What it does
Enter a regex expression and find semantically similar code snippets from popular OSS projects.
A frontend built with React takes a regex expression which is spellchecked using the Bing Spell Check API from Azure Cognitive Services. A word2vec model generates embeddings for the words in the regex expression. The embeddings are used to generate a regex that includes semantically similar expressions. The altered regular expression is then sent to a cluster of Elastic search nodes hosted on the Azure cloud which searches for matches among popular OSS repositories previously loaded onto the cluster. The best matches are then displayed according to their similarity score.
How we built it
Challenges we ran into
- Deploying a 4 GB file to Azure cloud
- Working with 7 year old outdated API documentation
- Centering a HTML box with CSS
- 'I just deleted the fucking database, there is nothing wrong with that' ~Eavan 2020, 1 hour before submission deadline
Accomplishments that we're proud of
What we learned
- The python environment on 2/4 of our laptops resembles a post-apocalyptical scenerie
What's next for Glance
git rm -r