVE_NAME)
+current_sha256 = sha256sum(zip_archive_path)
+archive_version_path = os.path.join(TESTDATA_REL_DIR, VERSION_FILE)
+
+current_version = {
+ 'version': '',
+ 'sha256': '',
+}
+try:
+ with open(archive_version_path, 'r', encoding='utf8') as f:
+ current_version = json.load(f)
+except OSError:
+ pass
+
+if desired_version != current_version['version']:
+ print('Version "%s" requested by DEPS differs from current version "%s"' % (
+ desired_version, current_version['version']))
+elif current_sha256 != current_version['sha256']:
+ print('sha256 sum "%s" does not match last-downloaded version "%s"' % (
+ current_sha256, current_version['sha256']))
+else:
+ print('Already up to date.')
+ exit(0)
+
+print('Fetching %s at %s.' % (ZIP_ARCHIVE_NAME, desired_version))
+
+os.makedirs(os.path.dirname(zip_archive_path), exist_ok=True)
+with open(zip_archive_path, 'wb') as f:
+ url = f'https://chrome-infra-packages.appspot.com/dl/chromium/third_party/googlefonts_testdata/+/{desired_version}'
+ f.write(urlopen(url).read())
+
+with zipfile.ZipFile(zip_archive_path, 'r') as f:
+ f.extractall(TESTDATA_REL_DIR)
+
+# Write the downloaded version info to a file.
+current_version['version'] = desired_version
+current_version['sha256'] = sha256sum(zip_archive_path)
+with open(archive_version_path, 'w', encoding='utf8') as f:
+ json.dump(current_version, f, sort_keys=True, indent=2)
