xsd

main.py

import xmlschema
import json
import yaml
import glob

ns = '{http://www.w3.org/XML/1998/namespace}'
xs = '{http://www.w3.org/2001/XMLSchema}'


def get_docs(element, lang):
    if hasattr(element.annotation, 'documentation'):
        for doc in element.annotation.documentation:
            if f'{ns}lang' in doc.attrib and lang in doc.attrib[f'{ns}lang'] and doc.text:
                return doc.text.strip()
    if hasattr(element.type.annotation, 'documentation'):
        for doc in element.type.annotation.documentation:
            if f'{ns}lang' in doc.attrib and lang in doc.attrib[f'{ns}lang'] and doc.text:
                return doc.text.strip()
    if element.type.base_type and hasattr(element.type.base_type.annotation, 'documentation'):
        for doc in element.type.base_type.annotation.documentation:
            if f'{ns}lang' in doc.attrib and lang in doc.attrib[f'{ns}lang'] and doc.text:
                return doc.text.strip()
    return ''


def get_dictionary(element, path=[]):
    element_doc_pl = get_docs(element, 'pl')
    element_doc_en = get_docs(element, 'en')
    elements = []
    attributes = []
    element_path = '/'.join(['']+path)
    for attr in element.attributes:
        attr = element.attributes[attr]
        doc_pl = get_docs(attr, 'pl')
        doc_en = get_docs(attr, 'en')
        attr_path = '/'.join(['']+path+[f'@{attr}'])
        attributes.append({'name': attr, 'path': attr_path,
                          'doc_pl': doc_pl, 'doc_en': doc_en})
    for child in element:
        if child.local_name is None:
            continue
        if child.local_name == 'Signature':
            continue
        child_dict = get_dictionary(child, path+[child.local_name])
        elements.append(child_dict)
    return {'name': element.local_name,
            'path': element_path,
            'doc_pl': element_doc_pl,
            'doc_en': element_doc_en,
            'attributes': attributes,
            'elements': elements}


def get_document(fname):
    my_schema = xmlschema.XMLSchema(fname)
    for key in my_schema.elements:
        document = my_schema.elements[key]
        document_tree = get_dictionary(document, [document.local_name])
        doc_pl = None
        doc_en = None
        for element in document.source.parent_map:
            if element.tag == f'{xs}documentation':
                if f'{ns}lang' in element.attrib and element.attrib[f'{ns}lang'] == 'pl' and element.text:
                    doc_pl = element.text.strip()
                if f'{ns}lang' in element.attrib and element.attrib[f'{ns}lang'] == 'en' and element.text:
                    doc_en = element.text.strip()
        return {'name': document.name, 'document': document_tree, 'doc_pl': doc_pl, 'doc_en': doc_en}


documents = []
for fname in glob.glob('./xsd/full/AIS_Import_publiczna/*.xsd'):
    document = get_document(fname)
    if document:
        documents.append(document)

with open('ais.yml', 'w+', encoding='utf-8') as f:
    yaml.dump(documents, f, allow_unicode=True, Dumper=yaml.BaseDumper)
    f.close()