Alma Invoice Attachments API

invoices.py

import os, sys, requests, base64, re
import xml.etree.ElementTree as ET

NS = {'exl': 'http://com/exlibris/repository/acq/invoice/xmlbeans'} 
URL = 'https://api-na.hosted.exlibrisgroup.com'

def process_invoice( invoice ):
    id = invoice.find('exl:unique_identifier', NS).text
    number_of_attachments = invoice.find('exl:number_of_attachments', NS)
    if (number_of_attachments is None or int(number_of_attachments.text) > 0):
        get_attachments(id)

def get_attachments(invoice_id):
    attachments = get(f"/almaws/v1/acq/invoices/{invoice_id}/attachments")
    os.makedirs(invoice_id, exist_ok = True) 
    for attachment in attachments.findall('attachment'):
        print(attachment.attrib['link'])
        file_name, content = get_attachment(attachment.attrib['link'])
        if not re.match(r"^ERP\..*\.xml$", file_name): # exclude export file
            file = open(os.path.join(invoice_id, file_name) , 'wb')
            file.write(base64.b64decode(content))
            file.close()

def get_attachment(link): 
    attachment = get(link + "?expand=content")
    return attachment.find('file_name').text, attachment.find('content').text

def get(url):
    headers = {
        'authorization': f"apikey {os.environ['ALMA_APIKEY']}",
        'accept': 'application/xml'
    }
    if not url.startswith('https://'):
        url = URL + url
    response = requests.get(url, headers=headers)
    response.raise_for_status()
    return ET.fromstring(response.content)

def main():
    directory = os.fsencode(sys.argv[1])
    for file in os.listdir(directory):
        filename = os.fsdecode(file)
        if filename.endswith(".xml"):
            xml = ET.parse(os.path.join(directory, file)).getroot()
            for invoice in xml.findall('./exl:invoice_list/exl:invoice', NS):
                process_invoice(invoice)

if len(sys.argv) <= 1:
    print("Missing directory name")
    exit()

main()

Usage:

$ python3 invoices.py {{directory}}