Logo Lanfrica

basementpool/shilluk-language

Domain:

natural language processing

Record type:

softwaredataset
Creator:
bas
Host:
# Moro-database This project contains the source code for a simple database of a collection of texts and stories in the Moro language. The live version of the website is hosted on the Berkeley Linguistics Site. ## Features - Viewing stories in Moro and English, optionally with gloss lines. - Searching the corpus for a specific string. - A concordance of Moro morphemes in the corpus. # Design ## Constraints - The data backing this website was enterred and storred in a LingSync project. - It should be possible to easily update the data backing the website. - Stories were being enterred and cleaned while the website was under construction. - Needs to be hosted on existing berkeley linguistics website infrastructure. - Easiest just to have all static content being served. ## Approach - Data source: - LingSync uses couchDB to store it's backing data. - We set up a mirror on cloudant.com and add some views that clean up the data (ask @sarum90 if you need details, sorta hacky) - In the end there are two URLs that return live data from the LingSync project: - sande570.cloudant.com - sande570.cloudant.com - Data processing: - Concordance is constructed in JavaScript in the client browser every time they load the website. - This means there is no "export data" script, the site always has live data from LingSync. - It somewhat complicates the logic in moroScript.jsx. - It increases load times of the website. - This is usually being done in the background while people are reading the landing page. - Data rendering: - Use a single-webpage design using React / ReactRouter to manage the various links. - All data is loaded/processed once when the page is openned, clicking just changes the view of the data. - Use semantic-ui to make things look prettier. # Technical Stack - Semantic UI For responsive HTML widgets / loading bars / layout …

Languages

Licenses