Backtracks

:star: v0.3.3 now on pypi 💫

What's Changed
* Sphinx documentation by tomasstolker in https://github.com/wbalmer/backtracks/pull/28
* Added the gaia_release and gaia_epoch attributes by tomasstolker in https://github.com/wbalmer/backtracks/pull/34
* Cleanup by wbalmer in https://github.com/wbalmer/backtracks/pull/36


**Full Changelog**: https://github.com/wbalmer/backtracks/compare/v0.3.2...v0.3.3

tomasstolker saves us yet again. Major clarifications and improvements to the package structure (no more binary files, for instance). Fixed a big bug where data with covariance between terms was not included in the likelihood function (this was defeating the purpose of including all our excellent GRAVITY data!) and a few more knobs and user options (like setting sampling style within dynesty - if your sampler is running slowly, try changing the `sample_method` keyword from `'unif'` to `'rwalk'` or `'rslice'` in `system.fit` !)

A quick hot-fix to ensure that installing directly from pypi, e.g.


pip install backtracks


includes the data files needed to run. Now, no need to clone the repo! Just run


conda create python=3.9 -n backtrack
conda activate backtrack
conda install pip
pip install backtracks


download our example data and script (or make your own), and run


wget https://raw.githubusercontent.com/wbalmer/backtrack/main/tests/scorpions1b_orbitizelike.csv
wget https://raw.githubusercontent.com/wbalmer/backtrack/main/tests/hd131339a.py
python hd131339a.py

Thanks to tomasstolker , `backtrack` can now be used on clusters using `schwimmbad.MPIPool`

new also is the `set_prior_attr` function and new options in the `fit` function, like nlive=500, mpi_pool=False, resume=False.

More to come (maybe someday I'll write docstrings)!

Hello world!

`backtrack` is here, and I'm version-ing the code here so we can continue development. Lots of work yet to be done (like docstrings, lol) but we can reproduce the results of [Nielsen et al. 2017](https://arxiv.org/abs/1705.06851), and the code appears applicable to at least one other case (citation TBD), which was the intention of the code in the first place.