# Install Julia
curl -fsSL https://install.julialang.org | sh -s -- -y
export PATH="/root/.julialup/bin:${PATH}"
# Set up Julia environment
julia -e 'using Pkg; packages = ["IJulia", "DataSets", "XLSX", "Plots", "Downloads", "DataFrames", "ImageShow", "FileIO", "JSON3", "CSV" ]; Pkg.add(packages);'
# Install Mimi Julia library
julia -e 'using Pkg; Pkg.add("Mimi"); using Mimi'
# Install Mimi FUND
julia -e 'using Pkg; Pkg.add(url="https://github.com/fund-model/MimiFUND.jl.git"); using MimiFUND'
# install beaker-climate contexts
pip install -e .
export OPENAI_API_KEY=your key here
export GEMINI_API_KEY=your key here
export ANTHROPIC_API_KEY=your key hereRun with beaker notebook
There are two contexts: beaker_climate and mimi-modeling. beaker_climate is used for general climate science questions in Python and mimi-modeling is used for questions about the Mimi integrated assessment models in Julia. The mimi-modeling context has only been extensively tested using Anthropic's Claude 3.5 Sonnet model as the primary model. Though it should work with other models, it may not perform as well.
To run the docker container, first copy .beaker.conf.example to .beaker.conf and set the LLM keys both there and in the .env file. Then use the following command:
docker compose build
docker compose upbeaker-climate is distributed under the terms of the MIT license.