sblunt / orbitize / 105 / 1
74%
main: 86%

LAST BUILD BRANCH: dummy
DEFAULT BRANCH: main
Ran
Files 8
Run time 1s
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst
coverage: 61.207% (-7.1%) from 68.269%
105.1

push

travis-ci

sblunt 
Draft API implementation

First draft of API.

Here are some notes about the structure of the code:

1. The standard way that I imagine users using orbitize is as follows:
   - Using Henry’s `read_input.py`, read in a data file, and produce a standardized data table.
   - Initialize a `System` object, with Henry’s data table as an attribute. The System object will initialize a list of `Prior` objects in a standard order that will be used by the `Sampler` object. The user can then overwrite priors they wish to customize.
   - Initialize a `Sampler` object (either a `PTMCMC` object or an `OFTI` object), with the `System` object and a `lnlike()` function as attributes. Right now we will only have a chi squared likelihood, but in the future we will add Gaussian process likelihoods & other fancy ones.
   - Run the `run_sampler()` method of the initialized Sampler object. This will involve calling `kepler.py` to compute orbits, and calling a helper function that will match the outputs of `kepler.py` with their appropriate data points (e.g. pairing “radec” data with position outputs, “rv” data with velocity outputs).
   - Save the results, and optionally make some plots.
2.  I’ve added a file, `driver.py` that contains a class to run through all the standard steps listed above automatically. At the highest level, users should just be able to feed all the information/data they have about into a Driver class, hit run, and relax.

Notes about things I’ve changed:
1. I’ve taken out the `CelestialDuo` class. I think it will be more readable and extensible if instead of using this class, we simply add another column to Henry's data table, specifying whether each data point corresponds to planet a, planet b (etc), or all planets (as in the case of RV/stellar astrometry measurements).
2. I’ve added an abstract base class to the `sampler.py` classes, just for tidiness.

142 of 232 relevant lines covered (61.21%)

0.61 hits per line

Source Files on job 105.1