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

Heuristic processing: Enable individual actions by default

Browse Source
This commit is contained in:
Kovid Goyal 2011-01-19 10:54:26 -07:00
parent 1ad9b6cbfd
commit 1f9007b9a9
2 changed files with 23 additions and 18 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

25
src/calibre/ebooks/conversion/cli.py
View File

@ -42,6 +42,12 @@ option.
For full documentation of the conversion system see
''') + 'http://calibre-ebook.com/user_manual/conversion.html'
HEURISTIC_OPTIONS = ['markup_chapter_headings',
'italicize_common_cases', 'fix_indents',
'html_unwrap_factor', 'unwrap_lines',
'delete_blank_paragraphs', 'format_scene_breaks',
'dehyphenate', 'renumber_headings']
def print_help(parser, log):
help = parser.format_help().encode(preferred_encoding, 'replace')
log(help)
@ -83,6 +89,8 @@ def option_recommendation_to_cli_option(add_option, rec):
if opt.long_switch == 'verbose':
attrs['action'] = 'count'
attrs.pop('type', '')
if opt.name in HEURISTIC_OPTIONS and rec.recommended_value is True:
switches = ['--disable-'+opt.long_switch]
add_option(Option(*switches, **attrs))
def add_input_output_options(parser, plumber):
@ -129,18 +137,15 @@ def add_pipeline_options(parser, plumber):
'asciiize',
]
),
'HEURISTIC PROCESSING' : (
_('Modify the document text and structure using common patterns.'),
[
'enable_heuristics', 'markup_chapter_headings',
'italicize_common_cases', 'fix_indents',
'html_unwrap_factor', 'unwrap_lines',
'delete_blank_paragraphs', 'format_scene_breaks',
'dehyphenate', 'renumber_headings',
]
_('Modify the document text and structure using common'
' patterns. Disabled by default. Use %s to enable. '
' Individual actions can be diable with the %s options.')
% ('--enable-heuristics', '--disable-*'),
['enable_heuristics'] + HEURISTIC_OPTIONS
),
'SEARCH AND REPLACE' : (
_('Modify the document text and structure using user defined patterns.'),
[

16
src/calibre/ebooks/conversion/plumber.py
View File

@ -490,19 +490,19 @@ OptionRecommendation(name='enable_heuristics',
'heuristic processing to take place.')),
OptionRecommendation(name='markup_chapter_headings',
recommended_value=False, level=OptionRecommendation.LOW,
recommended_value=True, level=OptionRecommendation.LOW,
help=_('Detect unformatted chapter headings and sub headings. Change '
'them to h2 and h3 tags. This setting will not create a TOC, '
'but can be used in conjunction with structure detection to create '
'one.')),
OptionRecommendation(name='italicize_common_cases',
recommended_value=False, level=OptionRecommendation.LOW,
recommended_value=True, level=OptionRecommendation.LOW,
help=_('Look for common words and patterns that denote '
'italics and italicize them.')),
OptionRecommendation(name='fix_indents',
recommended_value=False, level=OptionRecommendation.LOW,
recommended_value=True, level=OptionRecommendation.LOW,
help=_('Turn indentation created from multiple non-breaking space entities '
'into CSS indents.')),
@ -515,28 +515,28 @@ OptionRecommendation(name='html_unwrap_factor',
'be reduced')),
OptionRecommendation(name='unwrap_lines',
recommended_value=False, level=OptionRecommendation.LOW,
recommended_value=True, level=OptionRecommendation.LOW,
help=_('Unwrap lines using punctuation and other formatting clues.')),
OptionRecommendation(name='delete_blank_paragraphs',
recommended_value=False, level=OptionRecommendation.LOW,
recommended_value=True, level=OptionRecommendation.LOW,
help=_('Remove empty paragraphs from the document when they exist between '
'every other paragraph')),
OptionRecommendation(name='format_scene_breaks',
recommended_value=False, level=OptionRecommendation.LOW,
recommended_value=True, level=OptionRecommendation.LOW,
help=_('Left aligned scene break markers are center aligned. '
'Replace soft scene breaks that use multiple blank lines with'
'horizontal rules.')),
OptionRecommendation(name='dehyphenate',
recommended_value=False, level=OptionRecommendation.LOW,
recommended_value=True, level=OptionRecommendation.LOW,
help=_('Analyze hyphenated words throughout the document. The '
'document itself is used as a dictionary to determine whether hyphens '
'should be retained or removed.')),
OptionRecommendation(name='renumber_headings',
recommended_value=False, level=OptionRecommendation.LOW,
recommended_value=True, level=OptionRecommendation.LOW,
help=_('Looks for occurrences of sequential <h1> or <h2> tags. '
'The tags are renumbered to prevent splitting in the middle '
'of chapter headings.')),