a better way to search code
Web | Docs | Status | NPM | AUR
Note: All source files are protected by the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International license, included in this directory. Users of this source code (located in this current directory and any sub-directories) may not violate the terms of said license.
Note: Netlify "failed" means either deployment failed or no applicable changes found in last commit.
ensted is slow becuase we eprform multi match over each field and theyre all trigrams so it ends up beign a lot fo computation we may need to do optimizaiton of elastic past what the out of box functionality is
apprently you can convert from h5 to ast (we use h5 for antlr4 currently)
https://github.com/tree-sitter/tree-sitter
https://github.com/github/semantic#technology-and-architecture
Elatic nested fields are: comments variables imports functions classes each nested field ahs a parent as laid out in nested obejct [id parent and location] in elastic they are handled differently this is how we highlight the individual matching object instead of the whole file
think about optimizing the elastic fields for computation time instead of disk footprint possibly more than one search type (classe, functions, libraries, etc...)
Everything will still be stored in a flat fashion in the database, just with graphql we will simulate making it a phat object for ease of query
This is what flat looks like: { repostiory this id array of ids for children
file
this id
array of ids for each type of child
parent id
...
}
This is what phat looks like: repository -> [folder] -> [file] -> [class] -> [class functions] -> [standalone funcitons] -> [imports] -> package path
A field resolver is a block of code which runs when you query for a certain field -> basically lazy evaluation of a field in your data object so we want to write lazy resolvers for each layer of this query
frontload the keyword search with keywords extracted from file documentaitno and definition names and use that to filter out the files along with public access need a compressed representation of data +
Login Search Repository Account Profile Projects Repositories About Explore
Bugs When indexing with the CLI and making a new repository, it will throw an error that the repo does not exist and then if you run it again the repo will show up and work as expected SOEMTIMES seems like a race condition Argument Validation Error : what is it?