def postprocess_book(self, oeb, opts, log):
from ebook_converter.ebooks.oeb.base import XPath, XHTML
for item in oeb.spine:
root = item.data
if not hasattr(root, 'xpath'):
continue
for bad in ('metadata', 'guide'):
metadata = XPath('//h:'+bad)(root)
if metadata:
for x in metadata:
x.getparent().remove(x)
body = XPath('//h:body')(root)
if body:
body = body[0]
if len(body) == 1 and body[0].tag == XHTML('pre'):
pre = body[0]
from ebook_converter.ebooks.txt.processor import \
convert_basic, separate_paragraphs_single_line
from ebook_converter.ebooks.chardet import xml_to_unicode
self.log('LIT file with all text in singe <pre> tag '
'detected')
html = separate_paragraphs_single_line(pre.text)
html = convert_basic(html).replace('<html>',
'<html xmlns="%s">' %
const.XHTML_NS)
html = xml_to_unicode(html, strip_encoding_pats=True,
resolve_entities=True)[0]
if opts.smarten_punctuation:
# SmartyPants skips text inside <pre> tags
from ebook_converter.ebooks.conversion import \
preprocess
html = preprocess.smarten_punctuation(html, self.log)
root = etree.fromstring(html)
body = XPath('//h:body')(root)
pre.tag = XHTML('div')
pre.text = ''
for elem in body:
ne = copy.deepcopy(elem)
pre.append(ne)
def txt_process(self, match):
from ebook_converter.ebooks.txt.processor import convert_basic, separate_paragraphs_single_line
content = match.group('text')
content = separate_paragraphs_single_line(content)
content = convert_basic(content, epub_split_size_kb=0)
return content
def convert(self, stream, options, file_ext, log, accelerators):
from ebook_converter.ebooks.conversion.preprocess import DocAnalysis, Dehyphenator
from ebook_converter.ebooks.chardet import detect
from ebook_converter.utils.zipfile import ZipFile
from ebook_converter.ebooks.txt.processor import (
convert_basic, convert_markdown_with_metadata,
separate_paragraphs_single_line,
separate_paragraphs_print_formatted, preserve_spaces,
detect_paragraph_type, detect_formatting_type,
normalize_line_endings, convert_textile, remove_indents,
block_to_single_line, separate_hard_scene_breaks)
self.log = log
txt = b''
log.debug('Reading text from file...')
length = 0
base_dir = self.output_dir = os.getcwd()
# Extract content from zip archive.
if file_ext == 'txtz':
zf = ZipFile(stream)
zf.extractall('.')
for root, _, fnames in os.walk('.'):
for x in fnames:
x = os.path.join(root, x)
if os.path.splitext(x)[1].lower() in ('.txt', '.text'):
with open(x, 'rb') as tf:
txt += tf.read() + b'\n\n'
else:
if getattr(stream, 'name', None):
base_dir = os.path.dirname(stream.name)
txt = stream.read()
if file_ext in {'md', 'textile', 'markdown'}:
options.formatting_type = {
'md': 'markdown'
}.get(file_ext, file_ext)
log.info(
'File extension indicates particular formatting. '
'Forcing formatting type to: %s', options.formatting_type)
options.paragraph_type = 'off'
# Get the encoding of the document.
if options.input_encoding:
ienc = options.input_encoding
log.debug('Using user specified input encoding of %s', ienc)
else:
det_encoding = detect(txt[:4096])
det_encoding, confidence = det_encoding['encoding'], det_encoding[
'confidence']
if det_encoding and det_encoding.lower().replace(
'_',
'-').strip() in ('gb2312', 'chinese', 'csiso58gb231280',
'euc-cn', 'euccn', 'eucgb2312-cn',
'gb2312-1980', 'gb2312-80', 'iso-ir-58'):
# Microsoft Word exports to HTML with encoding incorrectly set to
# gb2312 instead of gbk. gbk is a superset of gb2312, anyway.
det_encoding = 'gbk'
ienc = det_encoding
log.debug(
'Detected input encoding as %s with a confidence of '
'%s%%', ienc, confidence * 100)
if not ienc:
ienc = 'utf-8'
log.debug(
'No input encoding specified and could not auto detect '
'using %s', ienc)
# Remove BOM from start of txt as its presence can confuse markdown
import codecs
for bom in (codecs.BOM_UTF16_LE, codecs.BOM_UTF16_BE, codecs.BOM_UTF8,
codecs.BOM_UTF32_LE, codecs.BOM_UTF32_BE):
if txt.startswith(bom):
txt = txt[len(bom):]
break
txt = txt.decode(ienc, 'replace')
# Replace entities
txt = entities.ENT_PAT.sub(entities.xml_entity_to_unicode, txt)
# Normalize line endings
txt = normalize_line_endings(txt)
# Determine the paragraph type of the document.
if options.paragraph_type == 'auto':
options.paragraph_type = detect_paragraph_type(txt)
if options.paragraph_type == 'unknown':
log.debug('Could not reliably determine paragraph type using '
'block')
options.paragraph_type = 'block'
else:
log.debug('Auto detected paragraph type as %s',
options.paragraph_type)
# Detect formatting
if options.formatting_type == 'auto':
options.formatting_type = detect_formatting_type(txt)
log.debug('Auto detected formatting as %s',
options.formatting_type)
if options.formatting_type == 'heuristic':
setattr(options, 'enable_heuristics', True)
setattr(options, 'unwrap_lines', False)
setattr(options, 'smarten_punctuation', True)
# Reformat paragraphs to block formatting based on the detected type.
# We don't check for block because the processor assumes block.
# single and print at transformed to block for processing.
if options.paragraph_type == 'single':
txt = separate_paragraphs_single_line(txt)
elif options.paragraph_type == 'print':
txt = separate_hard_scene_breaks(txt)
txt = separate_paragraphs_print_formatted(txt)
txt = block_to_single_line(txt)
elif options.paragraph_type == 'unformatted':
from ebook_converter.ebooks.conversion.utils import HeuristicProcessor
# unwrap lines based on punctuation
docanalysis = DocAnalysis('txt', txt)
length = docanalysis.line_length(.5)
preprocessor = HeuristicProcessor(options,
log=getattr(self, 'log', None))
txt = preprocessor.punctuation_unwrap(length, txt, 'txt')
txt = separate_paragraphs_single_line(txt)
elif options.paragraph_type == 'block':
txt = separate_hard_scene_breaks(txt)
txt = block_to_single_line(txt)
if getattr(options, 'enable_heuristics', False) and getattr(
options, 'dehyphenate', False):
docanalysis = DocAnalysis('txt', txt)
if not length:
length = docanalysis.line_length(.5)
dehyphenator = Dehyphenator(options.verbose, log=self.log)
txt = dehyphenator(txt, 'txt', length)
# User requested transformation on the text.
if options.txt_in_remove_indents:
txt = remove_indents(txt)
# Preserve spaces will replace multiple spaces to a space
# followed by the entity.
if options.preserve_spaces:
txt = preserve_spaces(txt)
# Process the text using the appropriate text processor.
self.shifted_files = []
try:
html = ''
input_mi = None
if options.formatting_type == 'markdown':
log.debug('Running text through markdown conversion...')
try:
input_mi, html = convert_markdown_with_metadata(
txt,
extensions=[
x.strip()
for x in options.markdown_extensions.split(',')
if x.strip()
])
except RuntimeError:
raise ValueError(
'This txt file has malformed markup, it cannot be'
' converted by calibre. See https://daringfireball.net/projects/markdown/syntax'
)
html = self.fix_resources(html, base_dir)
elif options.formatting_type == 'textile':
log.debug('Running text through textile conversion...')
html = convert_textile(txt)
html = self.fix_resources(html, base_dir)
else:
log.debug('Running text through basic conversion...')
flow_size = getattr(options, 'flow_size', 0)
html = convert_basic(txt, epub_split_size_kb=flow_size)
# Run the HTMLized text through the html processing plugin.
from ebook_converter.customize.ui import plugin_for_input_format
html_input = plugin_for_input_format('html')
for opt in html_input.options:
setattr(options, opt.option.name, opt.recommended_value)
options.input_encoding = 'utf-8'
htmlfile = self.shift_file('index.html', html.encode('utf-8'))
odi = options.debug_pipeline
options.debug_pipeline = None
# Generate oeb from html conversion.
oeb = html_input.convert(open(htmlfile, 'rb'), options, 'html',
log, {})
options.debug_pipeline = odi
finally:
for x in self.shifted_files:
os.remove(x)
# Set metadata from file.
if input_mi is None:
from ebook_converter.customize.ui import get_file_type_metadata
input_mi = get_file_type_metadata(stream, file_ext)
from ebook_converter.ebooks.oeb.transforms.metadata import meta_info_to_oeb_metadata
meta_info_to_oeb_metadata(input_mi, oeb.metadata, log)
self.html_postprocess_title = input_mi.title
return oeb