neveragaindottech.github.io/_bin/generate-signature-files

#!/usr/bin/env python

import commands
import re
import requests
import sys

import os
token = os.environ['GITHUB_ACCESS_TOKEN']

def errprint(s):
    sys.stderr.write(s)
    sys.stderr.flush()

prs = []

for page in range(1, 10):
    more_prs = requests.get('https://api.github.com/repos/neveragaindottech/neveragaindottech.github.io/pulls?access_token=%s&page=%d' % (token, page)).json()
    prs += more_prs
    errprint('fetched %d PRs\n' % len(prs))
    if not more_prs:
        break

for pr in reversed(prs):
    pr_num = pr['number']
    author = pr['user']['login']
    errprint('PR #%s from %s: ' % (pr_num, author))
    diff_lines = commands.getoutput('./pr-diff %s' % pr_num).strip().split('\n')
    if len(diff_lines) == 1 and diff_lines[0].startswith('+<'):
        line = diff_lines[0].lstrip('+')
        props = {}

        props['github_user'] = author
        props['pull_request_number'] = str(pr_num)

        match = re.search(r'<a *href *= *"([^"]*)', line, re.I)
        props['link'] = match and match.group(1)

        text_line = re.sub(r'<[^>]*>', '', line)
        parts = text_line.split(',', 2)
        if len(parts) == 3:
            props['name'], props['occupation_title'], props['affiliation'] = parts
        elif len(parts) == 2:
            props['name'], props['affiliation'] = parts
        else:
            props['name'] = text_line

        filename = '../_signatures/%s.md' % author
        if os.path.exists(filename):
            errprint('%s ALREADY EXISTS\n' % filename)
            continue

        with open(filename, 'w') as file:
            file.write('---\n')
            for key, value in props.items():
                value = value and value.strip()
                if value:
                    file.write('  %s: %s\n' % (key, value))
            file.write('---\n')
        errprint('ok, made %s\n' % filename)

    else:
        errprint('BAD; diff is:\n')
        errprint('  ' + '\n  '.join(diff_lines) + '\n')