def __init__(self, seen_before):

self._seen_before = os.path.expanduser(seen_before) if seen_before else None

@property

def feed_url(self):

return os.environ['FEED_URL']

@property

def email_from(self):

return os.environ['EMAIL_FROM']

@property

def email_to(self):

return os.environ['EMAIL_TO']

@property

def email_cc(self):

return os.environ.get('EMAIL_CC')

@property

def email_bcc(self):

return os.environ.get('EMAIL_BCC')

@property

def email_subject(self):

return os.environ['EMAIL_SUBJECT']

@property

def smtp_server(self):

return os.environ['SMTP_SERVER']

@property

def html_preamble(self):

return etree.parse(open(os.environ['HTML_PREAMBLE_FILE'], 'rb')).getroot()

@property

def html_css(self):

return open(os.environ['HTML_CSS_FILE'], 'rb').read().decode('utf-8')

@property

def text_preamble(self):

return open(os.environ['TEXT_PREAMBLE_FILE'], 'rb').read().decode('utf-8')

def __call__(self):

vacancies = self.get_vacancies()

current_vacancies = set(vacancies.xpath('/vacancies/vacancy/@id'))

if self._seen_before and os.path.exists(self._seen_before):

seen_vacancies = set(open(self._seen_before, 'r').read().split())

new_vacancies = current_vacancies - seen_vacancies

else:

new_vacancies = set()

html, text = self.generate_email_bodies(vacancies, new_vacancies)

msg = self.compose_email(html, text)

self.send_email(msg)

if self._seen_before:

with open(self._seen_before, 'w') as f:

for vacancy_id in sorted(current_vacancies):

f.write('{}\n'.format(vacancy_id))

def get_vacancies(self):

return etree.parse(urlopen(self.feed_url))

def generate_email_bodies(self, vacancies, new_vacancies):

html_vacancies = E('div')

html = E('html',

E('head', E('style', self.html_css, type='text/css')),

E('body', self.html_preamble, html_vacancies))

text_body = [self.text_preamble]

if len(vacancies.xpath('/vacancies/vacancy')) == 0:

raise SystemExit(0)

for i, vacancy in enumerate(vacancies.xpath('/vacancies/vacancy')):

vacancy_id = vacancy.attrib['id']

html_description = etree.fromstring(vacancy.xpath("description[@media_type='text/html']")[0].text,

parser=etree.HTMLParser())[0][0]

first_para = html_description.xpath(".//text()[normalize-space(.) and not(contains(., 'INTERNAL') or contains(., 'ADVERTISEMENT'))]")[0].strip()

html_first_para = E('div', first_para, **{'class': 'description'})

tags = []

if 'INTERNAL' in html_description.text:

tags.append('internal applicants only')

if 'ADVERTISEMENT' in html_description.text:

tags.append('re-advertisement')

tags = ' ({0})'.format(', '.join(tags)) if tags else ''

try:

closes = dateutil.parser.parse(vacancy.find('closes').text)

except Exception:

closes, closes_soon = 'unknown', False

else:

closes_soon = (closes - datetime.datetime.now(pytz.utc)).total_seconds() < 3600 * 24 * 2

closes = closes.strftime('%a, %d %b %Y, %I:%M %p')

html_vacancy = E('div',

E('h1', vacancy.find('label').text),

E('div',

E('span', vacancy.find('salary').find('label').text, **{'class': 'salary'}),

'; closes: ',

E('span', closes, **{'class': 'closes' + (' closes-soon' if closes_soon else '')}),

tags,

**{'class': 'byline'}

),

html_first_para,

E('div',

E('a', 'More details\N{HORIZONTAL ELLIPSIS}', href=vacancy.find('webpage').text)),

**{'class': 'vacancy'}

)

if vacancy_id in new_vacancies:

html_vacancy[0].text += " "

html_vacancy[0].append(E('span', '\N{BLACK STAR} new', **{'class': 'new'}))

html_vacancies.append(html_vacancy)

text_body.extend([

'_' * 70,

'\n\n',

textwrap.fill(vacancy.find('label').text),

' *NEW*' if vacancy_id in new_vacancies else '',

'\n\n',

vacancy.find('salary').find('label').text,

'\nCloses: ',

closes,

tags,

'\n\n',

textwrap.fill(first_para),

'\n\n'

'More details: https://data.ox.ac.uk/v/',

vacancy.attrib['id'],

'\n'

])

html_body = inline_css(etree.tostring(html, method='html'))

text_body = ''.join(text_body).encode('utf-8')

return html_body, text_body

def compose_email(self, html_body, text_body):

msg = email.mime.multipart.MIMEMultipart('alternative')

msg['Subject'] = self.email_subject

msg['From'] = self.email_from

msg['To'] = self.email_to

if self.email_cc:

msg['CC'] = self.email_cc

msg.attach(email.mime.text.MIMEText(text_body, 'plain', 'utf-8'))

msg.attach(email.mime.text.MIMEText(html_body, 'html', 'utf-8'))

return msg

def send_email(self, msg):

s = smtplib.SMTP(self.smtp_server)

recipients = self.email_to.split(',') + (self.email_cc or '').split(',') + (self.email_bcc or '').split(',')

recipients = [email.utils.parseaddr(r.strip())[1] for r in recipients if r.strip()]

s.sendmail(email.utils.parseaddr(self.email_from)[1],

recipients,

msg.as_string())

bank_holidays_json_url = 'https://www.gov.uk/bank-holidays.json'

def __call__(self, *arg, **kwargs):

bank_holidays = json.loads(requests.get(self.bank_holidays_json_url).text)

bank_holidays = set(datetime.datetime.strptime(d['date'], "%Y-%m-%d").date()

for d in bank_holidays['england-and-wales']['events'])

today = datetime.date.today()

def working_day(d):

return (d.weekday() < 5 # Not a weekend

and d not in bank_holidays # Not a bank holiday

and not (d.month == 12 and 25 <= d.day <= 31)) # Not in Christmas closedown

# Only send if today is actually a working day, and...

if not working_day(today):

logger.info("Today is not a working day; not continuing")

return

# Look back through the week. If any of those days was also a working day, stop now.

for i in range(today.weekday(), 0, -1):

previous_date = today - datetime.timedelta(i)

if working_day(previous_date):

logger.info("%s was a working day; not continuing", previous_date)

return