All notable changes to this project will be documented in this file.
The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.
- Add "tweet_mode = extended" in Twitter API calls to retrieve full text of tweet
- fix frontend bug on classification when click "split", "train", "predict" button
- use forever library to auto-restart servers when it fails
- populate more graphql errors back to SMILE server then to the browser
- fix bugs and change logic in search pagination
- update citation to the future generation computer publication
- UI interface change based on Hubzero usability report
- POST /dryRun endpoint to do a "pre" search instead of save the search results directly
- GET /citation endpoint to list the citation
- bugfix the urls schema type in twtTweetType.js from String to List
- Add ids to analysis output (hence rotate the aws lambda for NW, SA, PP)
- Add restrictions and hint text for upload user dataset
- Add features to collect images from social media sources
- Modify the Authorization page when search from social media data sources
- Add sentiment classification using debiased word embedding algorithm
- Add Gensim LDA topic modeling as a tool in SMILE analytics
- Add an identity check on the server side to check if the requested path (from frontend) to access a certain s3 folder matches the user's identity
- Add SVC, Decision Tree, and Adaboost classifier alogrithm for text classification
- Highlight the pipeline download in the popup of text classification
- Fix UID bug in text classification
- Add Twitter timeline as a datasource in GraphQL.
- Query interface updated.
- Search for twitter user feature removed.