translations_codegen.py

"""Static code analysis to generate translations.py from blender/addon.py

    - Strings are marked with T() function.
    - Non-translated strings are pushed to the top of the list, so that they can be easily identified and translated.
"""

from translations import translations_tuple
from collections import defaultdict
import os, ast

SOURCES = ["blender/addon.py"]
LANGS = ["zh_HANS"]


def parse_ast(filename):
    with open(filename, "r", encoding="utf-8") as f:
        src = ast.parse(f.read(), filename=filename)
        for pa in ast.walk(src):
            for node in ast.iter_child_nodes(pa):
                if isinstance(node, ast.Call) and getattr(node.func, "id", None) == "T":
                    yield f"scripts/addons/sssekai_blender_io/{filename}:{pa.lineno}", node.args[
                        0
                    ].value


translations = defaultdict(lambda: defaultdict(list))

for (
    (msgctxt, msgid),
    (sources, gen_comments),
    (lang, translation, (is_fuzzy, comments)),
) in translations_tuple:
    translations[msgid][lang].append((msgctxt, sources, translation))

for source in SOURCES:
    for sources, text in parse_ast(source):
        for lang in LANGS:
            has_lang = lang in translations[text]
            has_tl = (
                any((elem[-1] for elem in translations[text][lang]))
                if has_lang
                else False
            )
            if not has_lang or not has_tl:
                print(f'* Missing <{lang}> "{text}"')
            if not has_lang:
                translations[text][lang].append(("*", ((sources,),), ""))

# ((msgctxt, msgid), (sources, gen_comments), (lang, translation, (is_fuzzy, comments)), ...)
translations_tuple = list()
for msgid, translations_dict in translations.items():
    for lang, translations_list in translations_dict.items():
        for msgctxt, sources, translation in translations_list:
            translations_tuple.append(
                ((msgctxt, msgid), (sources, ""), (lang, translation, (False, ())))
            )
translations_tuple = sorted(translations_tuple, key=lambda x: x[0][1])  # key=msgid
translations_tuple = sorted(
    translations_tuple, key=lambda x: bool(x[2][1])
)  # key=translation
# Sorted by msgid, and then translation status
with open("translations.py", "w", encoding="utf-8") as f:
    from pprint import pformat

    f.write(
        """# This file is auto-generated by translations_codegen.py

translations_tuple = """
    )
    f.write(pformat(translations_tuple))
    f.write(
        """

translations_dict = {}
for msg in translations_tuple:
    key = msg[0]
    for lang, trans, (is_fuzzy, comments) in msg[2:]:
        if trans and not is_fuzzy:
            translations_dict.setdefault(lang, {})[key] = trans           
"""
    )