Download all github repos that you have open in firefox tabs

tabs-to-repos.py

#!/usr/bin/env python3
"""
_____________
tabs-to-repos -- Quickly gnt all of the github repos that you have open in firefox tabs.

Pipe the output to your shell to actually clone them, otherwise it just prints the commands and some comments ;)

@trevorj 12/2017
https://gist.github.com/akatrevorjay/3fb5bc753edd42b2abd0accb4dde4cdc
"""

import glob
import logging
import os
import re
import sys

import lz4.block  # noqa: bug in firefox_tabs requires this, it only imports bare `lz4`
import firefox_tabs

log = logging.getLogger(__name__)  # noqa

FF_PROF_GLOB = '~/.mozilla/firefox*/*.*/'
FF_PROF_SESS_REC_GLOB = '%s/sessionstore-backups/recovery.jsonlz4' % FF_PROF_GLOB

RE_GITHUB_REPO = re.compile(
    r'(?P<scheme>https?)://(www\.|)github.com/(?P<user>[-_0-9a-zA-Z]+)/(?P<repo>[-_0-9a-zA-Z]+)/?$',
    re.IGNORECASE,
)


class AttrDict(dict):
    """
    >>> m = AttrDict(omg=True, whoa='yes')
    """

    def __init__(self, *args, **kwargs):
        super(AttrDict, self).__init__(*args, **kwargs)
        self.__dict__ = self


def _glob_me_prof(pat=FF_PROF_GLOB, sort_by=lambda x: -os.stat(x).st_mtime):
    pat = os.path.expanduser(pat)

    paths = glob.glob(pat)
    sorted_paths = sorted(paths, key=sort_by)

    if not sorted_paths:
        log.warn('Could not glob %r. What OS are you on?', glob)

    for p in sorted_paths:
        yield p


def _load_firefox_data(path):
    path = os.path.expanduser(path)
    session_data = firefox_tabs.load_data(path)
    return session_data


def _iter_firefox_tabs(session_data, container_factory=AttrDict):
    for window_idx, window in enumerate(session_data['windows']):
        for tab in window['tabs']:
            current_tab = tab['entries'][-1]

            tab = container_factory(
                url=current_tab['url'],
                title=current_tab['title'],
                window=window_idx,
            )

            yield tab


def main(re_url=RE_GITHUB_REPO):
    path = next(_glob_me_prof(FF_PROF_SESS_REC_GLOB))

    log.info('Loading from firefox session store backup: %r', path)
    session_data = _load_firefox_data(path)

    tabs = list(_iter_firefox_tabs(session_data))
    log.info('Loaded session data: num_tabs=%d', len(tabs))

    for tab in tabs:
        log.debug('Checking tab: url=%s', tab.url)

        m = re_url.match(tab.url)
        if not m:
            continue

        md = AttrDict(m.groupdict())

        print('# %s' % tab.url)
        print('git clone https://github.com/%s/%s.git' % (md.user, md.repo))

    # avoids gc overhead at exit.
    sys.exit(0)


if __name__ == '__main__':
    logging.basicConfig(
        level=logging.DEBUG,
        # Match squid log output
        format='%(asctime)s| %(name)s/%(process)d: %(message)s '
        '@%(funcName)s:%(lineno)d #%(levelname)s',
        datefmt='%Y/%m/%d %H:%M:%S'
    )

    main()