We are not required to provide a refund or replacement if you change your mind.īut you can choose a refund or exchange if an item has a major problem. update_disk ( new_disk_model_convolved ) fmout = diskobj. shape, numbasis, dataset, disk_model_convolved, aligned_center = aligned_center, basis_filename = 'path/to/dir/klip-basis.h5', load_from_basis = True ) # do the forward modelling on a new model new_disk_model_convolved = convolve ( new_disk_model, instrument_psf, boundary = 'wrap' ) diskobj. klip_dataset ( dataset, diskobj, outputdir = "path/", fileprefix = "my_favorite_disk", numbasis = numbasis, maxnumbasis = 100, aligned_center = aligned_center, mode = 'ADI', annuli = 2, subsections = 1, minrot = 3 ) # - # starting from here you can close the session and reopen later if you want # - # load Klip parameters and FM basis diskobj = DiskFM ( dataset. shape, numbasis, dataset, disk_model_convolved, aligned_center = aligned_center, basis_filename = 'path/to/dir/klip-basis.pkl', save_basis = True ) # run klip to find and save FM basis fm. aligned_center = # indicate the position of the star # convolved the disk model disk_model_convolved = convolve ( disk_model, instrument_psf, boundary = 'wrap' ) # initialize the DiskFM class diskobj = DiskFM ( dataset. spectral_collapse ( collapse_channels = 1, align_frames = True ) numbasis = # different KL numbers we applied to the disk. GPIData ( filelist ) # in case of multiWL data, you might want to stack them first to speed things up dataset. glob ( "path/to/dataset/*.fits" )) dataset = GPI. Import glob import numpy as np import as GPI from nvolution import convolve from import DiskFM import pyklip.fm as fm # read in the data into a dataset filelist = sorted ( glob. This has now beed solved and all the information necessary is saved inside the. #Klipped to save code#This caused problems because you could run the same KL coefficients with slightly differentĭatasets (for example the order of the frames were not identical) the code would run but provide wrong forward models. In previous version, the dataset itself (input images) were not saved in the. Need diskFM with load_from_basis = True). You still need to re-create the DiskFM object and load it ( ie, you still ( ie even if you have runned diskFM with save_basis = True) in this python session, Note that even if you have already created a DiskFM object to save the FM These last 3 lines are specifically what should be repeated withinin the MCMC update_disk ( third_disk_model_convolved ) fmout = diskobj. fm_parallelized () # do the forward modelling on a third model third_disk_model_convolved = convolve ( third_disk_model, instrument_psf, boundary = 'wrap' ) diskobj. Only supports KLIP ADI, SDI ADI+SDI and RDI reductions (but currently not RDI+ADI/SDI or NMF).ĭiskobj = DiskFM ( dataset. These routines are implemented in PyKLIP and showed on this page. For a new disk model, the forward modelling is therfore only aĪrray reformating and a matrix multiplication, which can be optimized to be only a few One can save the KL vectors in a file once so they do not have to be However, once measured for a set of reduction parameters, the Karhunen-Loeve (KL) basisĭo not change. To the klipped reduced image of the data, within an MCMC or a Chi-square wrapper. Geometries, the forward modelling have to be repeated a lot of time on disks Modeling with disks is complicated by theįact that cannot be simplified using a simple PSF as the astrophysical model:Įvery hypothetical disk morphology must be explored”. As noted in Pueyo (2016), “ in practice Forward
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |