Backup_Repos/FanFicFare
1
0
Fork 0
mirror of https://github.com/JimmXinu/FanFicFare.git synced 2025-12-06 08:52:55 +01:00
Code Issues Projects Releases Wiki Activity

Replace &#XXX; entities in file names, allow BR tags--some stories use multiple BRs are section breaks.

Browse source
This commit is contained in:
retiefjimm 2010-09-28 19:26:19 -05:00
parent 84fea5b896
commit 3be9b3331e
1 changed files with 8 additions and 5 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

13
output.py
View file

@ -25,6 +25,7 @@ import zipdir
import html_constants
from constants import *
import html2text
@ -59,7 +60,7 @@ class HTMLWriter(FanficWriter):
def __init__(self, base, name, author, inmemory=False, compress=False):
self.basePath = base
self.name = name.replace(" ", "_").replace(":","_")
self.name = re.sub('&\#[0-9]+;', '_', name.replace(" ", "_").replace(":","_"))
self.storyTitle = name
self.fileName = self.basePath + '/' + self.name + '.html'
self.authorName = author
@ -133,7 +134,7 @@ class EPubFanficWriter(FanficWriter):
def __init__(self, base, name, author, inmemory=False, compress=True):
self.basePath = base
self.name = name.replace(" ", "_").replace(":","_")
self.name = re.sub('&\#[0-9]+;', '_', name.replace(" ", "_").replace(":","_"))
self.storyTitle = name
self.directory = self.basePath + '/' + self.name
self.inmemory = inmemory
@ -149,6 +150,7 @@ class EPubFanficWriter(FanficWriter):
self.writeToFile = False
if not self.inmemory:
if os.path.exists(self.directory):
shutil.rmtree(self.directory)
@ -194,8 +196,9 @@ class EPubFanficWriter(FanficWriter):
text = self._removeEntities(text)
# BeautifulStoneSoup doesn't have any selfClosingTags by default.
# hr needs to be if it's going to work.
self.soup = bs.BeautifulStoneSoup(text.decode('utf-8'), selfClosingTags=('hr'))
# hr & br needs to be if they're going to work.
# Some stories do use multiple br tags as their section breaks...
self.soup = bs.BeautifulStoneSoup(text.decode('utf-8'), selfClosingTags=('br','hr'))
allTags = self.soup.findAll(recursive=True)
for t in allTags:
@ -208,7 +211,7 @@ class EPubFanficWriter(FanficWriter):
if p.string != None and (len(p.string.strip()) == 0 or p.string.strip() == ' ' ) :
p.extract()
allBrs = self.soup.findAll(recursive=True, name = ["br", 'div'])
allBrs = self.soup.findAll(recursive=True, name = ['div'])
for br in allBrs:
if (br.string != None and len(br.string.strip()) != 0) or (br.contents != None):
br.name = 'p'