Wikt
A dictionary client which displays entries from
the Wiktionary project. The
entries are stored locally, so they can be accessed very quickly and
no internet connection is needed.
Wiktionary aims to include all words and idioms of all human
languages, including etymology, pronounciation (IPA and SAMPA
notation), thesaurus, and translations. For each word, all its senses
are explained, supported by usage examples and citations. The
explanation is followed by synonyms, antonyms, and related
terms. Then, translations to various languages are provided for each
sense. Many entries also contain recordings with pronounciation by
native speaker, and illustrative pictures.
Wikt comes in two flavours: a graphical user interface (GUI) and a
command line interface (CLI). Both variants allow searching for a word
or phrase instantly. GUI also displays similar entries (with
diacritics etc.)
Wikt is free software; you can redistribute it and/or modify it
under the terms of the GNU General Public License as published by the
Free Software Foundation; either version 3 of the License, or (at your
option) any later version.
Obtaining
The source code of released versions can be downloaded from
ftp://ftp.klic.name/wikt.
The source code repository with the latest development code and
history is publicly available via Git.
To get the source code of the development version:
$ git clone git://klic.name/wikt.git
To browse the source code online,
visit gitweb interface for
Wikt.
Known bugs
- Template:PAGESINCATEGORY not implemented (used in Category:!Kung language).
- Expression parser does not support expression “0 or 0” (used in Category:!Kung language).
Preparing dictionary for Wikt
1. Download XML dump of Wiktionary. Wiktionary dumps are available
on Wikimedia
Downloads website in a number of variants. Data usable for Wikt
are present in Articles, templates, image descriptions, and primary
meta-pages variant. The XML dump should be unpacked after
download:
$ wget http://download.wikimedia.org/enwiktionary/20111008/enwiktionary-20111008-pages-articles.xml.bz2
$ bunzip2 enwiktionary-20111008-pages-articles.xml.bz2
$ mv enwiktionary-20111008-pages-articles.xml wikt-dict-en.xml
2. Convert the XML file to PREP file:
$ wikt-dict xmltoprep wikt-dict-en.xml wikt-dict-en.prep wikt-dict-en.errata
3. Convert the PREP file to MID file:
$ wikt-dict preptomid wikt-dict-en.prep wikt-dict-en.mid
4. Download media linked from MID file to a directory:
$ wikt-dict downloadmedia wikt-dict-en.mid wikt-dict-en.media
5. Resize images in the directory to reasonable sizes. Usually,
pictures uploaded to Wikipedia Commons are unnecessarily large.
$ wikt-dict resizeimages wikt-dict-en.mid wikt-dict-en.media
6. Pack media from directory to MED file:
$ wikt-dict packmedia wikt-dict-en.media wikt-dict-en.med
7. Convert the MID file to DICT file:
$ wikt-dict midtodict wikt-dict-en.mid wikt-dict-en.eim wikt-dict-en.dict
News
Wikt 0.20 was released on 27 October 2011
- dictionary: CheckTrans sections are not displayed in entries.
- libwikt: Support for more languages added.
- wikt-gui:
- Wikt was renamed to wikt-gui.
- Fixed invalid contents generation from headers. Header
Abbreviation in entry
[[&c.]] is a link, and
that was not handled properly. Affects Contents panel.
- Find dialog (Ctrl+F) works again.
- Improved focus shifting and handling.
- Translation section visibility (collapse status) depends on
user's settings (dialog Options).
- Added application icons. Icons from Tango project were used,
SVG format.
- Error page is displayed when user enters a word to the
lookup panel and the word is not found.
- wikt-dict:
- All command line tools such as convdict12, convdict23 were
merged to single tool wikt-dict.
- convdict12:
- Bugfix: No longer removes the content between
<noinclude> open and close tags.
- convdict23:
- Bugfix in evaluating templates
uc: ucfirst: lc:
lcfist:, fixes invalid trimming of result. Affects
entry [[acaeciera]].
- Finished implementation of expression parser: now
the
#expr supports also numbers as .123e4,
functions exp, ln, abs, trunc, floor, ceil, sin, cos, tan,
asin, acos, atan, logical not, and, or, inequality
operators <> and !=.
- Properly handles
<noinclude>
tags. Removes the tagged content in templates during template
evaluation, and removes only the tags in all other cases.
- convdict34:
- Generates form-of/gloss entries statistics.
- Initial implementation of Category listings.
- Initial implementation of
<ref> and
<references/> tag handling. The change
affects entries [[abeam]],
[[aberrant]], [[absolutely]],
[[paraconsistent]].
Wikt 0.19.0 was released on 24 September 2009
- Source code was heavily refactored and simplified, which results
in significantly smaller source code line count.
- Some parts of Wikt application were moved to separate CLI
programs. Dynamic shared library libwikt was created.
- Memory usage of dictionary processing tools was
decreased. Several memory leaks were discovered and removed.
- Unit tests were moved to separate applications outside of main
source code.
Bugfixes
- Tags <includeonly>, <onlyinclude>, and
<noinclude> are detected using regular expressions instead of
fixed strings, eg. old <includeonly>, new:
<includeonly\s*>. Affects entry "apple".
- Template evaluation fixed: {{Template:Appendix:Contents}} text
appeared in the entry [[Category:Appendices]]. Now it is handled as
{{Appendix:Contents}} in convdict23 pass because Appendix is a valid
namespace.
New Features
- Support of 2009-07-11 version of Wiktionary data.
- Application can be installed in any subdirectory (/opt, /usr, ~)
of user's computer.
- User interface:
- The lookup panel receives focus when any appropriate key is
pressed in the application. This means that when you write a
word using keyboard, the word is written to the lookup panel no
matter which UI control was focused.
- Whole application is localizable. Czech localization is
available.
- Statistics language table is sortable (Javascript).
- Library:
- The parser is heavily optimized and major bottlenecks
reported by valgrind are removed.
- Support for dictionary data files larger than 2 GB: changed
32-bit signed integer file pointers to 64-bit signed integer
file pointers. This feature changes the dictionary format.
- Support for more languages added.
- Support of remaining namespaces added: Concordance(Talk),
CitationsTalk, Transwiki(Talk), WT(Talk).
- Tools:
- convdict23: Template evaluation cache results in better
performance.
- convdict23: Implementation of Template:formatnum. Affects
[[Category:!Kung language]].
Wikt 0.18.0 was released on 26 July 2009
Bug Fixes
- Fixed template evaluation of
template
{{given_name}} used in
entry Cheyenne.
- Nonexisting categories are removed from the text of each
entry. Affects entry
[[car]].
- Links
[[help]], [[file]] associated
with each audio link are removed from the dictionary.
New Features
- Image resizing process for media archive. It diminishes the size
of media file from 4 GB to 660 MB.
- Debian packaging files that allow to create Ubuntu Jaunty
installable binary package.
- Searching a whole sentence is supported and allowed.
- Errata files can modify the dictionary contents. Wikisaurus
entries improved this way.
- Better title page.
- Statistics page.
- Pages with dictionary licenses.
- Redirect support.
- Implemented
HIDDENCAT category removal.
Wikt 0.17.0 was released on 28 May 2009
Bug Fixes
- Images in thumbnail style contain icon which now links to the
image on wiktionary (external link). Affects a picture in
the
dog entry for example.
- Interwiki links are extracted from whole document, also from
inside paragraphs etc. Removed the "interwiki paragraph"
concept. Affects all entries.
- Parsing of table headers fixed. Affects a Danish inflection table
in the
bud entry.
New Features
- External links parsed properly, and open in external browser
window when activated.
- Links to wikipedia such as
[[w:IPA]] are converted to
external links such as [http://en.wiktionary.org/wiki/IPA
IPA]. Clicking to the external link opens external
browser.
- Gallery extension is properly handled and displayed. Affects
entry
car.
- User hides translations to languages he is not interested in.
- Table captions are now parsed and displayed correctly. Affects
water entry.
Template:SERVER
implemented. Affects Category:English nouns.Template:localurl
implemented. Affects Category:English nouns.- New languages added and language support reworked.
Template:#ifexpr implemented. Affects 啊
entry.Template:#expr implemented support for
multiplication, division, parenthesis, round, pow, mod, and
comparison.
Wikt 0.16.0 was released on 18 May 2009
Bug Fixes
- Find buttons disabled when the edit box has no content. Also
Find Next menu item works now.
- Subwindows are no longer blinking on application start.
- Fixed a write order/sorting bug in Format3Writer, Format4Writer,
MediaWriter. It caused reading problems in media file.
- Empty list items are ignored when parsing wiki text (class
ListParser). Affects
catalog entry.
- Paragraph boundary detection takes links (characters inside
links) into account. It results in proper parsing
of
stroke#Etymology 1.
- Parser handles links embedded in links; this is used in
images. Fixes the parsing of
A#Translingual->Etymology
1, dog.
New Features
- Images in entries are displayed.
- Pronunciation audio in entries is played.
- Better layout of Find dialog.
- Added Low Saxon and Lojban to interwiki support code. All
missing localized interwiki language names were added.
- Support of underline HTML tag, which is used in
Appendix:English nouns.
Wikt 0.15.0 was released on 13 May 2009
This is the first release that allows to create the media pack with
pronounciation audio files and images. The media pack occupies 3.8
gigabytes.
Many parser bugs were fixed in this release, and there are many new
user-facing changes.
Bug Fixes:
- Fixed the "stadiums" entry, where the error was caused by bad
interpretation of "Template:plural of", which contained a link inside
a link as a parameter in template call. Template argument parser
improperly parsed [[[[link a]] link b]].
- Paragraph parser now detects beginning of a table and closes the
paragraph (entry "carr#Irish").
- Table cell parser now recognizes links with parameters on first
line of a cell (entry "carr#Irish", "Appendix:Variations of
"a"").
- Category processing now handles various links well. Entry "myslet"
does not list "Category:Czech verbs" twice. The entry contains links
[[Category:Czech verbs]] and [[Category: Czech verbs]].
- Fixed parsing of lists: when a list item line is followed by a
line that (after any number of whitespaces) contains a category link,
that line with category link is considered as a part of list item (for
example see definition list in "stadia"). This is not logical, but
let's follow the Way of MediaWiki.
New Features:
- New "Media Packer" tool packs all media files to a single
file.
- Optional visibility of sections such as Etymology, Translations,
Inflections, and Anagrams.
- Cherokee language supported in interwiki. Added another seven
localized names for languages to interwiki support.
- Copy text to clipboard, cut, paste works with Ctrl+C, Ctrl+X,
Ctrl+V.
- Page search (Ctrl+F). Design took from Firefox; it's the best way
of text searching I know.
Wikt 0.14.0 was released on 8 May 2009
Bug Fixes:
- External links in wiki syntax are no longer converted to A
tags.
- A crash in Format3Reader caused by looking for '￥'
fixed. Analogous error in Format4Reader fixed too.
- Allowed nested tables (entry "-aseinen").
- Many methods and classes were documented by comments in source
code.
New Features:
- Links to missing entries are removed in format4 (only the link
text remains).
- Added "Appendix" to supported namespaces.
- Added interwiki name for Macedonian, "Македонски".
- Conversion from F3 to F4 is faster, it takes minutes instead of
hours.
- All format conversion dialogs remember file paths between
application runs.
- Status bar displays links.
- Media downloader downloads all images and sounds used in a
dictionary to local filesystem.
Wikt 0.13.0 was released on 2 May 2009
This release brings no visible features or fixes, and thus it is
still not ready to be used by common users. However, many new lines of
code were written as a pedestal for upcoming features.
- Link node reworked to support new features (entry existence
checking, conversion of outgoing links to external form)
- Entry namespaces, languages, Wikimedia projects handled explicitly
in code.
Wikt 0.12.0 was released on 24 April 2009
Bug Fixes:
DIV tags do not wrap the first embedded paragraph
in P tag.- Table headers and table cells do not wrap the first embedded
paragraph in
P tag.
New Features:
- New compressed dictionary format (Format4, .ei4). Dictionary
size went from 1,1GB to 427 MB.
- Support for inline HTML tags
del, em,
ins, strike, strong, var,
code.
- Support for block HTML
tags
p, h1-h6,
blockquote.
- Main window title contains displayed word.
Wikt 0.11.0 was released on 22 April 2009
Bug Fixes:
- Fixed parsing of bold and italic text. Added 7 tests that cover
basic cases. Corrects handling of example in
entry
A.
- Wikt does not crash when user clicks on a link to missing
entry.
New Features:
- Added
COPYING file with full license.
- Wiki parser handles
CENTER tags. This tag is used
in entry cognoscere. DivHtmlParser is
changed to BlockHtmlParser because
of CENTERs.
- Wiki parser handles
FONT tags. This tag is used
in a#Min Nan.
Wikt 0.10.0 was released on 21 April 2009
This is the initial release of source code.
The project has been started in 2008-11. The first larger iteration
was implemented in gtkmm, but all subsequent attempts to embed Gecko
or WebKit into GTK failed. It was simply too much work to do. At the
end the whole library was rewritten to Qt.
Wishlist
- Add screenshot to this page
- Package wikt-dict-en
- Fix crash in dictionary preparation
- Process the old wishlist:
- Format3ToFormat4 pass will generate a list of appendices and
other lists as necessary.
- Categories store list of entries pointing to them.
- User chooses whether the translation section is collapsed or
not.
- Error pages
- Icons in the application's menu bar.
- convdict23 add commandline parameters to report Expr errors and
to report nonexisting templates.
- Support for ref and references
tags. Entry
-ent
Wiktionary
Wiktionary is a collaborative project to produce a free-content
multilingual dictionary. The project is doing well as the dictionary
is very usable for some use-cases. For Czech language it will be the
best choice soon, as Czech nonfree dictionaries are lacking detail,
and their software do not run on common free operating systems.
It makes sense to collaborate on a dictionary. Languages are
changing slowly, so the dictionary is not losing its value when
developed. It's easy to contribute, and contributions add up. Just
like with software, freedom is very valuable in dictionary area.
Contributing
- Wiktionary:Entry layout explained
- Guidelines for creating and maintaining entries
- Wiktionary:About Czech
- Guidelines both for creating Czech entries on English Wiktionary
as well as for adding Czech translations to English words
- Category:Context labels
- Wikislovník:Výslovnost/čeština
- IPA for Czech entries.
- Wikislovník:Užitečné informace
- REJZEK, Jiří. Český etymologický slovník. Leda, 2001. ISBN 80-85927-85-3.
- Nový akademický slovník cizích slov. Academia 2005, 2006.
Czech contributors
- Dan
[contributions]
- Mirek
[contributions]
- Duncan
[contributions]
- Lenka
[contributions]
- -xfi-
[contributions]
Todo list
Add all missing entries from Index:Czech.
Treat all entries from Tbot entries (Czech).
Improve the handling of Czech adjectives: the gender issue {{g/m/f}}, fix [[Wiktionary:About Czech]], declension template.
Fix some syntax errors from User:Robert Ullmann/Mismatched wikisyntax.
Add entries from Wiktionary:Requested entries (Czech).
Development
My contributions: en, cs
- 2010-10-05: en 3281
- 2009-07-25: en 1916
- 2009-02-06: en 615, cs 570
| Datum | Czech | English |
|---|
| 2010-10-05 | 16 226 | 2 030 745 |
| 2010-06-15 | 14 371 | 1 749 745 |
| 2010-05-06 | 14 159 | 1 714 228 |
| 2010-02-07 | 13 703 | 1 574 486 |
| 2009-11-20 | 12 869 | 1 474 540 |
| 2009-10-11 | 11 244 | 1 426 330 |
| 2009-07-25 | 9 193 | 1 330 916 |
| 2009-02-04 | 6 180 | 1 145 264 |
| 2009-01-05 | 6 015 | 1 109 630 |
| 2008-12-01 | 5 573 | 1 073 598 |
| 2008-11-04 | 5 276 | 1 026 681 |
| 2008-10-16 | 5 001 | 988 103 |
| 2008-09-29 | 4 547 | 926 570 |
| 2008-09-15 | 4 294 | 917 586 |
| 2008-07-21 | 3 649 | 866 907 |
Homepage