Artha - An cross-platform thesaurus based on WordNet
Artha is an open thesaurus based on the WordNet database, created with
simplicity in mind. Once executed, Artha monitors for a user preset global
hotkey combination. When the user selects some text in any window, and presses
this hotkey combo, Artha looks up WordNet thesaurus for the selected text and
pops-up with the results.
When executed for the first time Artha tries to register a hotkey automatically,
in the order, Ctrl + Alt + [W or A or T or Q]. You can view/change it via the
'Hotkey' button in the toolbar. It can also be disabled.
Definitions are categorized based on the PoS (Part of Speech - Noun, Verb,
Adjective and Adverb). Apart from showing the definitions/senses of a searched
string with usage examples, Artha also shows a word's relatives like Synonyms,
Antonyms, Derivatives, Pertainyms (Related noun/verb), Attributes, Similar
Terms, Domain/Domain Terms, Causes, Entails, Hypernyms (is a kind of),
Hyponyms (Kinds), Holonyms (is a part of), Meronyms (Parts).
A word can have more then one sense i.e. it can convey more than a single
meaning/definition. Relative words are words that are related to one or more
senses of the searched word, by a relationship like Synonym, Derivative, etc.
To know which all sense a relative is related to, just select the it, the
corresponding senses it maps to are highlighted. As per WordNet, depending on
the number of senses a word has (polysemy count), it's familiarity is
determined. It gets displayed next to the PoS in the definition area. There
are 7 types: extremely rare, very rare, rare, uncommon, common, familiar, very
familiar and extremely familiar.
Artha has 2 modes. Simple and Detailed. Artha enters Detailed mode when the
'Detailed' button in the toolbar is pressed. When toggled again, it returns
back to simple mode. For relatives like Antonyms, Pertainyms, Hypernyms,
Hyponyms, Holonyms and Meronyms, where more than one level of relatives may be
present, is showed in a tree fashion, in detailed mode. If in simple mode,
only one level of relatives are shown even when more levels are present. E.g.
'rich' has 'poor' and 'lean' alone as antonyms in simple mode. While in
detailed mode, 'poor' further infers broke, skint, etc. which are shown as
children of 'poor'.
Regular expressions can be used to search for a term when you vaguely know it
and want to locate it in the thesaurus. Artha's regular expression pattern
closely follows Wildmat syntax by Rich Salz owing to its simplicity.
- * (wildcard) matches any number of (including 0) unknown
characters
- ? (joker) matches one unknown character
- [...] (range) matches one unknown character within the
range specified
- {m, n} (limits) upper & lower limits of the number of
characters in a range
- [^...] (not in the range) matches one unknown character NOT
within the range specified
- EXAMPLES:
- Expr. `cro*p` means the term you want to corner starts with
`cro` and ends with `p` while the number of characters in between are
unknown. It fetches crop, crop up, croup, crock up and crow step.
- Expr. `*chester` means the searched word ends with a
`chester` while the beginning and its number characters are unknown. It
fetches chester, manchester, rochester, winchester and toy
manchester.
- Expr. `can????r` means the term sought starts with `can`
and ends with `r` while you are sure that there are 5 unknown characters
in between. It fetches canister and cannular.
- Expr. `andre*[x|y|z]` means the word searched for starts
with andre and ends with either an x or y or z, and there could be any
number of terms in betweem these. It fetches andre malraux, andrei
tarkovsky, andres martinez, etc.
- Expr. `a[c|d|e]{2,}` means the word looked for starts with
a and then there are minimum 2 or more occurrences of c, d or e. It
fetches acc, accede, ace, add, ade and aec.
Look-ups made in Artha are stored permanently as a per-user setting, and can be
saved via the popup menu of the search bar or can be cleared too. Other
options like showing term familiarity based on polysemy count, display of
status icon, etc. can be changed via the Options window.
Should the user prefer passive desktop notifications (balloon tips), rather than
the application popping up with the definitions, it can be done by enabling
Notifications. This is done via the Notify tool button or by right-clicking on
Artha's system tray icon, and tick off the 'Notifications' check box in the
menu. When notifications are enabled, and the user selects text in a window
and presses the hotkey combo, Artha takes the prime definition of that term
from WordNet and shows that definition as a system tray notification.
Note: For the notifications feature to be present, notify library's binary
(libnotify.so.1) should be available on your system. If not, Artha will not
expose the feature at all. Also the notification-daemon should installed for
the notifications to show up.
Suggestions is a feature that gives out possible near matches when a misspelled
word is searched for. To have this feature, your system should have libenchant
binary (libenchant.so.1) installed and an English dict file for the spell
engine to refer (locale doesn't matter).
Artha has a World Wide Web site at
http://artha.sourceforge.net. From
this web site users can know more about the Artha project and also download
its source and binary distributions for various distros.
Sundaram Ramaswamy <
[email protected]>