Astrobase

Fixes

- checkplot: use axes_grid1 because axes_grid is deprecated apparently
- features.all_nonperiodic_features: add back magsarefluxes kwarg
- lcmath: add (optional for now) iterative sigclip to sigclip_magseries
- lcmodels.eclipses: fixes to docstring
- lcmodels.eclipses, lcfit: implemented invgauss_eclipses model & fit fns
- lcproc.plot_variability_thresholds: yscale -> log
- lcproc: stetson_threshold -> variability_threshold, now per mag-bin
- lcproc: stupid bug fix
- lcproc.variability_threshold: get and plot LC RMS as well
- recoverysim: added make_fakelc_collection
- recoverysim: fixed up make_fakelc, make_fakelc_collection
- varbase.features: some more measures added
- varbase.lcfit: fix confusing fourierorder, fourierparams kwargs. GH PR 27
- varbase.signals: fixes for GH PR 25 and 29
- varbase/signals: fix long-broken gls_prewhiten iterative prewhiten function
- varclass.recoverysim -> two modules: fakelcgen and fakelcrecovery
- Wrapped LOGWARNING in `if verbose:` statement


Work that may be in progress

- [WIP] moved various LC models into lcmodels, TODO: finish these
- [WIP] recoverysim: redo make_fakelc, now gets mag-rms from previous results
- [WIP] setting up the varclass package
- [WIP] various implementation notes
- [WIP] working on recoverysim

Fixes

- lcfit: fix error handling
- lcfit: implement traptransit_fit_magseries, to be tested
- lcfit: tested/fixed traptransit_fit_magseries, seems to work OK
- lcfit.traptransit_fit_magseries: get formal 1-sig errs for final params
- lcfit.traptransit_fit_magseries: get formal 1-sig errs for fit as well
- lcproc/checkplot: use full lcfpath in makelclist, getlclist, nbr procs
- lcproc, features: fix broken beyond1std, add bestmagcol to results
- lcproc.makelclist: add maxlcs kwarg
- lcproc.makelclist: fixing bugs
- lcproc, periodbase: update period-finder registry dicts
- periodbase: fix falsealarmprob for PDM
- periodbase.make_combined_periodogram: addmethods kwarg
- README: added zenodo DOI badge
- smav: bump up default nharmonics to 6 from 4
- smav: use correct dot product function (np.vdot not np.dot)

Fixes

- checkplot: add neighbor stuff to checkplot png as well
- checkplot.checkplot_pickle_to_png: actually fix mistaken pkl outfile ext
- checkplot: fix handling of no neighbors case
- checkplot: get neighbors of target, their radecs, xys, and lcfpaths
- checkplotlist: added some TODOs for multi-checkplot-per-object stuff
- checkplotlist: more notes on planned features
- finally added coding: utf-8 to everything
- glsp: remove because all moved to zgls.py
- lcmath: added a fill_magseries_gaps function
- lcmath, autocorr: add some guards against weirdness
- lcmath.fill_magseries_gaps: fix normalization of fluxes
- lcmath.fill_magseries_gaps: use 0.0 to fill gaps instead of noiselevel by default
- lcmath.fill_magseries_gaps: use scipy's perfectly cromulent mode fn
- lcmath.fill_magseries_gaps: use scipy's perfectly cromulent mode fn instead
- lcproc: added update_checkplotdict_nbrlcs to get neighbor LCs
- lcproc: add neighbor stuff to parallel_cp workers and driver
- lcproc: add serial_varfeatures
- lcproc: better kwargs for xmatchexternal
- lcproc: fix default fileglobs for HAT and HPX LCs
- lcproc.makelclist: do concurrent (threaded) info collection instead
- lcproc.makelclist: do parallel info collection instead
- lcproc.makelclist: get actual ndets per magnitude column
- lcproc.makelclist: remove useless fileglob from tasklist
- lcproc.parallel_cp_lcdir: make output directory if it's doesn't exist
- lcproc.parallel_varfeatures: use ProcessPoolExecutor instead of mp.Pool
- lcproc.runpf: bugfix
- lcproc.runpf: make BLS SNR optional because it takes forever
- lcproc.runpf/runcp: allow any of the period-finders from periodbase
- lcproc: simplify parallel task list for makelclist
- macf: added macf_period_find; seems to work, TODO: more testing
- macf: added plot_acf_results to see smoothed/unsmoothed ACF and peaks
- macf: fixed, tested; reproduces McQuillian+ (2014) results on KeplerLCs
- macf: fix handling of kwargs
- macf: fix peak detection
- macf: search interval for ACF peaks now depends on smoothing
- macf: some more docstring work
- macf: update docstrings
- periodbase: add a make_combined_periodogram fn
- periodbase: add macf to base namespace
- periodbase.macf: initial bits for McQuillian+ (2013a, 2014) period-finder
- plotbase: get the FITS header for the cutout for WCS purposes later
- README: updated for some new stuff
- README, various: add links to ADS for period-finders, fix spelling of names
- smav: bugs fixed, seems to be working OK
- smav: implemented aovhm_periodfind, to be tested
- smav: initial bits
- smav: some more progress
- varbase.autocorr: actually return acf as np.array
- varbase.autocorr: autocorr_magseries fn using lcmath.fill_magseries_gaps
- varbase.autocorr: fix _autocorr_func3 and set as default
- varbase.autocorr: return acf as np.array
- varbase.autocorr: use 0.0 to fill gaps instead of noiselevel by default
- varbase.features: fix annoyingly subtle bug causing np.polyfit crashes
- varbase.features: np.nonzero -> npnonzero
- zgls, lcfit: guard against zero errors here too
- zgls: move the glsp functions into here

Fixes

- lcproc: add external cross-matching to getlclist
- lcproc.getlclist: also copy matching light curves to requested dir
- lcproc: getlclist, makelclist now have column-filtering and cone-search

Fixes

- lcproc: don't crash runpf if BLS SNR fails

Fixes

- cplist: add 'ne' -> not-equal-to filter operator for --filter arg
- cplist/cpserver: now takes multiple filters to be ANDed together; yet to test
- cplist: fix docstring
- cplist: fixed multiple filter keyword arguments
- cplist: fix incorrect filtering
- cplist: fixing multiple filter args