Specifically the column ImpRespBW does not appear to be correct.

Right now just cloning the repo and 'setup.py test' fails all over the place. The error messages make clear where the test data is expected to be found, but not where a user can download it from in order to put it there.

The link in the README about the sample ntf is helpful, but not complete. I also had to hunt on my own to find required test data at

• https://gwg.nga.mil/ntb/baseline/software/testfile/Nitfv2_1/scen_2_1.html
• https://geographiclib.sourceforge.io/html/geoid.html
• ftp://data.asc-csa.gc.ca/users/OpenData_DonneesOuvertes/pub/RCM/Montreal/RCM1_OK1049807_PK1049808_1_16M11_20191028_224339_HH_HV_SLC/

and I still can't find data for the Sentinel and RS2 tests.

It would be good if each test failure message would say like "The data needed for this test can be downloaded from http://..."

Even better, how about if there were a 'setup.py download' that automatically downloaded and organized all the required test data? (with a confirmation like 'This will download X gigabytes of test data into ...\Desktop\sarpy_testing, is that OK? [Y/n]' )

I need to extract the SICD XML string from a SICD file. I will also be using the pixels, so want to use the SICDReader to do it all. I assume I am using it correctly - I am effectively doing:
sicdXml = SICDReader("testImage.ntf").sicd_meta.to_xml_string().
The resulting string gave me an error in my proprietary code. Before attempting to use sarpy, I was using a proprietary sicd string reader. Upon examination, I noticed that the the top-level tag name had been changed from "SICD" to "SICDType" in the string returned by sarpy. Is this intentional? It would be convenient if to_xml_string() returned the XML string as it appears in the file.
Making a cursory glance at the rest of the returned string, I do not see any other element name substitutions.

After installing sarpy into a bare virtualenv, trying to open a SICD file using sarpy.io.complex.open(fname) fails with:

reader = sarpy.io.complex.open(input_file)
File "xxx/venv/lib/python3.6/site-packages/sarpy/io/complex/init.py", line 70, in open
modules = [sys.modules[names] for names in module_names if import(names)]
File "xxx/venv/lib/python3.6/site-packages/sarpy/io/complex/init.py", line 70, in
modules = [sys.modules[names] for names in module_names if import(names)]
File "xxx/venv/lib/python3.6/site-packages/sarpy/io/complex/csm.py", line 20, in
import h5py
ModuleNotFoundError: No module named 'h5py'

So csm wants h5py and even though I'm trying to read a SICD, this is tripping me up. Reasonable solutions are:

1. Just list h5py as a required dependency.
2. try/catch around reader module setup and ignore readers that we don't have dependencies for.
3. For CSM only import h5py after checking that the input is a H5 file by examining it and checking its magic number.

There is an issue with calculating the correct byte indexing for each CPHD PVP field during reads. The issues can be found in io/phase_history/cphd.py line 381.

Currently field_offset is calculated as the PVP offset as an 8 byte offset value; e.g. field_offset = 1 for the second PVP value in the vector. This issue can be fixed by multiplying line 381 by 8 for the correct byte offset.

current_offset = pvp_block_offset + pvp_offset + the_range[0]*vector_size + field_offset * 8

Using the following setup setting the third argument to 'True' should produce a 1.1 SICD according to the documentation, but it fails to change the NITF subheader fields 'FTITLE' and 'IID2' to comply with the 1.1 specification as follows:
SarPy SICDWriter setup:

 x = SICDWriter('newSicd.nitf',meta,True,False)

According to the 1.1 spec, FTITLE and IID2 should be prefixed with "SICD:"

NGA STANDARDIZATION DOCUMENT SENSOR INDEPENDENT COMPLEX DATA (SICD) Volume 2
File Format Description Document Specification of the placement of SICD data products in the allowed image file formats.
(2014-09-30) Version 1.1

From Table 3.2:

From Table 3.4:

These prefixes are removed with SICD 1.2, but 1.1 must have them to follow spec.

More details should be forth-coming in the near future. This is to support the RGIQE effort.

• We need a good naming scheme.
• These should be incorporated smoothly into the sphinx documentation.
• We need to get the sphinx documentation published - this is already a task...

Immediate documentation and example needs:
1 - basic data reading:
i.) - sarpy.io.complex basic complex data reading and associated tasks (presently sarpy_example.py)
ii.) - sarpy.io.product basic derived data reading and associated tasks
iii.) - sarpy.io.phase_history basic phase history data reading and associated tasks

2 - sarpy.consistency usage - simple script and command line example for each of:
i.) sicd_consistency
ii.) sidd_consistency
iii.) cphd_consistency

3 - sarpy.geometry
i.) point_projection examples
ii.) geocoords basics?
iii.) geometry_elements? - (presently polygon_example.py)

4 - basic file creation/writing:
i.) sarpy.io.complex conversion from one complex format to sicd
ii.) sarpy.io.product creation of derived products
iii.) sarpy.io.phase_history example for writing a cphd file

5 - command line utils (sarpy.utils) - capability summary and basic usage overview for each of:
i.) convert_to_sicd (beef this utility up too?)
ii.) create_product
iii.) create_kmz (beef this utility up too?)
iv.) nitf_utils
v.) cphd_utils - pretty feeble right now, anything to beef this up with?
vi.) review_class -> review_classification (this one is trivial)
vii.) sicd_des_head - should almost certainly be junked

The main targets for this are readers and writers for SICD, SIDD, and CPHD.

This could potentially be accompanied by allowing the readers to be opened as r+ format, so that the image data could potentially be modified. This is tied up in another issue, and may not even address the needs.

Can someone say whether the algorithm employed to compute a subregion in the aperture_filter() class, found in /sarpy/processing/, is equivalent to the digital spotlighting algorithm as implemented in MATLAB?

The algorithm is described in
https://www.spiedigitallibrary.org/conference-proceedings-of-spie/8746/1/SAR-digital-spotlight-implementation-in-MATLAB/10.1117/12.2020884.short?SSO=1

I am new to sarpy, but when I tried to project 2 points using ground_to_image it called _ground_to_image in the num_points < block_size section.

iters comes out of _ground_to_image as an int and that does not work in the resize command at the bottom of grount_to_image.

I was able to fix this locally by adding this line after the call to _ground_to_image

iters = numpy.full((num_points,), iters)

That will just replicate that value.

srp comes from the PVP parameters which are double precision floating point.

iarp comes from the XML string representation which has limited significant digits.

There are a few solutions:

• np.around(srp, digits=9)
• print more significant digits in the XML

I lean towards the first one. If someone doesn't use sarpy to create the CPHD (as I did) this could still fail.

Refactoring color subaperture image creation, and using this as a SIDD creation test case.

The Spacenet 6 data was from Capella Space's Airborne testbed. The data is available on AWS S3. Information on getting the data is available at
https://medium.com/the-downlinq/spacenet-6-expanded-dataset-release-e1a7ddaf030

I attempted reading the data with the CapellaReader and this data is missing the field delta_line_time in the image_geometry

Calculation of the size of the ENGDATA field in the ENGRDA TRE appears to be incorrect. I believe it depends on both ENGDATC, the count, as well as ENGDTS, the size. In other words, rather than this:

self.add_field('ENGDATA', 'b', self.ENGDATC, value)

I believe the calculation should be:

self.add_field('ENGDATA', 'b', self.ENGDATC * self.ENGDTS, value)

This fixed it for me locally. I can't find a spec for ENGRDA to refer to, but the Nitro ENGRDA source backs this up.

Trying to test some image handling code, I wanted to generate some fake SICDs with multiple NITF image segments. To do this I scraped the SICD metadata from an existing image and used it to writeout a new image with a different number of rows/cols. The metadata won't make sense, but this ought to work.

import numpy as np
import argparse

import sarpy.io.complex.sicd as sicd
import sarpy.io.complex as cf


parser = argparse.ArgumentParser(description='Fake SICD Generator')
parser.add_argument('InFile', help='Any SICD file to scrape metadata from')
parser.add_argument('OutFile', help='Output SICD')
parser.add_argument('Rows', type=int, help='Rows of output SICD')
parser.add_argument('Cols', type=int, help='Cols of output SICD')

args = parser.parse_args()

reader = cf.open(args.InFile)
meta = reader.sicd_meta

d1 = np.random.rand(args.Rows, args.Cols)
d2 = np.random.rand(args.Rows, args.Cols)
data = d1 + d2 * 1j
meta.ImageData.NumRows = args.Rows
meta.ImageData.NumCols = args.Cols
meta.ImageData.FullImage.NumRows = meta.ImageData.NumRows
meta.ImageData.FullImage.NumCols = meta.ImageData.NumCols
meta.ImageData.PixelType = 'RE32F_IM32F'

writer = sicd.SICDWriter(args.OutFile, meta)
writer.write_chip(data)
del writer

python make_multi_seg_image.py ~/sample-sicds/sicd_example_1_PFA_RE32F_IM32F_HH.nitf multi-seg.nitf 200100 2

WARNING:root:Got string input value of length 97 for attribute IID2 of class ImageSegmentHeader, which is longer than the allowed length 80, so truncating
WARNING:root:Got string input value of length 97 for attribute IID2 of class ImageSegmentHeader, which is longer than the allowed length 80, so truncating
WARNING:root:Got string input value of length 97 for attribute IID2 of class ImageSegmentHeader, which is longer than the allowed length 80, so truncating
WARNING:root:Got string input value of length 11 for attribute ILOC of class ImageSegmentHeader, which is longer than the allowed length 10, so truncating
WARNING:root:Got string input value of length 97 for attribute FTITLE of class NITFHeader, which is longer than the allowed length 80, so truncating
Traceback (most recent call last):
  File "make_multi_seg_image.py", line 29, in <module>
    writer.write_chip(data)
  File "/home/kjurka/git/kjurka/sarpy/sarpy/io/general/nitf.py", line 2515, in write_chip
    self.__call__(data, start_indices=start_indices, index=index)
  File "/home/kjurka/git/kjurka/sarpy/sarpy/io/general/nitf.py", line 2573, in __call__
    (this_inds[0], this_inds[2]))
  File "/home/kjurka/git/kjurka/sarpy/sarpy/io/general/base.py", line 1895, in __call__
    self._call(start1, stop1, start2, stop2, data_view)
  File "/home/kjurka/git/kjurka/sarpy/sarpy/io/general/base.py", line 1902, in _call
    self._memory_map[start1:stop1, start2:stop2] = data
ValueError: could not broadcast input array from shape (99999,2,2) into shape (0,2,2)
CRITICAL:root:Image segment 1 has only written 0 of 199998 pixels
CRITICAL:root:Image segment 2 has only written 0 of 204 pixels
Exception ignored in: <bound method AbstractWriter.__del__ of <sarpy.io.complex.sicd.SICDWriter object at 0x7f0bb1e595e8>>
Traceback (most recent call last):
  File "/home/kjurka/git/kjurka/sarpy/sarpy/io/general/base.py", line 1175, in __del__
  File "/home/kjurka/git/kjurka/sarpy/sarpy/io/general/nitf.py", line 2611, in close
OSError: The NITF file multi-seg.nitf image data is not fully written, and the file is potentially corrupt.
Image segment 1 has only written 0 of 199998 pixels
Image segment 2 has only written 0 of 204 pixels

In the CPHD spec, FXBWNoise is optional and required only if the optional parameters FXN1 and FXN2 are present in the PVPs. The precondition here is assuming they are all present for any FX domain CPHD files.

I am attempting to create a SIDD from an inmemory SICD. Creating a SIDD from an inmemory SICD seems to not work due to an issue in OrthorectificationIterator

new_meta=meta.copy()
apod_shape=np.shape(tImage3a)
new_meta.ImageData.NumRows=apod_shape[0]
new_meta.ImageData.NumCols=apod_shape[1]
new_reader=FlatSICDReader(new_meta, tImage3a)

ortho_method=NearestNeighborMethod(new_reader, index=0, complex_valued=True, apply_radiometric=None, subtract_radiometric_noise=False)

create_detected_image_sidd(ortho_method, '.', block_size=10, version=2, remap_function=High_Contrast())

I get the following error:
File "/opt/miniconda3/lib/python3.7/site-packages/sarpy/io/product/sidd_product_creation.py", line 159, in create_detected_image_sidd
remap_function=remap_function, recalc_remap_globals=False)
File "/opt/miniconda3/lib/python3.7/site-packages/sarpy/processing/ortho_rectify.py", line 2294, in init
if os.path.abspath(ortho_helper.reader.file_name) !=
File "/opt/miniconda3/lib/python3.7/posixpath.py", line 378, in abspath
path = os.fspath(path)
TypeError: expected str, bytes or os.PathLike object, not NoneType

setup.py specifies the minimum version of numpy to be 1.9, however the SICD converters do not work correctly with 1.9.

I created a RS2 SICD and the resulting xsd:dateTime fields are improperly formatted. They contain my local time zone in addition to the Z. As a result the XML does not pass the schema.

        <DateTime>2015-11-30T18:13:53.000000-0700Z</DateTime>

        <CollectStart>2008-05-04T01:30:48.413940-0700Z</CollectStart>

This looks to be caused by the datetime behavior change introduced in numpy 1.11.

The simplest fix is probably to change the minimum numpy version to 1.11. If 1.9 is still needed, then the timezone needs to be stripped out, maybe in serialize_plain?

More details to be provided.

Hi all.

If I have a SICD Meta object from sarpy e.g.

>>> type(my_meta)
<class 'sarpy.io.complex.sicd.MetaNode'>

I can convert it to a string:

>>> meta_string = str(my_meta)

or

>>> meta_string = my_meta.__str__()

or

>>> meta_string = my_meta.__repr__()

But I can't convert that string back to a MetaNode

>>> meta_from_string = sarpy.io.complex.sicd.MetaNode(meta_string) 
TypeError: object() takes no parameters

Is there a way to do this?

As more and more large imagery is kept in cloud object storage, is there any plan for sarpy to be able to access these resources via smart_open or otherwise? Specifically I'm interested in S3 support for reading SICDs, but looking at the reader implementation, the file path is passed down pretty deeply into things.

Recently, we upgraded from an older commit on GitHub to the latest SarPy (1.1.78) available from PyPI. For the most part, the transition was straightforward; however, pickling has presented an issue. For example:

>>> reader = SICDReader('foo.ntf')
>>> meta = reader.sicd_meta
>>> with open('foo.pkl', 'wb') as fid:
...     pickle.dump(meta, fid)
>>> with open('foo.pkl', 'rb') as fid:
...     meta_pkl = pickle.load(fid)
>>> meta_pkl
SICDType(**OrderedDict())

The deserialized object is a SICDType containing an empty dictionary. Am I doing something wrong here? Or does SICDType not support pickling? Thanks!

The corresponding error is not hugely useful, but the likely problem is that the parsing/structure of the TRE is not defined correctly.

Would it be possible to add sarpy to PyPI so that it can more easily be used as a dependency in other python projections, e.g. with version specificity etc.? https://pypi.org/

bradh@minutae:~/sarpy/docs/examples$ python3 ./fft_example.py 
File /home/bradh/Data/sarpy_data/nitf/sicd_example_1_PFA_RE32F_IM32F_HH.nitf is determined to be a SICD (NITF format) file.
ERROR:root:ImpRespBW (15.17029) must be <= DeltaK2 - DeltaK1 (15.17028)
ERROR:root:Issue discovered with Col attribute of type <class 'sarpy.io.complex.sicd_elements.Grid.DirParamType'> of class GridType.
ERROR:root:Issue discovered with Grid attribute of type <class 'sarpy.io.complex.sicd_elements.Grid.GridType'> of class SICDType.
compute the fft to display in range / polar azimuth
Traceback (most recent call last):
  File "./fft_example.py", line 15, in <module>
    cdata = ro.read_chip((0, ro.data_size[0] - 1, 1), (0, ro.data_size[1] - 1, 1))
TypeError: unsupported operand type(s) for -: 'tuple' and 'int'

it looks like its opening the file OK, but that read_chip is returning an tuple of tuple: ((3975, 6724),).

It works better if I make this change:

diff --git a/docs/examples/fft_example.py b/docs/examples/fft_example.py
index 09f7c55..d7e79ae 100644
--- a/docs/examples/fft_example.py
+++ b/docs/examples/fft_example.py
@@ -12,7 +12,7 @@ fname = os.path.expanduser(os.path.join('~/Data/sarpy_data/nitf', 'sicd_example_
 ro = cf.open(fname)
 
 print("compute the fft to display in range / polar azimuth")
-cdata = ro.read_chip((0, ro.data_size[0] - 1, 1), (0, ro.data_size[1] - 1, 1))
+cdata = ro.read_chip((0, ro.data_size[0][0] - 1, 1), (0, ro.data_size[0][1] - 1, 1))
 
 cdata = cdata[0:1000, 0:1000]

I'm not familiar with the code base, so can't say if I'm fixing or working around the actual issue here. Let me know if a PR would help.

When I read a chip using slices, I get what I expect:

import sarpy.io.complex as cf

ro = cf.open(fn) #some .sicd file
ro[0:100,0:100].shape
(100, 100)

When I use the read_chip function without the 3rd parameter in each bound range defined, I would expect to get the same output as above:

ro.read_chip((0,100),(0,100)).shape
(100, 100)

Instead, I do not get any data back:

ro.read_chip((0,100),(0,100)).shape
(0, 0)

If I include the 3rd parameter, things work the way I would expect again

ro.read_chip((0,100,None),(0,100,None)).shape
 (100, 100)

I see the note under the _reorder_arguments function that says that if range = (int,int), then range = (stop, step), and I can get around this by just always making the 3rd argument in the bounds = 1, but this seems like unintuitive behavior.

Currently, the sarpy source code package does logging with the root logger. Something like -

logging.info("This is a comment")

In this current implementation I cannot explicitly set the log level for the sarpy application. If we define a custom logger for sarpy we can explictly tell 3rd party applications that use sarpy to use a specific log level in relation to sarpy logging.

My proposed change is to add some basic logging config to the top level __init__.py. Here is an example of something I use.

import logging

# Setup logging
logger = logging.getLogger(__name__)  # ** __name__ evaluates to "sarpy" here
logger.setLevel(logging.WARNING)

# create console handler and set level to debug
ch = logging.StreamHandler()

# create formatter
formatter = logging.Formatter("%(asctime)s [%(levelname)s] %(name)s: %(message)s")

# add formatter to ch
ch.setFormatter(formatter)

# add ch to logger
logger.addHandler(ch)

In the sarpy source code we can just create a logger like the below and use that anywhere we had previously used logging.XXX.

import logging
logger = logging.getLogger(__name__)  # ** __name__ evalues to "sarpy.my_module" here. The log is passed up the tree to the configured sarpy logger

logger.INFO("Some log")

And in 3rd party apps (notebooks, other src, etc.), we can change the log level of sarpy specifically using something similar.

import logging
import sarpy

logger = logging.getLogger("sarpy")
logger.setLevel(logging.ERROR)  # Set just the sarpy log level, not my current application

Thanks, and let me know if this is something I can help with a PR for.
Zach

With a clean install, the sarpy python modules don't seem to load properly. Please assist

$ git clone https://github.com/ngageoint/sarpy/
$ cd sarpy
$ python3 setup.py install --user
$ cd docs/examples
$ python3 sarpy_example.py
Traceback (most recent call last):
  File "sarpy_example.py", line 6, in <module>
    import sarpy.io.complex as cf
ModuleNotFoundError: No module named 'sarpy.io'

scipy.misc.comb is used in csk.py, radarsat.py, and sentinel.py under io/complex/utils, and maybe others. This scipy function has been deprecated since version 1.0.0 (documentation here).

The following solution (from here) is an alternative that requires no external dependencies:

from math import factorial
def comb(n, k):
return factorial(n) / factorial(k) / factorial(n - k)

Could you create a python package or at least a wheel for this so people can pip install? The ReadMe also gives very little guidance at the moment as to how to actually use the code.

Without a package, I tried to navigate into the io/complex and follow the "Example of how to use this package" steps, but I'm getting an error saying the following because I am trying to import from geometry, which is above my current root folder of io.

`>>> import complex

reader_object = complex.open('test.h5')
Traceback (most recent call last):
File "", line 1, in
File "complex/init.py", line 70, in open
modules = [sys.modules[names] for names in module_names if import(names)]
File "complex/radarsat.py", line 4, in
from .sicd import MetaNode
File "complex/sicd.py", line 8, in
from ...geometry import geocoords as gc
ValueError: Attempted relative import beyond toplevel package`

If I go up a level and import io.complex as complex, I'm getting the following error, which I don't understand as well.
`>>> import io.complex as complex

reader_object = complex.open('test.h5')
Traceback (most recent call last):
File "", line 1, in
File "io/complex/init.py", line 70, in open
modules = [sys.modules[names] for names in module_names if import(names)]
File "io/complex/radarsat.py", line 4, in
from .sicd import MetaNode
File "io/complex/sicd.py", line 6, in
from .utils import bip
File "io/complex/utils/bip.py", line 3, in
import numpy as np
File "/home/kate/miniconda3/envs/py27/lib/python2.7/site-packages/numpy/init.py", line 142, in
from . import add_newdocs
File "/home/kate/miniconda3/envs/py27/lib/python2.7/site-packages/numpy/add_newdocs.py", line 13, in
from numpy.lib import add_newdoc
File "/home/kate/miniconda3/envs/py27/lib/python2.7/site-packages/numpy/lib/init.py", line 23, in
from .npyio import *
File "/home/kate/miniconda3/envs/py27/lib/python2.7/site-packages/numpy/lib/npyio.py", line 14, in
from ._datasource import DataSource
File "/home/kate/miniconda3/envs/py27/lib/python2.7/site-packages/numpy/lib/_datasource.py", line 220, in
_file_openers = _FileOpeners()
File "/home/kate/miniconda3/envs/py27/lib/python2.7/site-packages/numpy/lib/_datasource.py", line 162, in init
self._file_openers = {None: io.open}
AttributeError: 'module' object has no attribute 'open'
reader_object = complex.open('test.h5')`

I'm excited to use the package once I get it up and running. Thanks!

One of the latest changes in the BaseChipper class introduced different set of arguments to __init__ member of the class. The subclass ICEYEChipper is still using old 'complex_type' named argument in its own __init__ method:

    def __init__(self, file_name, data_size, symmetry, complex_type=True, real_group='s_i', imaginary_group='s_q'):
        self._file_name = file_name
        self._real_group = real_group
        self._imaginary_group = imaginary_group
        super(ICEYEChipper, self).__init__(data_size, symmetry=symmetry, complex_type=complex_type)

This results with an error every time ICEYEChipper is used:

File "(...)\sarpy\io\complex\iceye.py", line 469, in init
super(ICEYEChipper, self).init(data_size, symmetry=symmetry, complex_type=complex_type)
TypeError: init() got an unexpected keyword argument 'complex_type'

I believe the correct version should be:

    def __init__(self, file_name, data_size, symmetry, transform_data='complex', real_group='s_i', imaginary_group='s_q'):
        self._file_name = file_name
        self._real_group = real_group
        self._imaginary_group = imaginary_group
        super(ICEYEChipper, self).__init__(data_size, symmetry=symmetry, transform_data=transform_data)

This fixed the problem for me, but I can't push to the repo.

I used setup.py build and install to have things properly installed under Ubuntu 16.04.4. I then tried to run the sarpy_example.py but I'm stuck in this error:

AttributeError: 'module' object has no attribute 'BufferedIOBase'

After some reading, it looks to be a conflict with the 'io' directory within 'sarpy':

https://stackoverflow.com/questions/32657580/class-gzipfileio-bufferediobase-attributeerror-module-object-has-no-attrib?noredirect=1&lq=1

I'm really not familiar with python and I don't know how to fix similar issues. It seems that renaming such 'io' directory to something else should solve the problem, but then all the code and package paths should be updated accordingly.

• Ensure the sphinx docs are getting built for the appropriate modules.
• Host the documentation at ReadTheDocs.

I have a SICD with with a RadarCollection.Area.Plane segment defined in the DES, but I get an error when I try loading the file using sarpy.
I created this example XML (modified from the SIX repo) which validates against the XML schema:
sicd_example_RMA_RGZERO_RE32F_IM32F_cropped_multiple_image_segments_PlaneArea-XML_DATA_CONTENT0.txt

xmllint --schema sarpy/sarpy/io/complex/sicd_schema/SICD_schema_V1_1_0_2014_07_08.xsd sicd_example_RMA_RGZERO_RE32F_IM32F_cropped_multiple_image_segments_PlaneArea-XML_DATA_CONTENT0.txt > xmllint.txt
sicd_example_RMA_RGZERO_RE32F_IM32F_cropped_multiple_image_segments_PlaneArea-XML_DATA_CONTENT0.txt validates

When I manually read this XML using sarpy, I get an error:

>>> from sarpy.io.complex.sicd_elements.SICD import SICDType
>>> with open('sicd_example_RMA_RGZERO_RE32F_IM32F_cropped_multiple_image_segments_PlaneArea-XML_DATA_CONTENT0.txt','rb') as f:
>>>      des_bytes = f.read()
>>> root_node, xml_ns = utils.parse_xml_from_string(des_bytes.decode('utf-8').strip())
>>> sicd_meta = SICDType.from_node(root_node, xml_ns, ns_key='default')
ERROR:root:Failed converting <Element '{urn:SICD:1.1.0}Plane' at 0x7f04263467c8> of type <class 
xml.etree.ElementTree.Element'> to Serializable type <class 
'sarpy.io.complex.sicd_elements.RadarCollection.ReferencePlaneType'> for field Plane of class AreaType 
with exception <class 'ValueError'> - Attribute coords of array type functionality belonging to class 
SerializableArray got a ElementTree element with size attribute 3, but has 0 child nodes with tag 
SegmentList.. Setting value to None, which may be against the standard.

It looks like if you change sarpy.io.complex.sicd_elements.RadarCollection.ReferencePlaneType._collections_tags to {'SegmentList': {'array': True, 'child_tag': 'Segment'}}, it seems to import.

Reading a SICD was pretty straightforward:

from sarpy.io.complex import open
sp = open(sicd_file)
meta = sp.sicd_meta
cimg = sp.read_chip(1, 1)

Is there an example for writing a SICD in the source code? Having trouble finding something like

import sarpy.io.complex.SICDWriter
SICDWriter(image=cimg, meta=meta, filename='out.sicd')

SICD type reader -> SIDD products
SICD type reader -> kmz products

Iterate over all the file contents of the directory

When trying to process a NITF with an ENGRDA TRE, I'm finding that the embedded Records fail to parse. Rather than starting at an offset into the TRE data (at ENGLN, 23 bytes in, after RESRC and RECNT), they're starting at the beginning (i.e. the RESRC) field and failing because because what's parsed is not an integer, as expected, but character data.

This is possibly a bug in TRELoop. Changing

 entry = child_type(value, *args, **kwargs)

to

 entry = child_type(value[loc:], *args, **kwargs)

fixed this for me locally.

Details for exactly what this actually looks like should be forthcoming in the near future.