KevExplorer

Project organization

This is a Javascript monorepo composed of the following packages:

• core: Core features (scan a folder)
• cli: The CLI program
• api: Web API exposing core features
• app: Web application using the API

Installation

Just run:

yarn

Run the CLI

yarn cli

Run the app

Start the backend

With Docker:

docker-compose up

In development mode:

yarn api

The backend listens on http://localhost:8080.

A GraphiQL UI is available here: http://localhost:8080/graphiql

Start the Web app

yarn app

Run the tests

All tests can be run with:

yarn test

You can also get into a specific package and run the same command.

For instance, run tests for the cli packages:

cd packages/cli
yarn test

TODO

Frontend

• Tests (unit, e2e)
• Logs (Sentry)
• Analytics?
• Environment config
• Accessibility refinements
• Virtual scroll would be nice for very long lists
• i18n
• Better responsive
• Navigation history
•  Caching folders content
•  Authentication

Backend (including the file scanner)

• Logs
• Environement configuration
•  E2e tests

More about the scanner:

• [PERF] Better concurrency management. Currently, the limit of concurrency is 4... for each scanned folder. We could get a better management by using node Workers and push jobs into a pool that can provide a global limitation for those parallel tasks.
• [SECURITY] Give a folder as a root path (a jail) from which the user can't get out

Known issue

Memory: when the tree of files is too big, we hit a JavaScript heap out of memory error. If you want run Node with more memory, you can set this environment variable:

NODE_OPTION='--max-old-space-size=8192'

Symlinks loops: sometimes

CLI

• Improve UX by handling autocompletion, history...
• The table could be better