Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/iancze/imlup
Example Repository for DSHARP IM Lup MPoL Analysis
https://github.com/iancze/imlup
Last synced: about 1 month ago
JSON representation
Example Repository for DSHARP IM Lup MPoL Analysis
- Host: GitHub
- URL: https://github.com/iancze/imlup
- Owner: iancze
- License: mit
- Created: 2023-12-01T10:26:45.000Z (about 1 year ago)
- Default Branch: main
- Last Pushed: 2024-05-21T23:32:20.000Z (8 months ago)
- Last Synced: 2024-05-22T00:31:43.590Z (8 months ago)
- Language: Python
- Size: 86.9 KB
- Stars: 0
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# IMLup
Repository for DSHARP IM Lup MPoL Analysis
# Directory Structure
IMLup/
casa_analysis/
src/
data/
raw/
temp/
Snakefile
mpol_analysis/
src/
data/
raw/
temp/
Snakefile
The repository is configured the way it is because I need to approach this folder using two separate Docker images, mounting the `IMLup` folder as a volume. This is because (as of December 2023) the Docker images that will run CASA without bugs (Python 3.8) are so old that they won't run the optimized NVIDIA GPU files (Python 3.10, at oldest).
The main idea is that Docker image #1 is non-GPU and configured with CASA modular and works with the scripts in `casa_analysis`. The goal of these scripts is to double-check we're understanding the DSHARP data correctly, (possibly) rescale the weights, and export the visibilities to a file format that does not depend on CASA, e.g. `casa_analysis/data/temp/data.asdf`.
Docker image #2 is configured with MPoL, other modern PyTorch analysis scripts, and picks up from the data exported from the CASA analysis (copied to `mpol_analysis/data/raw/data.asdf`).
The Dockerfiles are outside of the scope of this particular repository, but can be provided upon request, if useful.