refactor: change project layout

- move script as asset because they are not build script
- create new script directory for "User Compile" method because github action need to build with gtest.
- change compile manual for this change.
- modify external dependency location in github action and gitignore.

asset/pycodec/conv_encoding_table.py

@@ -0,0 +1,54 @@
+import csv
+from pathlib import Path
+import jinja2
+
+
+class LanguageToken:
+    name: str
+    alias: tuple[str, ...]
+    code_page: str | None
+    iconv_code: str | None
+
+    def __init__(self, row: list[str]):
+        """Init language token from CSV row."""
+        self.name = row[0].lower()
+        code_page = row[1]
+        self.code_page = None if len(code_page) == 0 else code_page
+        iconv_code = row[2]
+        self.iconv_code = None if len(iconv_code) == 0 else iconv_code
+        # For alias, we strip and to lower them first, and remove all empty entries
+        alias = row[3:]
+        self.alias = tuple(
+            filter(lambda x: len(x) != 0,
+                   map(lambda x: x.strip().lower(), alias)))
+
+
+def _get_self_dir() -> Path:
+    return Path(__file__).resolve().parent
+
+
+def _extract_tokens() -> list[LanguageToken]:
+    rv: list[LanguageToken] = []
+    csv_file = _get_self_dir() / 'encoding_table.csv'
+
+    with open(csv_file, 'r', encoding='utf-8', newline='') as f:
+        reader = csv.reader(f, delimiter='\t')
+        for row in reader:
+            rv.append(LanguageToken(row))
+
+    return rv
+
+
+def _render_cpp(tokens: list[LanguageToken]) -> None:
+    loader = jinja2.FileSystemLoader(_get_self_dir())
+    environment = jinja2.Environment(loader=loader)
+    template = environment.get_template('encoding_table.cpp.jinja')
+
+    cpp_file = _get_self_dir() / 'encoding_table.cpp'
+    with open(cpp_file, 'w', encoding='utf-8') as f:
+        f.write(template.render(tokens=tokens))
+
+
+if __name__ == '__main__':
+    tokens = _extract_tokens()
+    _render_cpp(tokens)