dix2text.py

import argparse
import xml.etree.ElementTree as ET

def dix2text(input, source, target):
    if input.startswith("http"):
        import urllib.request

        with urllib.request.urlopen(input) as response:
            # Read the content of the file
            content = response.read().decode('utf-8')
            root = ET.fromstring(content)
    else:
        tree = ET.parse(input)
        root = tree.getroot()

    if root.tag != "dictionary":
        raise IOError("Invalid .dix file (root key is not <dictionary>)")

    def extract_text(el):
        text = " ".join(el.itertext())
        return text.strip()

    i = 0
    added = {}
    duplicates = 0
    with open(source, "w", encoding="utf-8") as src, \
        open(target, "w", encoding="utf-8") as tgt:
        for c in root:
            if c.tag == 'section':
                section = c
                for ch in section:
                    if ch.tag == 'e':
                        p = ch[0]
                        if p.tag != 'p':
                            continue
                        l = extract_text(p[0])
                        r = extract_text(p[1])
                        if len(l) > 0 and len(r) > 0:
                            if not l in added:
                                src.write(l + "\n")
                                tgt.write(r + "\n")
                                i += 1
                                added[l] = True
                            else:
                                duplicates += 1
    return i, duplicates

if __name__ == "__main__":
    parser = argparse.ArgumentParser(description='Convert Apertium .dix dictionary files to text')
    parser.add_argument('input',
        type=str,
        help='Input .dix file path or URL')
    parser.add_argument('source',
        type=str,
        default="source.txt",
        help='Source output file. Default: %(default)s')
    parser.add_argument('target',
        type=str,
        default="target.txt",
        help='Target output file. Default: %(default)s')
    args = parser.parse_args()

    i, duplicates = dix2text(args.input, args.source, args.target)
    print(f"{i} entries (skipped {duplicates} duplicates)")
    print(f"W\t{args.source}")
    print(f"W\t{args.target}")