jinson

#!/usr/bin/env python3

"""Jinson - Convert JIRA's XML to JSON

Usage:
  jinson <entities.xml>

Options:
  -h --help     Show this screen.
  --version     Show version.
"""

from docopt import docopt
from datetime import datetime
import json
import sys
import xml.etree.ElementTree as ET


class Converter:
    def __init__(self):
        self.bytype = {}

    def load(self, tree):
        def elementdict(el):
            d = dict(el.attrib)
            for child in el:
                if not child.attrib and not len(child):
                    if child.text.strip():
                        d[child.tag] = child.text.strip()
                else:
                    d[child.tag] = elementdict(child)
            return d

        root = tree.getroot()
        for child in root:
            if 'id' in child.attrib:
                self.bytype.setdefault(child.tag, {})[child.attrib['id']] = elementdict(child)
            else:
                self.bytype.setdefault(child.tag, []).append(elementdict(child))

    def convert(self):
        r = self.bytype

        def nest(data, child, key, parent, prop, dictkey=None):
            if child not in data:
                return
            for i, item in sorted(data[child].items()):
                dest = data[parent][item[key]]
                if dictkey:
                    dest.setdefault(prop, {})[item[dictkey]] = item
                else:
                    dest.setdefault(prop, []).append(item)
                del item[key]
            del data[child]

        nest(r, 'Membership', 'childId', 'User', 'membership')
        nest(r, 'ChangeItem', 'group', 'ChangeGroup', 'items')
        nest(r, 'ChangeGroup', 'issue', 'Issue', 'changes')
        nest(r, 'Action', 'issue', 'Issue', 'actions')
        nest(r, 'Worklog', 'issue', 'Issue', 'worklog')
        nest(r, 'FileAttachment', 'issue', 'Issue', 'attachments', 'id')
        nest(r, 'AuditChangedValue', 'logId', 'AuditLog', 'changes')
        nest(r, 'AuditItem', 'logId', 'AuditLog', 'items')

        # generated the path pointing to the attachment data saved in the zip file
        for i in r['Issue'].values():
            p = r['Project'][i['project']]
            for a in i.get('attachments', {}).values():
                unknown_thing = '10000'
                a['data'] = '/'.join(('data/attachments', p['key'], unknown_thing, i['key'], a['id']))

        # store users in a dict keyed by ApplicationUser.userKey
        users = { u['lowerUserName']: u for u in r['User'].values() }
        for appuser in r['ApplicationUser'].values():
            u = users.get(appuser['lowerUserName'])
            r.setdefault('users', {})[appuser['userKey']] = u
        del r['ApplicationUser']
        del r['User']

        for g in r['Group'].values():
            r.setdefault('groups', {})[g['lowerGroupName']] = g
        del r['Group']

        # store actors in the appropriate projects, keyed by role type
        for i, role in list(sorted(r['ProjectRoleActor'].items())):
            if 'pid' in role:
                r['Project'][role['pid']].setdefault('roles', {}).setdefault(role['projectroleid'], []).append(role)
                del role['pid']
                del role['projectroleid']
                del r['ProjectRoleActor'][i]

        return r

def main(arguments):
    data = arguments['<entities.xml>']
    if data == '-':
        data = sys.stdin
    tree = ET.parse(data)
    converter = Converter()
    converter.load(tree)
    data = converter.convert()
    print(json.dumps(data, sort_keys=True, indent=4))

if __name__ == '__main__':
    arguments = docopt(__doc__, version='Jinson 0.1')
    main(arguments)