#!/usr/bin/env python

# Trivet
# Copyright (c) 2023 by Stacy Prowell.  All rights reserved.
# https://gitlab.com/binary-tools/trivet

"""
Read the XML Unicode database file and construct a Rust version for use
by Trivet.  This converts the XML into a literal data structure in Rust
using the following process.

The result is a BTree mapping &str to char.  The strings come from the
names (NA attribute) and aliases (name attributes of contained name-alias
elements).

Obtain the Unicode database from https://unicode.org/ucd/.  The desired
file is the flat complete set, and is named ucd.all.flat.xml.

This only needs to be run if the Unicode database needs to be updated.
Run with:

    $ python3 build_unicode_db.py > ucd.rs

Check the output to make sure it looks reasonable, then move it into place
in the src/parse/strings folder.

    $ mv ucd.rs ../src/parse/strings/ucd.rs
"""

import xml.etree.ElementTree as ET

def main():
    """Entry point when run from the prompt."""
    # Open and read the Unicode XML data file.
    tree = ET.parse("ucd.all.flat.xml")
    root = tree.getroot()

    # Extract all char children.
    ucd = {}
    for character in root.iter('{http://www.unicode.org/ns/2003/ucd/1.0}char'):
        # If there is a code point in the attribute, then get it.
        if 'cp' in character.attrib:
            code_point = character.attrib['cp']
            # Now try to find the name.
            if 'na' in character.attrib:
                name = character.attrib['na']
                if name != '':
                    # Found the name.
                    ucd[name] = f'\\u{{{code_point}}}'
            # Now print any aliases.
            for alias in character.iter('{http://www.unicode.org/ns/2003/ucd/1.0}name-alias'):
                if 'alias' in alias.attrib:
                    alias = alias.attrib['alias']
                    if alias != '':
                        ucd[alias] = f'\\u{{{code_point}}}'
    
    # Now construct the Rust file.
    print("// trivet")
    print("// copyright")
    print("")
    print("//! Provide the Unicode database.")
    print("")
    print("/// The Unicode database.")
    print("""///
/// This is generated from the complete database, which can be obtained from
/// [unicode.org](https://unicode.org/ucd/).  It is an array of pairs, with
/// each pair consisting of a Unicode name or alias and the corresponding code
/// point.
""")
    print("pub const UCD: &[(&str, char)] = &[")
    for name,code_point in ucd.items():
        print(f"    (\"{name.upper()}\", '{code_point}'),")
    print("];")



if __name__ == '__main__':
    main()