Re-examining the radial velocity detection of L 98-59 b

Animation of the L98-59 system

Written by Max Talberg

email: mt942@cam.ac.uk

June, 2024

Updated: July, 2024

Introduction

In recent years the radial velocity (RV) method has proven exceptionally effective at detecting exoplanets. Advancements in instruments like HARPS (Mayor et al. 2003) and ESPRESSO (Pepe et al. 2021) have pushed the detection threshold to new lows. Introducing L98-59b, a terrestrial planet with half the mass of Venus, the lowest mass planet detected using RVs to date (Demangeon et al. 2021). We aim to re-examine the detection of L98-59b, utilising a Gaussian Process (GP) framework. This is the Data Analysis Pipeline for the project.

Run the L98-59 Model Locally

The Data Analysis Pipeline contains the code to preprocess and produce periodograms from the HARPS and ESPRESSO dataets, run the L98-59 model and generate results.

Setting Up the Project

1. Clone the repository:

   • Shallow clone the repository from GitLab:

     git clone --depth 1 git@gitlab.developers.cam.ac.uk:phy/data-intensive-science-mphil/projects/mt942.git

2. Set up the virtual environment:

   • Navigate to the project directory:

       cd mt942

   • Create virtual environment:

     conda env create -f environment.yml

   • Activate virtual environment:

     conda activate l9859-env

   • The L98-59 model requires PolyChordLite (Handley et al. (2015)) be installed first:

     git clone https://github.com/PolyChord/PolyChordLite.git
     cd PolyChordLite
     make
     pip install .
     cd ..

3. Running the script:

   • Running the main script to preprocess and produce periodograms from the HARPS and ESPRESSO datasets:

     python src/main.py

   • The L98-59 model is ran from, the run_l9859.py script where one can be tune the L98-59 model with different parameters:

     python src/run_l9859.py

Documentation

The documentation can be accessed here or generated locally using the following steps:

Generating Documentation

1. Navigate to the docs directory:

    cd docs

2. Generate the documentation:

    make html

3. Open the documentation:

    open build/html/index.html
   

Testing

Running Unit Tests

1. Navigate to the project directory tests:

    cd ../tests

2. Run the tests:

    export JUPYTER_PLATFORM_DIRS=1
    pytest

Running the L98-59 Model on Docker

Setting Up the Project

1. Clone the repository:

   • Shallow clone the repository from GitLab:

     git clone --depth 1 git@gitlab.developers.cam.ac.uk:phy/data-intensive-science-mphil/projects/mt942.git

2. Build the Docker image:

   • Navigate to the project directory:

     cd mt942

   • Build image:

     docker build -t l9859-project .

3. Running the script:

   • Run the main script:

     docker run -it --rm -v host_directory:/app/plots l9859-project

     • Replace host_directory with the path to the directory where you want to save the plots, for example: /path/to/plots and all the images will be saved into a folder named plots, acompanying information will be in the terminal output.
   • Running PolyChordLite:
     • Currently, we have not been able to successfully configure Docker to work with PolyChord Lite. We are investigating this issue and plan to provide updates as soon as a solution is found.

Specifications

On a laptop with the following specifications:

• Model: MacBook Pro
• Processor: Apple M2 chip
• Memory: 16GB RAM

The smallest L98-59 model:(planet b: False, FWHM: False, S-index:False) took approximately 20 minutes to complete.

The full L98-59 model:(planet b: True, FWHM: True, S-index:True) took approximately 2 hours to complete.

These times can be reduced by increasing the prescion criteria and decreasing the number of repeats in the run_l9859.py script.

Notes

The executive summary report/ExecutiveSummary.pdf contains a summary of the project and the results in the larger context of exoplanet detection. word count: 973

The full report report/Report.pdf contains a detailed description of the project, the methods used, the results and the implications of the results. word count: 6,879

Acknowledgements

I want to thank my thesis supervisor, Dr. Vinesh Rajpaul, for his guidance, constant support and GP wisdom. His expertise was key to finishing this thesis. Our late-night meetings, despite the frequent bad call quality, were especially memorable and instrumental in my progress.

I’m also very grateful to Dr. James Fergusson, the course director, for giving me this fantastic opportunity. This past year has been unforgettable and has significantly changed the trajectory of my life.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Use of generative tools

This project has utilised auto-generative tools in the development of documentation that is compatible with auto-documentation tools, latex formatting and the development of plotting functions.

Example prompts used for this project:

• Generate doc-strings in NumPy format for this function.
• Generate Latex code for a subplot.
• Generate Latex code for a 3 by 3 matrix.
• Generate Python code for a 2 by 1 subplot.