Output
Add intermediate output steps
inputs
msout.antennacompression
- type:
bool
- doc:
If set to true, a special AntennaStMan storage manager will be used that (losslessly) compresses the antenna values of the ANTENNA1 and ANTENNA2 columns. Normally, the pair of values from these columns form a repeating pattern; the storage manager just stores the repeating pattern once.
- default:
depends on compile-time option METADATA_COMPRESSION_DEFAULT
msout.chunkduration
- type:
double
- default:
0.0
- doc:
Duration (in seconds) after which to start writing a next measurement set. The new measurement set is named with the number of the chunk (with leading zeros to make it at least contain 3 digits) added in front of the extension, prepended by a dash. For example, if msout is “myobs.ms”, the first chunk will be called “myobs-000.ms”, the next one “myobs-001.ms”, etc. When the values is set to zero (the default), no time chunking is performed..
msout.clusterdesc
- type:
string
- default:
""
- doc:
If not empty, create the VDS file using this ClusterDesc file.
msout.datacolumn
- type:
string
- default:
DATA
- doc:
The column in which to write the data. When creating a new MeasurementSet, only column DATA can be used. When updating the input MeasurementSet, any column can be used. If not existing, it will be created first. When this setting is used, the output step always writes the data.
msout.flagcolumn
- type:
string
- default:
FLAG
- doc:
The column in which to write the flags. When creating a new MeasurementSet, only the name FLAG can be used. When updating the input MeasurementSet, any column name can be used. If not existing, it will be created first. When this setting is used, the output step always writes the flags.
msout.overwrite
- type:
bool
- default:
False
- doc:
When creating a new MS, overwrite if already existing?
msout.scalarflags
- type:
bool
- default:
depends on compile-time option METADATA_COMPRESSION_DEFAULT
- doc:
If set to true, only one flag will be stored for four correlations. For LOFAR, all correlations have equal flags, so storing all flags is redunandant. This typically saves a couple of percent storage.
msout.storagemanager.databitrate
- type:
integer
- default:
10
- doc:
Number of bits per float used for columns containing visibilities. Can be set to zero to compress weights only.
msout.storagemanager.distribution
- type:
string
- default:
TruncatedGaussian
- doc:
Assumed distribution for compression; “Uniform”, “TruncatedGaussian”, “Gaussian” or “StudentsT”.
msout.storagemanager.disttruncation
- type:
double
- default:
2.5
- doc:
Truncation level for compression with the Truncated Gaussian distribution.
msout.storagemanager.normalization
- type:
string
- default:
AF
- doc:
Compression normalization method: AF, RF or Row.
msout.storagemanager.weightbitrate
- type:
integer
- doc:
Number of bits per float used for WEIGHT_SPECTRUM column. Can be set to zero to compress data only. Note that compressing weights will set all polarizations to the same weight (determined by the minimum weight over the polarizations).
- default:
12
msout.storagemanager/msout.storagemanager.name
- type:
string
- default:
""
- doc:
What storage manager to use. Options: empty (=default), “stokes_i” or “dysco”. When empty, the data will be stored uncompressed. “stokes_i” will use the StokesIStMan, which stores only one (Stokes I) value for four correlations. For this case, the 2nd and 3rd correlations (normally the XY and YX values) must be zero. When set to “dysco”, the data will be compressed using Dysco. Settings below will set the compression settings; see the Dysco wiki and the paper for more info. The default settings are reasonably conservative and safe.
msout.tilenchan
- type:
integer
- doc:
For expert user: maximum number of channels per tile in output MS.
- default:
64
msout.tilesize
- type:
integer
- default:
1024
- doc:
For expert user: tile size (in Kbytes) for the data columns in the output MS.
msout.uvwcompression
- type:
bool
- default:
depends on compile-time option METADATA_COMPRESSION_DEFAULT
- doc:
If set to true, a special UvwStMan storage manager will be used that (losslessly) compresses UVW values. It does this by taking advantage of the fact that values within of the same correlation matrix can be determined by just storing one uvw-value per antenna. With this compression, UVW values do not need to be recalculated from the antenna positions.
msout.vdsdir
- type:
string
- default:
""
- doc:
Directory where to put the VDS file; if empty, the MS directory is used.
msout.weightcolumn
- type:
string
- default:
WEIGHT_SPECTRUM
- doc:
The column in which to write the weights. When creating a new MeasurementSet, only WEIGHT_SPECTRUM can be used. When updating the input Measurementset, any column can be used. If not existing, it will be created first. When this setting is used, the output step always writes the weights.
msout/msout.name
- type:
string
- doc:
Name of new output MeasurementSet; if empty, the input MS is updated. The other msout parameters are not applicable (apart from
countflag
). Normally an update is only done if a step is given that can change the data (e.g. PreFlagger). However, a name ‘.’ or a name equal to the name of the input MS means that the input MS will always be updated, even if no step is given. This is useful if only flagging of NaN-s in the MS needs to be done. Note that when doing averaging, the input MS cannot be updated.