Cutlery/calibre
1
0
Fork 0
mirror of https://github.com/kovidgoyal/calibre.git synced 2025-07-09 03:04:10 -04:00
Code Issues Packages Projects Releases Wiki Activity

CHM Input: Store extracted files in the input/ sub dir for easy debugging when --debug-pipeline is specified

Browse Source
This commit is contained in:
Kovid Goyal 2011-04-27 14:19:25 -06:00
parent 631afb7eac
commit 8887b7f057
3 changed files with 16 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
src/calibre/ebooks/chm/input.py
View File

@ -19,12 +19,12 @@ class CHMInput(InputFormatPlugin):
description = 'Convert CHM files to OEB' description = 'Convert CHM files to OEB'
file_types = set(['chm']) file_types = set(['chm'])
def _chmtohtml(self, output_dir, chm_path, no_images, log): def _chmtohtml(self, output_dir, chm_path, no_images, log, debug_dump=False):
from calibre.ebooks.chm.reader import CHMReader from calibre.ebooks.chm.reader import CHMReader
log.debug('Opening CHM file') log.debug('Opening CHM file')
rdr = CHMReader(chm_path, log, self.opts) rdr = CHMReader(chm_path, log, self.opts)
log.debug('Extracting CHM to %s' % output_dir) log.debug('Extracting CHM to %s' % output_dir)
rdr.extract_content(output_dir) rdr.extract_content(output_dir, debug_dump=debug_dump)
self._chm_reader = rdr self._chm_reader = rdr
return rdr.hhc_path return rdr.hhc_path
@ -47,7 +47,12 @@ class CHMInput(InputFormatPlugin):
stream.close() stream.close()
log.debug('tdir=%s' % tdir) log.debug('tdir=%s' % tdir)
log.debug('stream.name=%s' % stream.name) log.debug('stream.name=%s' % stream.name)
mainname = self._chmtohtml(tdir, chm_name, no_images, log) debug_dump = False
odi = options.debug_pipeline
if odi:
debug_dump = os.path.join(odi, 'input')
mainname = self._chmtohtml(tdir, chm_name, no_images, log,
debug_dump=debug_dump)
mainpath = os.path.join(tdir, mainname) mainpath = os.path.join(tdir, mainname)
metadata = get_metadata_from_reader(self._chm_reader) metadata = get_metadata_from_reader(self._chm_reader)
@ -56,7 +61,6 @@ class CHMInput(InputFormatPlugin):
#from calibre import ipython #from calibre import ipython
#ipython() #ipython()
odi = options.debug_pipeline
options.debug_pipeline = None options.debug_pipeline = None
options.input_encoding = 'utf-8' options.input_encoding = 'utf-8'
# try a custom conversion: # try a custom conversion:

9
src/calibre/ebooks/chm/reader.py
View File

@ -97,7 +97,7 @@ class CHMReader(CHMFile):
raise CHMError("'%s' is zero bytes in length!"%(path,)) raise CHMError("'%s' is zero bytes in length!"%(path,))
return data return data
def ExtractFiles(self, output_dir=os.getcwdu()): def ExtractFiles(self, output_dir=os.getcwdu(), debug_dump=False):
html_files = set([]) html_files = set([])
for path in self.Contents(): for path in self.Contents():
lpath = os.path.join(output_dir, path) lpath = os.path.join(output_dir, path)
@ -123,6 +123,9 @@ class CHMReader(CHMFile):
self.log.warn('%r filename too long, skipping'%path) self.log.warn('%r filename too long, skipping'%path)
continue continue
raise raise
if debug_dump:
import shutil
shutil.copytree(output_dir, os.path.join(debug_dump, 'debug_dump'))
for lpath in html_files: for lpath in html_files:
with open(lpath, 'r+b') as f: with open(lpath, 'r+b') as f:
data = f.read() data = f.read()
@ -249,8 +252,8 @@ class CHMReader(CHMFile):
if not os.path.isdir(dir): if not os.path.isdir(dir):
os.makedirs(dir) os.makedirs(dir)
def extract_content(self, output_dir=os.getcwdu()): def extract_content(self, output_dir=os.getcwdu(), debug_dump=False):
self.ExtractFiles(output_dir=output_dir) self.ExtractFiles(output_dir=output_dir, debug_dump=debug_dump)

3
src/calibre/ebooks/conversion/plumber.py
View File

@ -854,7 +854,8 @@ OptionRecommendation(name='sr3_replace',
if isinstance(ret, basestring): if isinstance(ret, basestring):
shutil.copytree(output_dir, out_dir) shutil.copytree(output_dir, out_dir)
else: else:
os.makedirs(out_dir) if not os.path.exists(out_dir):
os.makedirs(out_dir)
self.dump_oeb(ret, out_dir) self.dump_oeb(ret, out_dir)
if self.input_fmt == 'recipe': if self.input_fmt == 'recipe':
zf = ZipFile(os.path.join(self.opts.debug_pipeline, zf = ZipFile(os.path.join(self.opts.debug_pipeline,