Developer notes

JETLS integration in VS Code

For this project, using JETLS is recommended to catch type and inference issues early while editing Julia code.

Remark: according to this repository's changelog, JETLS.jl currently requires Julia 1.12.

Install JETLS

1. Install Julia 1.12 (or newer) and make sure julia is on your PATH.
2. Install the VS Code extension jetls-client.
3. Install JETLS.jl from the command line:

julia -e 'using Pkg; Pkg.Apps.add(; url="https://github.com/aviatesk/JETLS.jl", rev="release")'

1. Verify the CLI is available:

jetls --help

If this command is not found, ensure your Julia binaries/scripts are on the PATH used by VS Code.

Required .vscode/settings.json content

Use the following configuration in your workspace settings:

{
  "jetls-client.executable": {
    "path": "jetls",
    "threads": "1,0"
  },
  "jetls-client.initializationOptions": {
    "n_analysis_workers": 1
  }
}

This configuration tells the extension to run the jetls executable and keeps analysis parallelism conservative.

Project-specific JETLS files

This repository also contains:

• .JETLSConfig.toml
• .JETLSConfig.toml.default

These files configure diagnostics and formatting behavior for this code base. After checkout from git, run

cp .JETLSConfig.toml.default .JETLSConfig.toml

to get an initial configuration file, that you can then modify according to your needs.

Run checks from the repository

You can run the helper scripts shipped with this project:

bin/jetls
bin/jetls_examples

They run jetls check for src/ and examples/ respectively.

More effective development with AI tools using Kaimon

It is suggested to install Kaimon if you are using AI tools like Claude or GitHub Copilot. It allows faster development and reduces the token usage (energy and costs).

See: Kaimon.jl

It is suggested to start kaimon before starting VSCode and keep it open in a terminal window.

The bin/run_julia script will use it automatically if started.

Sometimes you need to tell the agent manually: "Please, use kaimon!".

Installation of Kaimon

Install the app:

]app add Kaimon

Run kaimon --help to see if it works.

Add Qdrant for semantic search:

First, install docker.

Then, run the following code in the terminal:

# This code starts the Qdrant vector database in a Docker container. It creates a Docker 
# volume named "qdrant_data" to persist the data and runs the container with the appropriate 
# port mapping and restart policy.

# It needs to be executed only once. After that, the Qdrant container will automatically 
# restart on system reboot or if it crashes.

# The data is stored in /var/lib/docker/volumes/qdrant_data/_data/collections

docker volume create qdrant_data

docker run -d \
  --name qdrant \
  --restart unless-stopped \
  -p 6333:6333 \
  -v qdrant_data:/qdrant/storage \
  qdrant/qdrant

Install Ollama from ollama.com, then pull the default embedding model:

ollama pull qwen3-embedding:0.6b

Start kaimon with:

kaimon

or with

kaimon --revise

You can quit it by typing q.

The Search tab (press 4) will show a health indicator for both services. If either is not running, the indicator turns red with an error message.

Before you can search, index the project. From the Search tab press m (manage), then a (add) and enter the path for each project you want to index. Then press i (index). To index one project might take 10 minutes, but you have to do that only once.