On first startup the Kura main window is blank, only allowing you to connect to a database.
After you've connected to a database the main window always shows a list with data. Double-clicking on a line will open a dialog window where the data can be edited. Pressing F2 or the search icon in the toolbar will open a filter dialog window. The percent sign (%) can be used as a wildcard in searches. For instance, searching for 'a%' will show all items that begin with an 'a', and searching for '%an%' will show all items that contain 'an'.
F2 | find |
Enter | open a record for changes |
DEL | delete a record |
INS | create a new record |
As explained in the installation guides, Kura needs a database to store its data. This database, MySQL is a fairly professional affair and needs to run as a separate application, a server.
You can connect to a database running on your own computer, or one that is running on another computer, as long as you can reach it over a network and have a username and password for it.
Actually, if MySQL is installed without serious security, just the hostname and databasename will do.
The connection information will be stored in the file .kurarc
, which is either in
your home directory (Unix), or somewhere in the Kura tree (Windows, depending on the value of the HOME variable in kura.bat).
If you have trouble working with MySQL you can ask me for advice; but if things get complicated I might have to refer to the MySQL manual.
There are many options in Kura a user can alter to fit Kura to his needs.
The user Kura will tag every datum with.
The language the user works with most often.
The project the user works on most often.
If you check this, Kura will pre-filter the data it presents using some default values. This can be handy, for instance in the case of a lexicon, where you only want to see data for one language. It can be a hindrance when working on the configuration, where Kura will show only some tags, not all, because some default value filters the others out.
When tagging a text, you can enter a free value for some tags. Kura can give you a picklist with all the values that have already been entered for you to choose from. However, if the list is large, Kura will slow down to a crawl.
This is a fairly complicated way of defining what exactly is shown in the interlinear text view. First you define the text, then the sentences, and then the elements in the sentences - and elements are recursive.
Example:
text; text.title stream.stream text_tag.NOTE
stream: element.element stream_tag.TR stream_tag.NOTE
element: element.text element_tag.GL
This means that a text will show its title, then a list of streams (as defined in 'stream', and then the notes to the text. A stream will show the constituent elements (as defined in 'element'), the translation (any record with the tag TR to 'stream') and notes. Every element will be shown with its text and a glosse below the text.
The list with texts available in Kura can be sorted by language, project, or user responsible for that text. However, you can also apply a filter, and only show the texts that conform to that filter.
If you prefer you can have Kura look like a MacIntosh application, a Windows application or a Unix/X11 application.
The font used throughout Kura for the presentation of language data (except for interlinear texts: the presentation of interlinear texts is more flexible). Be sure to choose a Unicode character set (iso-10646-1).
The Documents menu brings together all the pertinent linguistic data. These are: lexemes, recordings, manuscript scans and analyzed texts.
You can enter new lexemes, tag lexemes with optional data and connect lexemes to each other. For instance, all members of a paradigm might be related to a root form using the lexical relations tab.
To add new tags to a lexeme, press INS (Insert) when in the Tags tab. Depending on the defintion of the tag, you can enter a value, choose a value from a list or enter a longer note.
Sound recordings. In a future version, Kura will be able to play these recordings and synchronize them with the display of interlinear text.
Picture scans of manuscripts. In a future version, Kura will be able to link regions on a scan with elements in an analyzed text.
Texts from a corpus are the core of Kura. A text is parsed, broken up in sentences and words; and the words might even be broken up in morphemes or syllables or even phonemes.
Double-clicking on the title of a text in the list causes the text to appear in the interlinear editor on the right.
For importing text, see Importing and exporting.
Entering a new text is done using the New Text Wizard. Use the "New" button or press Insert while in the text list to start it.
Currently it is not possible to add sentences to an already analyzed text; I think this will be possible in Kura 1.1.
The New Text Wizard can be used to enter new texts in Kura. Another possibility is to import texts that are marked up in XML. However, that currently only works on the command line. See Importing and exporting. The new text wizard is liberally annotated.
Please remember that Kura can only work with Unicode texts, more specifically texts in the standard utf-8 format. Any other format will present you with gibberish.
In the next version of Kura, it will be possible to create dedicated parsers for languages. For now, Kura is able to look for orthographically the same words in other texts in the same language and in the lexicon and use those to create glosses.
The interlinear text editor is a bit slow, but it does work. You can select any part of it by double clicking on text; Kura will open a dialog window where you can edit the particular element you selected. Please note that adding a tag in the element dialog will not add it on the display; however, it will be saved in the database and shown the next time you open the text in the interlinear text editor.
This is the place where basics like users, projects and languages are maintained. The entries should speak for themselves.
When editing these tables, make sure that you have the option "use default values in searches" off - otherwise you won't be able to see all records.
Configuring Kura is not easy. I suggest to go with the initial configuration
from the config.sql
file and just expand the set of tags and tag domains as you see fit, but to use the standard tags wherever possible. It's not neat, it's not good, but in some cases, Kura just expects those tags to be in use.
Tag domains are another kettle of fish. These define fixed lists of things you can choose from, for example parts of speech, or the infamous Penn-Treebank set of standard tags.
When editing these tables, make sure that you have the option "use default values in searches" off - otherwise you won't be able to see all records.
Previous: Importing and exporting | Next: The web server | Table of contents