chryss / pygaarst Goto Github PK

In raster.py, NetCDF class with basic functionality: Data access, lat, long.

Landsat metadata parsing is applicable to all datasets that use USGS's badly documented ODL (object description language) files. These are at the very least Landsat (4, 5, 7, 8), Hyperion and ALI.

Implement ODL MTL file parser independently of Landsat and subclass for application to individual sensors. Also implement HyperionScene and ALIScene like LandsatScene.

p3readiness: need to be able to test better. Make test data for:

• Landsat
• HDF5
• HDF4

SMALL files, with all the metadata.

Trying to run the test suite using nose (https://nose.readthedocs.org/en/latest/index.html) but it cannot find the tests:
nosetests -m "test_*" -w tests/

Finally found how here: http://pytest.org/latest/nose.html
py.test

Of course, they all fail b/c the test data is missing, but hey, they run! 👍

Change NBR and NDVI methods to use, by default, the reflectances of the respective bands. Leave the option to use DN. Maybe add option to use radiance?

Get all code in shape for py3 readiness with pylint.

Support for MODIS swath data analogous to VIIRSHDF5.

Documentation with Sphinx for Milestone 0.5 features.

The pygaarst.raster.Landsatscene object should have properties that return the NDVI and the NBR (raw, that is, calculated from DN, not reflectance/radiance).

Only radiance is currently implemented.

Methods/properties on GeoTIFF class may be presuming the scene is located in the northern hemisphere. Check.

Cloning functionality: clone open GeoTiff -- same CRS, same geotransform, same resolution, pixel size (shape) -- but replace data.

Optional enhancement: save only rectangular subset -- enter top-left and bottom-right indices. Requires calculation of new geotransform.

Hi Chris,

I'm hoping to be able to read landsat XML files associated with the landsat 8 surface reflectance product. I'm wondering if you have plans of incorporating this?

Nice work so far!

I am having trouble on a new build of pygaarst. An older build on another computer is still functioning properly. The Landsatscene class does not seem to be working. When I try:

sc = raster.Landsatscene(base_dir + ls_folder + ls_file_folder)

I get:

Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "pygaarst/landsat.py", line 31, in __init__
File "pygaarst/usgsl1.py", line 59, in __init__
File "pygaarst/mtlutils.py", line 377, in parsemeta
File "pygaarst/mtlutils.py", line 404, in _parsemetastream
NameError: global name 'metadatafn' is not defined

Has anyone seen this before and know what my problem may be? I might should mention I don't have the HDF libraries installed at the moment, but I can't see why that would be an issue. NumPy, GDAL, and PyProj are installed. Thanks.

For the modapsclient requests listSatelliteInstruments, listCollections and listMapProjections, the tags (namespace prefixes) in the response aren't stable. xmlns:ax21 and xmlns:ax:27 have been seen. Instead of hardcoding those, they need to be retrieved from the root node before parsing.

This should be saved to HDF5 file. Selectable wavelength range. Selectable i, j block (slices). Save projection in metadata.

OGR is a pain to use to access features. Rewrite, including tests (expand).

N_GEO_Ref nodes are empty when layers are embedded (no external georeference files). This leads to AttributeError: 'NoneType' object has no attribute 'nodeType' from xml.dom.minidom.

GeoTIIFF object public attributes:

• projection (proj4)
• units
• X and Y meshgrids

private attributes

• geotransform
• top-left and bottom-right corner coordinates in length units (m, km, degrees)

@Property attributes:

• Lat and Long meshgrids (centroids or corners)

Store test files externally and provide link.

When parsing userblock: ExpatError: not well-formed (invalid token): line 1, column 4589.

Needs to support ordering. OpenSearch is maybe not as important.

p3readiness: Use pytest-cov to get coverage up.