datatracker/bin/test-crawl

#!/usr/bin/env python

import os, sys, re, datetime, argparse, traceback, tempfile

# args
parser = argparse.ArgumentParser(
    description="""Perform a test crawl of the project. For each found URL, the HTTP
    response status is printed. If it's not OK/redirect, FAIL is
    printed - in case of errors, a stacktrace is also included.""")
parser.add_argument('urls', metavar='URL', nargs='*',
                    help='One or more URLs to start the crawl from')
parser.add_argument('--urls', '-u', dest='url_file',
                    help='file with URLs to start the crawl from')
parser.add_argument('--slow', dest='slow_threshold', type=float, default=1.0,
                    help='responses taking longer than this (in seconds) results in SLOW being printed')
parser.add_argument('--settings', dest='settings', help='custom settings file')

args = parser.parse_args()

# boilerplate
basedir = os.path.abspath(os.path.join(os.path.dirname(__file__), "../"))
sys.path = [ basedir ] + sys.path

settings_module = args.settings or "ietf.settings"

os.environ.setdefault("DJANGO_SETTINGS_MODULE", settings_module)

import django
import django.test

# prevent memory from leaking when settings.DEBUG=True
from django.db import connection
class DontSaveQueries(object):
    def append(self, x):
        pass
connection.queries = DontSaveQueries()

MAX_URL_LENGTH = 500

slow_threshold = args.slow_threshold

initial_urls = []
initial_urls.extend(args.urls)

if args.url_file:
    with open(args.url_file) as f:
        for line in f:
            line = line.partition("#")[0].strip()
            if line:
                initial_urls.append(line)

if not initial_urls:
    initial_urls.append("/")

visited = set()
urls = {} # url -> referrer
referrers = {}

def strip_url(url):
    if url.startswith("http://testserver"):
        url = url[len("http://testserver"):]
    return url

def extract_html_urls(content):
    for m in re.finditer(r'(<(?:a|link) [^>]*href=[\'"]([^"]+)[\'"][^>]*>)', content):
        if re.search(r'rel=["\']?nofollow["\']', m.group(1)):
            continue

        url = strip_url(m.group(2))
        if len(url) > MAX_URL_LENGTH:
            continue # avoid infinite GET parameter appendages

        if not url.startswith("/"):
            continue

        if url.startswith("//"):
            continue

        yield url

django.setup()
client = django.test.Client()

for url in initial_urls:
    urls[url] = "[initial]"

errors = 0
count = 0

start_time = datetime.datetime.now()
fh, fn = tempfile.mkstemp(prefix="test-crawl-", suffix=".log", dir="../")
logfile = open(fn, "w")
os.close(fh)

def log(s):
    print(s)
    logfile.write(s)
    logfile.write('\n')

def get_referrers(url):
    ref_list = []
    while url in referrers:
        url = referrers[url]
        if url in ref_list:
            log("Circular referral list, discovered at %s" % url)
            break
        ref_list.append(url)
    return ref_list

while urls:
    url, referrer = urls.popitem()

    visited.add(url)

    try:
        timestamp = datetime.datetime.now()
        r = client.get(url)
        elapsed = datetime.datetime.now() - timestamp
    except KeyboardInterrupt:
        log(" ... was fetching %s" % url)
        sys.exit(1)
    except:
        log("500 %.3fs %s FAIL (from: [ %s ])" % ((datetime.datetime.now() - timestamp).total_seconds(), url, (",\n\t".join(get_referrers(url)))))
        log("=============")
        log(traceback.format_exc())
        log("=============")
        errors += 1
    else:
        tags = []

        if r.status_code in (301, 302):
            u = strip_url(r["Location"])
            if u not in visited and u not in urls:
                urls[u] = referrer # referrer is original referrer, not redirected url
                referrers[u] = referrer

        elif r.status_code == 200:
            ctype = r["Content-Type"]
            if ";" in ctype:
                ctype = ctype[:ctype.index(";")]

            if ctype == "text/html":
                try:
                    for u in extract_html_urls(r.content):
                        if u not in visited and u not in urls:
                            urls[u] = url
                            referrers[u] = url
                except:
                    log("error extracting HTML urls from %s" % url)
                    log("=============")
                    log(traceback.format_exc())
                    log("=============")
        else:
            tags.append(u"FAIL (from %s)" % referrer)
            errors += 1

        if elapsed.total_seconds() > slow_threshold:
            tags.append("SLOW")

        acc_time = (timestamp - start_time).total_seconds()
        acc_secs = (timestamp - start_time).total_seconds()
        hrs = acc_secs // (60*60)
        min = (acc_secs % (60*60)) // 60
        sec = acc_secs % 60

        if (len(visited) % 100) == 1:
            log("\nElapsed  Visited  Queue   Code  Time  Url  ...  Notes")

        log("%2d:%02d:%02d %7d %6d  %s %.3fs %s %s" % (hrs,min,sec, len(visited), len(urls), r.status_code, elapsed.total_seconds(), url, " ".join(tags)))

logfile.close()
sys.stderr.write("Output written to %s\n\n" % logfile.name)

if errors > 0:
    sys.stderr.write("Found %s errors, grep output for FAIL for details\n" % errors)
    sys.exit(1)
else:
    sys.stderr.write("Found no errors.")