generic extractor (#735)
* Generic extractor, see issue #683 * Fix failed test_names test, no subcategory needed * Prefix directory_fmt with "generic" * Relax regex (would break some urls) * Flake8 compliance * pattern: don't require a scheme This fixes a bug when we force the generic extractor on urls without a scheme (that are allowed by all other extractors). * Fix using g: and r: on urls without http(s) scheme Almost all extractors accept urls without an initial http(s) scheme. Many extractors also allow for generic subdomains in their "pattern" variable; some of them implement this with the regex character class "[^.]+" (everything but a dot). This leads to a problem when the extractor is given a url starting with g: or r: (to force using the generic or recursive extractor) and without the http(s) scheme: e.g. with "r:foobar.tumblr.com" the "r:" is wrongly considered part of the subdomain. This commit fixes the bug, replacing the too generic "[^.]+" with the more specific "[\w-]+" (letters, digits and "-", the only characters allowed in domain names), which is already used by some extractors. * Relax imageurl_pattern_ext: allow relative urls * First round of small suggested changes * Support image urls starting with "//" * self.baseurl: remove trailing slash * Relax regexp (didn't catch some image urls) * Some fixes and cleanup * Fix domain pattern; option to enable extractor Fixed the domain section for "pattern", to pass "test_add" and "test_add_module" tests. Added the "enabled" configuration option (default False) to enable the generic extractor. Using "g(eneric):URL" forces using the extractor.pull/2197/head
parent
4376b39a2b
commit
96fcff182c
@ -0,0 +1,208 @@
|
||||
# -*- coding: utf-8 -*-
|
||||
|
||||
"""Extractor for images in a generic web page."""
|
||||
|
||||
from .common import Extractor, Message
|
||||
from .. import config, text
|
||||
import re
|
||||
import os.path
|
||||
|
||||
|
||||
class GenericExtractor(Extractor):
|
||||
"""Extractor for images in a generic web page."""
|
||||
|
||||
category = "generic"
|
||||
directory_fmt = ("{category}", "{pageurl}")
|
||||
archive_fmt = "{imageurl}"
|
||||
|
||||
# By default, the generic extractor is disabled
|
||||
# and the "g(eneric):" prefix in url is required.
|
||||
# If the extractor is enabled, make the prefix optional
|
||||
pattern = r"(?ix)(?P<generic>g(?:eneric)?:)"
|
||||
if config.get(("extractor", "generic"), "enabled"):
|
||||
pattern += r"?"
|
||||
|
||||
# The generic extractor pattern should match (almost) any valid url
|
||||
# Based on: https://tools.ietf.org/html/rfc3986#appendix-B
|
||||
pattern += r"""
|
||||
(?P<scheme>https?://)? # optional http(s) scheme
|
||||
(?P<domain>[-\w\.]+) # required domain
|
||||
(?P<path>/[^?&#]*)? # optional path
|
||||
(?:\?(?P<query>[^/?#]*))? # optional query
|
||||
(?:\#(?P<fragment>.*))?$ # optional fragment
|
||||
"""
|
||||
|
||||
def __init__(self, match):
|
||||
"""Init."""
|
||||
Extractor.__init__(self, match)
|
||||
|
||||
# Strip the "g(eneric):" prefix
|
||||
# and inform about "forced" or "fallback" mode
|
||||
if match.group('generic'):
|
||||
self.log.info("Forcing use of generic information extractor.")
|
||||
self.url = match.group(0).partition(":")[2]
|
||||
else:
|
||||
self.log.info("Falling back on generic information extractor.")
|
||||
self.url = match.group(0)
|
||||
|
||||
# Make sure we have a scheme, or use https
|
||||
if match.group('scheme'):
|
||||
self.scheme = match.group('scheme')
|
||||
else:
|
||||
self.scheme = 'https://'
|
||||
self.url = self.scheme + self.url
|
||||
|
||||
# Used to resolve relative image urls
|
||||
self.root = self.scheme + match.group('domain')
|
||||
|
||||
def items(self):
|
||||
"""Get page, extract metadata & images, yield them in suitable messages.
|
||||
|
||||
Adapted from common.GalleryExtractor.items()
|
||||
|
||||
"""
|
||||
page = self.request(self.url).text
|
||||
data = self.metadata(page)
|
||||
imgs = self.images(page)
|
||||
|
||||
try:
|
||||
data["count"] = len(imgs)
|
||||
except TypeError:
|
||||
pass
|
||||
images = enumerate(imgs, 1)
|
||||
|
||||
yield Message.Version, 1
|
||||
yield Message.Directory, data
|
||||
|
||||
for data["num"], (url, imgdata) in images:
|
||||
if imgdata:
|
||||
data.update(imgdata)
|
||||
if "extension" not in imgdata:
|
||||
text.nameext_from_url(url, data)
|
||||
else:
|
||||
text.nameext_from_url(url, data)
|
||||
yield Message.Url, url, data
|
||||
|
||||
def metadata(self, page):
|
||||
"""Extract generic webpage metadata, return them in a dict."""
|
||||
data = {}
|
||||
data['pageurl'] = self.url
|
||||
data['title'] = text.extract(page, '<title>', "</title>")[0] or ""
|
||||
data['description'] = text.extract(
|
||||
page, '<meta name="description" content="', '"')[0] or ""
|
||||
data['keywords'] = text.extract(
|
||||
page, '<meta name="keywords" content="', '"')[0] or ""
|
||||
data['language'] = text.extract(
|
||||
page, '<meta name="language" content="', '"')[0] or ""
|
||||
data['name'] = text.extract(
|
||||
page, '<meta itemprop="name" content="', '"')[0] or ""
|
||||
data['copyright'] = text.extract(
|
||||
page, '<meta name="copyright" content="', '"')[0] or ""
|
||||
data['og_site'] = text.extract(
|
||||
page, '<meta property="og:site" content="', '"')[0] or ""
|
||||
data['og_site_name'] = text.extract(
|
||||
page, '<meta property="og:site_name" content="', '"')[0] or ""
|
||||
data['og_title'] = text.extract(
|
||||
page, '<meta property="og:title" content="', '"')[0] or ""
|
||||
data['og_description'] = text.extract(
|
||||
page, '<meta property="og:description" content="', '"')[0] or ""
|
||||
|
||||
data = {k: text.unescape(data[k]) for k in data if data[k] != ""}
|
||||
|
||||
return data
|
||||
|
||||
def images(self, page):
|
||||
"""Extract image urls, return a list of (image url, metadata) tuples.
|
||||
|
||||
The extractor aims at finding as many _likely_ image urls as possible,
|
||||
using two strategies (see below); since these often overlap, any
|
||||
duplicate urls will be removed at the end of the process.
|
||||
|
||||
Note: since we are using re.findall() (see below), it's essential that
|
||||
the following patterns contain 0 or at most 1 capturing group, so that
|
||||
re.findall() return a list of urls (instead of a list of tuples of
|
||||
matching groups). All other groups used in the pattern should be
|
||||
non-capturing (?:...).
|
||||
|
||||
1: Look in src/srcset attributes of img/video/source elements
|
||||
|
||||
See:
|
||||
https://www.w3schools.com/tags/att_src.asp
|
||||
https://www.w3schools.com/tags/att_source_srcset.asp
|
||||
|
||||
We allow both absolute and relative urls here.
|
||||
|
||||
Note that srcset attributes often contain multiple space separated
|
||||
image urls; this pattern matches only the first url; remaining urls
|
||||
will be matched by the "imageurl_pattern_ext" pattern below.
|
||||
"""
|
||||
imageurl_pattern_src = r"""(?ix)
|
||||
<(?:img|video|source)\s.*? # <img>, <video> or <source>
|
||||
src(?:set)?=["']? # src or srcset attributes
|
||||
(?P<URL>[^"'\s>]+) # url
|
||||
"""
|
||||
|
||||
"""
|
||||
2: Look anywhere for urls containing common image/video extensions
|
||||
|
||||
The list of allowed extensions is borrowed from the directlink.py
|
||||
extractor; other could be added, see
|
||||
https://en.wikipedia.org/wiki/List_of_file_formats
|
||||
|
||||
Compared to the "pattern" class variable, here we must exclude also
|
||||
other special characters (space, ", ', >), since we are looking for
|
||||
urls in html tags.
|
||||
"""
|
||||
|
||||
imageurl_pattern_ext = r"""(?ix)
|
||||
(?:[^?&#"'>\s]+) # anything until dot+extension
|
||||
\.(?:jpe?g|jpe|png|gif
|
||||
|web[mp]|mp4|mkv|og[gmv]|opus) # dot + image/video extensions
|
||||
(?:[^"'>\s]*)? # optional query and fragment
|
||||
"""
|
||||
|
||||
imageurls_src = re.findall(imageurl_pattern_src, page)
|
||||
imageurls_ext = re.findall(imageurl_pattern_ext, page)
|
||||
imageurls = imageurls_src + imageurls_ext
|
||||
|
||||
# Resolve relative urls
|
||||
#
|
||||
# Image urls catched so far may be relative, so we must resolve them
|
||||
# by prepending a suitable base url.
|
||||
#
|
||||
# If the page contains a <base> element, use it as base url
|
||||
basematch = re.search(
|
||||
r"(?i)(?:<base\s.*?href=[\"']?)(?P<url>[^\"' >]+)", page)
|
||||
if basematch:
|
||||
self.baseurl = basematch.group('url').rstrip('/')
|
||||
# Otherwise, extract the base url from self.url
|
||||
else:
|
||||
if self.url.endswith("/"):
|
||||
self.baseurl = self.url.rstrip('/')
|
||||
else:
|
||||
self.baseurl = os.path.dirname(self.url)
|
||||
|
||||
# Build the list of absolute image urls
|
||||
absimageurls = []
|
||||
for u in imageurls:
|
||||
# Absolute urls are taken as-is
|
||||
if u.startswith('http'):
|
||||
absimageurls.append(u)
|
||||
# // relative urls are prefixed with current scheme
|
||||
elif u.startswith('//'):
|
||||
absimageurls.append(self.scheme + u.lstrip('/'))
|
||||
# / relative urls are prefixed with current scheme+domain
|
||||
elif u.startswith('/'):
|
||||
absimageurls.append(self.root + u)
|
||||
# other relative urls are prefixed with baseurl
|
||||
else:
|
||||
absimageurls.append(self.baseurl + '/' + u)
|
||||
|
||||
# Remove duplicates
|
||||
absimageurls = set(absimageurls)
|
||||
|
||||
# Create the image metadata dict and add imageurl to it
|
||||
# (image filename and extension are added by items())
|
||||
images = [(u, {'imageurl': u}) for u in absimageurls]
|
||||
|
||||
return images
|
Loading…
Reference in new issue