|
import os |
|
import csv |
|
import sys |
|
from glob import glob |
|
from xml.etree import ElementTree as ET |
|
|
|
|
|
xsd_dir_path = 'PUT BASE DIRECTORY HERE :)' |
|
no_duplicate_tables = '--no-duplicates' in sys.argv |
|
done_tables = set() |
|
|
|
def strip_ns(root_node): |
|
for el in root_node.iter(): |
|
if '}' in el.tag: |
|
el.tag = el.tag.split('}', 1)[1] # strip all namespaces |
|
|
|
|
|
def is_table(node): |
|
if not node.tag == 'complexType': |
|
return False |
|
el_type = node.attrib.get('name') |
|
if not el_type: |
|
return False |
|
return True if el_type.lower().startswith('t') else False |
|
|
|
|
|
def documentation(node): |
|
return next(node.iter('documentation')).text.strip().split('\n')[0] |
|
|
|
|
|
def handle_element(element_node): |
|
# Nome da Coluna |
|
name = element_node.attrib["name"] |
|
|
|
# Descrição da Coluna |
|
description = documentation(element_node) |
|
|
|
# Opcional? |
|
minoccurs = element_node.attrib.get("minOccurs") |
|
minoccurs = int(minoccurs) if minoccurs else 99 |
|
|
|
optional = True if int(minoccurs) < 1 else False |
|
|
|
return (name, optional, description,) |
|
|
|
|
|
def handle_xsd_file(filepath): |
|
root = ET.parse(filepath).getroot() |
|
strip_ns(root) |
|
|
|
root_documentation = documentation(root) |
|
|
|
tables = [el for el in root.iter() if is_table(el)] |
|
|
|
for tidx, t in enumerate(tables, start=1): |
|
table_name = t.attrib["name"] |
|
|
|
# skip if table was already handled |
|
if no_duplicate_tables: |
|
if table_name in done_tables: |
|
continue |
|
|
|
table_description = documentation(t) |
|
elements = [handle_element(e) for e in t.iter() |
|
if e.tag == 'element'] |
|
for eidx, e in enumerate(elements, start=1): |
|
ename, eoptional, edescr = e |
|
yield ( |
|
root_documentation, |
|
tidx, |
|
table_name, |
|
table_description, |
|
eidx, |
|
ename, |
|
eoptional, |
|
edescr, |
|
) |
|
done_tables.add(table_name) |
|
|
|
|
|
writer = csv.writer(sys.stdout) |
|
for f in glob(xsd_dir_path + '/*'): |
|
basename = os.path.basename(f) |
|
for row in handle_xsd_file(f): |
|
writer.writerow([basename, *row]) |
