modis.collect

ModisRawH5

Warning

only implemented for MOD/MYD 11/13 products!

Class representing HDF5 file containing raw MODIS data.

This class will create an HDF5 file and collect data from MODIS HDF files into it. This file can then be used for smoothing in a subsequent step.

__init__(self, files, targetdir, vam_product_code=None, interleave=False) special

Initialize instance ModisRawH5 class.

This creates an ModisRawH5 object. If the corresponding HDF5 file already exists, it'll be automatically linked to it. If not, the file will be created on the first update run. All HDF files in the files list will be collected.

The user needs to be make sure that files are of the same product, spatial extent and that temporal consistency is conserved!

To make sure the update workflow is functioning as intended, it's important that targetdir is set correctly. This way existing HDF5 files can be updated, and new ones created.

To select a specific subdataset, vam_product_code needs to be provided. If not, the defaults will be extracted (VIM / TDA/ TDT).

For VIM, 16 day composite products can be interleaved to form a synthetic 8 day product if both satellites (MOD & MYD) are present in files. The resulting HDF5 file will be named with MXD as product code.

Parameters:

Name	Type	Description	Default
files	List[str]	A list of absolute paths to MODIS raw hdf files to be processed	required
vam_product_code	str	VAM product code to be processed (default VIM/LTD)	None
targetdir	str	Target directory for raw MODIS HDF5 file	required
interleave	bool	Boolean flag if MOD/MYD 13 products should be interleaved	False

Exceptions:

Type	Description
ValueError	If other product than MOD/MYD 11/13 are provided.
AssertionError	If files from multiple products are provided (except interleave).
AssertionError	If duplicates are detected that can't be handled.
AssertionError	If vam_product_code is not supported.

create(self, compression='gzip', chunks=None)

Creates HDF5 file.

If the corresponding HDF5 is not found in the target directory, it's created. If no chunking scheme is specified using chunks, a generic one of (number rows // 25, 1) will be used where the rows represent the spatial dimension and the colums the temporal dimension.

Parameters:

Name	Type	Description	Default
compression	str	Compression for data (default = gzip).	'gzip'
chunks	Tuple[int]	Chunksize for data (tuple of 2 int; default = (rows//25, 1)).	None

Exceptions:

Type	Description
AssertionError	If chunks is not a Tuple containing 2 int.
HDF5CreationError	If creation of HDF5 file fails.

update(self, force=False)

Updates MODIS raw HDF5 file with raw data.

The files specified in __init__ get collected into the HDF5 file, which is either created before or already existed after a previous initialization. If a HDF file can't be read, the datapoints are filled using the internal nodata value.

Parameters:

Name	Type	Description	Default
force	bool	Force collect, using nodata for non-readable input files.	False

Exceptions:

Type	Description
AssertionError	If dates of files to be collected are not after the ones aleady contained in the file.
HDF5WriteError	If writing to the HDF5 file fails.
IOError	If process fails to read HDF input file and force = False.

Retruns

collected (List): List of collected HDF files.