There is now a way to register datasets entirely from the command line. qed requires four pieces of information:
the product type so that it can invoke the correct reader; specified using --reader
the pixel data type; specified using --cell
the shape of the dataset; specified using --shape
the name of one or more files to open; the chosen reader must be able to open all the files
The reader is specified with the reader command line argument, or its r alias. Certain readers, such NISAR, GDAL, and ISCE2 under the right circumstances, are able to determine the rest of the necessary information from the data product itself. The command
qed -r=nisar.rslc rslc.h5
opens rslc.h5 with the NISAR RSLC reader and registers the datasets it contains. Similarly,
qed -r=native.gdal dem.tiff
will load dem.tiff using GDAL as the reader. Similar considerations apply to ISCE2 datasets that have an XML companion file. The three ISCE2 readers, isce2.slc, isce2.int, and isce2.unw, can parse the companion file and extract pixel type and shape information.
The pixel type specification can happen in two slightly different ways. The cell command line argument accepts a value directly using the standard mnemonics:
signed integers: char, int8, int16, int32, and int64
unsigned integers: uint8, uint16, uint32, and uint64
floats: float and double, as well as float32, float64, real32, and real64
complex numbers: complex64 and complex128, as well as cfloat32 and cfloat64
Alternatively, one can use a set of flags to set the pixel type directly:
signed integers: int8, i1, or integer*1; int16, i2, or integer*2; int32, i4, or integer*4; and int64, i8, or integer*8
unsigned integers: uint8, u1, byte or b1; uint16, u2, or b2; uint32, or u4; and uint64 or u8
floats: float, r4, or real*32; and double, r8, or real*8
complex numbers: complex64 or c8; and complex128 or c16
where the less obvious aliases are borrowed from mdx. So the commands
qed -r=native.flat -cell=complex128 c16.dat
and
qed -r=native.flat -c16 c16.dat
both open the file c16.dat using the native.flat reader and specify that the image pixels are double precision complex values.
The image shape can be specified using the shape command line argument that takes a pair of values separated by a comma. The command
qed -r=native.flat -c16 -shape=(1160,1400) c16.dat
tells the native.flat reader to expect an image whose height is 1160 lines and whose width is 1400 samples. Alternatively, one can specify the lines and samples separately:
height: lines, rows, l
width: samples, columns, cols, s
One can mix and match these, as the regular pyre priority rules for configuration events are applied. If the command line does not provide enough information for the shape to be determined, qed will issue a warning and suggest plausible values for the shape.
There is now a way to register datasets entirely from the command line.
qed
requires four pieces of information:--reader
--cell
--shape
The reader is specified with the
reader
command line argument, or itsr
alias. Certain readers, such NISAR, GDAL, and ISCE2 under the right circumstances, are able to determine the rest of the necessary information from the data product itself. The commandopens
rslc.h5
with the NISAR RSLC reader and registers the datasets it contains. Similarly,will load
dem.tiff
using GDAL as the reader. Similar considerations apply to ISCE2 datasets that have an XML companion file. The three ISCE2 readers,isce2.slc
,isce2.int
, andisce2.unw
, can parse the companion file and extract pixel type and shape information.The pixel type specification can happen in two slightly different ways. The
cell
command line argument accepts a value directly using the standard mnemonics:char
,int8
,int16
,int32
, andint64
uint8
,uint16
,uint32
, anduint64
float
anddouble
, as well asfloat32
,float64
,real32
, andreal64
complex64
andcomplex128
, as well ascfloat32
andcfloat64
Alternatively, one can use a set of flags to set the pixel type directly:
int8
,i1
, orinteger*1
;int16
,i2
, orinteger*2
;int32
,i4
, orinteger*4
; andint64
,i8
, orinteger*8
uint8
,u1
,byte
orb1
;uint16
,u2
, orb2
;uint32
, oru4
; anduint64
oru8
float
,r4
, orreal*32
; anddouble
,r8
, orreal*8
complex64
orc8
; andcomplex128
orc16
where the less obvious aliases are borrowed from
mdx
. So the commandsand
both open the file
c16.dat
using thenative.flat
reader and specify that the image pixels are double precision complex values.The image shape can be specified using the
shape
command line argument that takes a pair of values separated by a comma. The commandtells the
native.flat
reader to expect an image whose height is 1160 lines and whose width is 1400 samples. Alternatively, one can specify the lines and samples separately:lines
,rows
,l
samples
,columns
,cols
,s
One can mix and match these, as the regular
pyre
priority rules for configuration events are applied. If the command line does not provide enough information for the shape to be determined,qed
will issue a warning and suggest plausible values for the shape.