This release finally allows `vespa` to be compatible with the latest versions of `isochrones`. This required many changes, but most of it is under the hood. There have also been a number of other bug fixes and general tidying up. The most significant usage change is that the `calcfpp` script now no longer runs the stellar model fits; it is now expected that you first execute `starfit --all <candidate_directory>` and then `calcfpp <candidate_directory>`.
Changes/improvements/bugfixes:
* Unit tests are updated and working again.
* Default stellar models are now MIST.
* `ColormatchMultipleStarPopulation` and `SpectroscopicStarPopulation` are now completely removed; they were outdated and have already been functionally replaced by `Observed_BinaryPopulation` and `Observed_TriplePopulation`.
* The `get_ichrone` convenience function now allows easier instantiation of the desired set of stellar models by passing a string, which allows for better lazy initialization of many objects, and removes the need to initialize `Isochrone` objects upon import.
* HDF file handling is improved, to avoid unintentionally leaving files open after pandas reads.
* Now uses `emee.autocorr` rather than `acor` to test convergence of trapezoidal model fit.
* Adds `TransitSignal.from_ascii` function to allow a `TransitSignal` to be loaded straight from a text file.
* Reads sample columns from `StarModel` appropriately for isochrones > 1.0 (e.g., `Teff_0_0`, etc.)
* `TransitSignal.from_ini` is cleaned up and slimmed down. This function no longer executes the stellar model fitting; it just assumes they have been done and reads them in (and complains if they are not).
* The `kepler` module is now compatible with the DR25 data release. Included in these updates are corrections to the TTV corrections and stellar parameter provenance when setting the priors for star fits.
Known issues
* Documentation is not yet fully updated.
* There is still no sanity-checking of isochrones stellar model fits built in; always check the quality of the model fits (e.g., by eyeballing `mist_corner_single_observed.png`) to avoid being misled by the model fits and [ending up with a wrongly validated planet](https://arxiv.org/abs/1708.08455).