Make HTMLParser handle Unicode.

[jira-zephyrbot.git] / jirabot.py

#!/usr/bin/python
import calendar
import feedparser
import formatter
import htmlentitydefs
import htmllib
import mechanize
import os
import random
import string
import StringIO
import time
import traceback
import urlparse
import zephyr

zephyr_sender = 'jira'
zephyr_class = 'andersk-test'
time_file = 'jirabot.time'

class UnicodeHTMLParser(htmllib.HTMLParser):
    entitydefs = dict((k, unichr(v)) for (k, v) in htmlentitydefs.name2codepoint.items())

    def convert_charref(self, name):
        try:
            n = int(name)
        except ValueError:
            return
        return self.convert_codepoint(n)

    def convert_codepoint(self, codepoint):
        return unichr(codepoint)

def jira_init():
    b = mechanize.Browser()
    b.set_handle_robots(False)
    b.add_client_certificate("https://idp.mit.edu:9443", "cert.pem", "cert.pem")
    b.addheaders = [("Accept-Language", "en-us,en;q=0.5"),]
    return b

def jira_login(b):
    b.open("https://jira.mit.edu/jira/secure/Dashboard.jspa")
    try:
        b.follow_link(text="MIT Touchstone")
    except mechanize.LinkNotFoundError:
        return
    if (urlparse.urlparse(b.geturl())[1] == "jira.mit.edu"):
        return
    b.select_form("wayfForm1")
    b.submit()
    b.select_form(predicate=lambda f: any(c.name == 'login_certificate'
                                          for c in f.controls))
    b.submit()
    b.select_form(nr=0)
    b.submit()

def feed_to_zephyrs(thing, rss, parse):
    zephyrs = []
    try:
        feed = feedparser.parse(rss)
        for e in feed.entries:
            global old_time, new_time
            t = int(calendar.timegm(e.date_parsed))
            if t <= old_time:
                continue
            if t > new_time:
                new_time = t
            try:
                z = parse(e)
            except:
                z = zerror("Error parsing " + thing + ":\n" + e.id + "\n" + traceback.format_exc())
            zephyrs.append((t, z))
    except:
        zephyrs.append((0, zerror("Error parsing " + thing + "s feed:\n" + traceback.format_exc())))
    return zephyrs

def parse_issue(e):
    issue = urlparse.urlparse(e.id)[2].rsplit('/', 1)[1]
    url = e.id
    msg = e.id + "\nThis issue was updated."

    return zephyr.ZNotice(
        sender=zephyr_sender,
        auth=False,
        opcode='auto',
        cls=zephyr_class,
        instance=issue,
        fields=[e.title, msg],
    )

def parse_comment(e):
    url = urlparse.urlunsplit(urlparse.urlparse(e.id)[0:3] + (None,None))
    issue = url.rsplit('/', 1)[1]

    s = StringIO.StringIO()
    parser = UnicodeHTMLParser(formatter.AbstractFormatter(formatter.DumbWriter(s)))
    parser.feed(e.summary.rsplit('<table>', 1)[0])
    parser.close()
    comment = s.getvalue()

    msg = e.author + " added a comment:\n" + comment.rstrip()

    return zephyr.ZNotice(
        sender=zephyr_sender,
        auth=False,
        opcode='auto',
        cls=zephyr_class,
        instance=issue,
        fields=[e.title, msg],
    )

def zerror(msg):
    return zephyr.ZNotice(
        sender=zephyr_sender,
        auth=False,
        opcode='auto',
        cls=zephyr_class,
        instance='jira-error',
        fields=['Jira bot error', msg]
    )

b = jira_init()
zephyr.init()

while True:
    time_file_new = time_file + '.' + ''.join(random.sample(string.letters, 8))

    try:
        os.rename(time_file, time_file_new)
    except OSError:
        print "warning: could not acquire timestamp lock"
        time.sleep(17)
        continue

    jira_login(b)
    b.open("https://jira.mit.edu/jira/sr/jira.issueviews:searchrequest-rss/temp/SearchRequest.xml?&pid=10185&updated%3Aprevious=-1w&sorter/field=updated&sorter/order=DESC&tempMax=1000")
    issues_rss = b.response().read()
    b.open("https://jira.mit.edu/jira/sr/jira.issueviews:searchrequest-comments-rss/temp/SearchRequest.xml?&pid=10185&updated%3Aprevious=-1w&sorter/field=updated&sorter/order=DESC&tempMax=1000")
    comments_rss = b.response().read()

    old_time = int(open(time_file_new).read())
    new_time = old_time

    zephyrs = (feed_to_zephyrs('issue', issues_rss, parse_issue) +
               feed_to_zephyrs('comment', comments_rss, parse_comment))
    zephyrs.sort(key=lambda tz: tz[0])
    for (t, z) in zephyrs:
        z.send()

    open(time_file_new, 'w').write(str(new_time))
    os.rename(time_file_new, time_file)

    time.sleep(60)