查看“Marvin”的源代码

==安装==
*pip install sdss-marvin
*pip install -U sdss-marvin

==配置==
* 注意设置一些环境变量，关键是要找到dapall文件
 setenv SAS_BASE_DIR $HOME/MaNGA
 setenv MANGA_SPECTRO_REDUX $SAS_BASE_DIR/mangawork/manga/spectro/redux 
 setenv MANGA_SPECTRO_ANALYSIS $SAS_BASE_DIR/mangawork/manga/spectro/analysis
*mode
 import marvin
 marvin.config.mode = 'local'   # or 'remote'，‘auto’
 marvin.config.access = 'collab' # 'DR15'
 marvin.config.setRelease("MPL-10")

* collab模式下要设置.netrc文件

==Maps==
*读取dap中的map
:map的data model 可以参看 https://sdss-marvin.readthedocs.io/en/latest/datamodel/mpl10.html?highlight=property%20name#dap-datamodel

        ha = dapmap['gflux ha']
        gflux=dapmap.getMap('spx_mflux',channel=None)

* 速度弥散度的改正 inst_sigma_correction() #marvin.tools.quantities.map.Map.inst_sigma_correction
 Ha_sig=dapmap.getMap('emline_gsigma',channel='ha')
 #Ha_sigcorr=dapmap.getMap('emline_gsigmacorr',channel='ha')
 Ha_vdis=Ha.inst_sigma_correction()
 st_sig=dapmap.getMap('stellar_sigma')
 #st_sigcorr=dapmap.getMap('stellar_sigmacorr',channel='fit')
 st_vdis=st_sig.inst_sigma_correction()

==Mask==
===pixmask===
*DAP中的map类都有pixmap的类属性，可以help查看
*比如获得ha的map之后，可以用ha.pixmask.schema查看
:pixmask的第0位是'NOCOV'，就是是否天区覆盖,gflux 并没有设置这个bitmask，但是画图同样可以显示NOCOV天区，原因是因为用了ivar==0判据
:np.where(gflux.ivar == 0)[0].shape
*ha.mask其实就是ha.pixmask.mask  

===manga_target1.mask===
*选源的时候的mask
 from marvin.utils.general.maskbit import Maskbit
 mngtarg1 = Maskbit('MANGA_TARGET1')
 mngtarg1.schema
*map也可以查看这个mask
 ha.manga_target1.mask
 ha.manga_target1.bits
 ha.manga_target1.labels 

====Translating Amongst Mask Values, Bits, and labels====
 ha.pixmask.values_to_bits(1073741843)  # [0, 1, 4, 30]
 ha.pixmask.values_to_labels(1073741843)  #['NOCOV', 'LOWCOV', 'NOVALUE', 'DONOTUSE']
* Translate one label
 ha.pixmask.labels_to_value('NOCOV')  # 1
 ha.pixmask.labels_to_bits('NOCOV')   # [0]
* Translate multiple labels
 ha.pixmask.labels_to_value(['NOCOV', 'UNRELIABLE'])  # 33
 ha.pixmask.labels_to_bits(['NOCOV', 'UNRELIABLE'])  # [0, 5]
====Making a Custom Mask====
* Mask of regions with no IFU coverage
 nocov = ha.pixmask.get_mask('NOCOV')
* Mask of regions with low Halpha flux and marked as DONOTUSE
 low_ha = (ha.value < 1e-17) * ha.pixmask.labels_to_value('DONOTUSE')
* Combine masks using bitwise OR (`|`)
 my_mask = nocov | low_ha
 fig, ax = ha.plot(mask=my_mask)