Export open Github issues for offline use

github_issues_exporter.py

#!/usr/bin/env python
# -*- coding: utf-8 -*-

'''
pip install --upgrade PyGithub

TODO:
- Configurable output directory
- Output more issue info
- Better rendering
'''

import argparse
import os
import sys
import errno

from github import Github


_OUTPUT_DIR = 'issues'
_DIVIDER_LEN = 79


def _makedirs(path):
    try:
        os.makedirs(path)
    except OSError as e:
        if e.errno == errno.EEXIST and os.path.isdir(path):
            pass
        else:
            raise


def _export_issue(issue, export_path):
    export_path = os.path.join(export_path, '%04d.txt' % (issue.number))

    f = open(export_path, 'w')

    f.write('#%d\n' % (issue.number,))
    f.write('%s\n%s\n\n' % (issue.title, '='*len(issue.title.encode('utf-8'))))

    f.write('%s %s\n\n' % (issue.created_at, issue.user.login))

    state = '%s %s' % (issue.state, issue.closed_at if issue.closed_at else '')
    f.write('%s\n\n' % (state,))

    assignee = 'assigned to %s' % (issue.assignee.login,) if issue.assignee else 'unassigned'
    f.write('%s\n\n' % (assignee,))

    labels = ', '.join([l.name for l in issue.labels])
    f.write('%s\n\n' % (labels,))

    f.write('%s\n\n' % (issue.body.replace('\r\n', '\n').encode('utf-8')))

    comment_pager = issue.get_comments()
    current_page = 0
    while True:
        comments = comment_pager.get_page(current_page)
        current_page += 1
        current_page += 1
        if not comments:
            break

        for comment in comments:
            f.write('%s\n\n' %('-'*_DIVIDER_LEN,))
            f.write('%s %s\n\n' % (comment.created_at, comment.user.login))
            f.write('%s\n\n' % (comment.body.replace('\r\n', '\n').encode('utf-8')))

    f.close()


def _export_issues_for_project(gh, projname, state, labels, exc_labels, export_path):
    repo = gh.get_repo(projname)

    real_labels = None
    if labels:
        real_labels = [repo.get_label(label_str) for label_str in labels.split(',')]

    if exc_labels:
        exc_labels = exc_labels.split(',')
    else:
        exc_labels = ()

    # It seems impossible to get assigned and unassigned issues in a single query,
    # so we'll do one for each state.
    for assignee in ['*', 'none']:
        for issue in repo.get_issues(state=state, labels=real_labels, assignee=assignee):
            if [l for l in issue.labels if l.name in exc_labels]:
                continue

            print '#%d %s' % (issue.number, issue.title)

            _export_issue(issue, export_path)


if __name__ == '__main__':
    parser = argparse.ArgumentParser(description='Export Github issues to disk')
    parser.add_argument('projnames', metavar='username/projectname', type=str, nargs='+',
                        help='fully qualified "username/projectname" of projects to export issues from')
    parser.add_argument('--auth-token', action='store',
                        help='Github auth token to use (optional, but helps prevent request rate denial)')
    parser.add_argument('--state', action='store',
                        help='Indicates the state of the issues to return. Can be either open, closed, or all. Default: open')
    parser.add_argument('--labels', action='store',
                        help='A list of comma separated label names. Example: bug,ui,@high')
    parser.add_argument('--exc_labels', action='store',
                        help='A list of comma separated label names TO EXCLUDE. Example: enhancement,@low')

    args = parser.parse_args()

    gh = Github(args.auth_token)

    for projname in args.projnames:
        export_path = os.path.normpath(os.path.join(_OUTPUT_DIR, projname))
        _makedirs(export_path)
        _export_issues_for_project(
            gh, projname, args.state, args.labels, args.exc_labels, export_path)

    sys.exit(0)