From f1b7e3005fa013d74bdbd3c0a19c163fd9dcfce2 Mon Sep 17 00:00:00 2001
From: Alexandre Aubin <alex.aubin@mailoo.org>
Date: Fri, 3 Sep 2021 16:35:03 +0200
Subject: [PATCH] ci: add actions to autoreformat locale files

---
 .github/workflows/i18n.yml        | 28 ++++++++++++++++++
 test/autofix_locale_format.py     | 47 +++++++++++++++++++++++++++++++
 test/reformat_locales.py          | 38 +++++++++++++++++++++++++
 test/remove_stale_i18n_strings.py |  2 +-
 4 files changed, 114 insertions(+), 1 deletion(-)
 create mode 100644 .github/workflows/i18n.yml
 create mode 100644 test/autofix_locale_format.py
 create mode 100644 test/reformat_locales.py

diff --git a/.github/workflows/i18n.yml b/.github/workflows/i18n.yml
new file mode 100644
index 00000000..d92b3f57
--- /dev/null
+++ b/.github/workflows/i18n.yml
@@ -0,0 +1,28 @@
+name: Check / autofix locales
+on:
+  push:
+      branches:
+          - dev
+jobs:
+  i18n:
+    name: Check / auto apply black
+    runs-on: ubuntu-latest
+    steps:
+        - uses: actions/checkout@v2
+        - name: Apply reformating scripts
+          id: action_reformat
+          run: |
+              python3 test/remove_stale_i18n_strings.py
+              python3 test/autofix_locale_format.py
+              python3 test/reformat_locales.py
+              git diff -w --exit-cide
+          continue-on-error: true
+      - name: Create Pull Request
+        if: ${{ failure() }}
+        uses: peter-evans/create-pull-request@v3
+        with:
+          token: ${{ secrets.GITHUB_TOKEN }}
+          title: "Reformat locale files"
+          commit-message: ":robot: Reformat locale files"
+          base: ${{ github.head_ref }}
+          branch: actions/i18n
diff --git a/test/autofix_locale_format.py b/test/autofix_locale_format.py
new file mode 100644
index 00000000..f777f06f
--- /dev/null
+++ b/test/autofix_locale_format.py
@@ -0,0 +1,47 @@
+import re
+import json
+import glob
+
+# List all locale files (except en.json being the ref)
+locale_folder = "locales/"
+locale_files = glob.glob(locale_folder + "*.json")
+locale_files = [filename.split("/")[-1] for filename in locale_files]
+locale_files.remove("en.json")
+
+reference = json.loads(open(locale_folder + "en.json").read())
+
+
+def fix_locale(locale_file):
+
+    this_locale = json.loads(open(locale_folder + locale_file).read())
+    fixed_stuff = False
+
+    # We iterate over all keys/string in en.json
+    for key, string in reference.items():
+
+        # Ignore check if there's no translation yet for this key
+        if key not in this_locale:
+            continue
+
+        # Then we check that every "{stuff}" (for python's .format())
+        # should also be in the translated string, otherwise the .format
+        # will trigger an exception!
+        subkeys_in_ref = [k[0] for k in re.findall(r"{(\w+)(:\w)?}", string)]
+        subkeys_in_this_locale = [k[0] for k in re.findall(r"{(\w+)(:\w)?}", this_locale[key])]
+
+        if set(subkeys_in_ref) != set(subkeys_in_this_locale) and (len(subkeys_in_ref) == len(subkeys_in_this_locale)):
+            for i, subkey in enumerate(subkeys_in_ref):
+                this_locale[key] = this_locale[key].replace('{%s}' % subkeys_in_this_locale[i], '{%s}' % subkey)
+                fixed_stuff = True
+
+    if fixed_stuff:
+        json.dump(
+            this_locale,
+            open(locale_folder + locale_file, "w"),
+            indent=4,
+            ensure_ascii=False,
+        )
+
+
+for locale_file in locale_files:
+    fix_locale(locale_file)
diff --git a/test/reformat_locales.py b/test/reformat_locales.py
new file mode 100644
index 00000000..90251d04
--- /dev/null
+++ b/test/reformat_locales.py
@@ -0,0 +1,38 @@
+import re
+
+def reformat(lang, transformations):
+
+    locale = open(f"locales/{lang}.json").read()
+    for pattern, replace in transformations.items():
+        locale = re.compile(pattern).sub(replace, locale)
+
+    open(f"locales/{lang}.json", "w").write(locale)
+
+######################################################
+
+godamn_spaces_of_hell = ["\u00a0", "\u2000", "\u2001", "\u2002", "\u2003", "\u2004", "\u2005", "\u2006", "\u2007", "\u2008", "\u2009", "\u200A", "\u202f", "\u202F", "\u3000"]
+
+transformations = {s: " " for s in godamn_spaces_of_hell}
+transformations.update({
+    "…": "...",
+})
+
+
+reformat("en", transformations)
+
+######################################################
+
+transformations.update({
+    "courriel": "email",
+    "e-mail": "email",
+    "Courriel": "Email",
+    "E-mail": "Email",
+    "« ": "'",
+    "«": "'",
+    " »": "'",
+    "»": "'",
+    "’": "'",
+    #r"$(\w{1,2})'|( \w{1,2})'": r"\1\2’",
+})
+
+reformat("fr", transformations)
diff --git a/test/remove_stale_i18n_strings.py b/test/remove_stale_i18n_strings.py
index 48f2180e..d3a4fe2a 100644
--- a/test/remove_stale_i18n_strings.py
+++ b/test/remove_stale_i18n_strings.py
@@ -2,7 +2,7 @@ import json
 import glob
 from collections import OrderedDict
 
-locale_folder = "../locales/"
+locale_folder = "locales/"
 locale_files = glob.glob(locale_folder + "*.json")
 locale_files = [filename.split("/")[-1] for filename in locale_files]
 locale_files.remove("en.json")