diff options
Diffstat (limited to 'gallery_dl/extractor/4archive.py')
| -rw-r--r-- | gallery_dl/extractor/4archive.py | 17 |
1 files changed, 7 insertions, 10 deletions
diff --git a/gallery_dl/extractor/4archive.py b/gallery_dl/extractor/4archive.py index d198369..c9be2a4 100644 --- a/gallery_dl/extractor/4archive.py +++ b/gallery_dl/extractor/4archive.py @@ -27,8 +27,7 @@ class _4archiveThreadExtractor(Extractor): self.board, self.thread = match.groups() def items(self): - url = "{}/board/{}/thread/{}".format( - self.root, self.board, self.thread) + url = f"{self.root}/board/{self.board}/thread/{self.thread}" page = self.request(url).text data = self.metadata(page) posts = self.posts(page) @@ -58,15 +57,14 @@ class _4archiveThreadExtractor(Extractor): for post in page.split('class="postContainer')[1:] ] - @staticmethod - def parse(post): + def parse(self, post): extr = text.extract_from(post) data = { "name": extr('class="name">', "</span>"), "date": text.parse_datetime( extr('class="dateTime postNum" >', "<").strip(), "%Y-%m-%d %H:%M:%S"), - "no" : text.parse_int(extr('href="#p', '"')), + "no" : text.parse_int(extr(">Post No.", "<")), } if 'class="file"' in post: extr('class="fileText"', ">File: <a") @@ -94,18 +92,17 @@ class _4archiveBoardExtractor(Extractor): def __init__(self, match): Extractor.__init__(self, match) - self.board = match.group(1) - self.num = text.parse_int(match.group(2), 1) + self.board = match[1] + self.num = text.parse_int(match[2], 1) def items(self): data = {"_extractor": _4archiveThreadExtractor} while True: - url = "{}/board/{}/{}".format(self.root, self.board, self.num) + url = f"{self.root}/board/{self.board}/{self.num}" page = self.request(url).text if 'class="thread"' not in page: return for thread in text.extract_iter(page, 'class="thread" id="t', '"'): - url = "{}/board/{}/thread/{}".format( - self.root, self.board, thread) + url = f"{self.root}/board/{self.board}/thread/{thread}" yield Message.Queue, url, data self.num += 1 |