Introducing snapista, a GPT wrapper for Python

Show Room
#22

Hello @AriJeannin and thanks for nice comment!

You can set the gpt options in the Graph run method:

def run(self, gpt_options=["-x", "-c", "1024M"]):

which is defined here: https://github.com/snap-contrib/snapista/blob/bd1209de84355f41538436d0af26c10014670561/src/snapista/graph.py#L378

Cheers,

Fabrice

1 Like
#23

Perfect, good to be in control of that!

Thank you!

#24

Hi mengdahl,

I try to install Snapista by using this [step](https://github.com/snap-contrib/snap-conda) but I get error.

image
image707×61 1.8 KB

if you have any suggestion let me know because I already installed snap package in my environment now I don’t know why I have to install. And also I checked the snapista package, it’s not available.

Greetings,
Karlmarx

Noisy interferometric SAR coherence
#25

@fabricebrito should be able to help.

#26

Seach for snapista at anaconda.org. You didn’t mention your platform. Snapista is only available for linux64, so on other OS’s you may need a docker container.

#27

Thank you gnwii. I am using windows, so I should use docker container?

greetings,
Karlmarx

#28

I’m getting 404 errors for the snapista installation guide and also for https://github.com/snap-contrib/snapista/blob/master/docs/installation.

I was able to install snapista using conda/mamba in Ubuntu 20.04.

You should be able to use either a Docker container or WSL (1 or 2 – each suitable for different use cases) Ubuntu 20.04.

#29

Installation - SNAPISTA works, I wonder why https://snap-contrib.github.io/snapista/installation/ doesn’t

#30

snapista relies on a linux installation of SNAP so, yes, for Windows/Mac users, containers are your best option. I suggest to use VS Code remote containers for the implementation phase and then containers for the exploitation phase.

#31

Hi Fabricebrito,

Thank you for your suggestion. I will try it with this way.

greetings,
Karlmarx

#32

Hi @fabricebrito ,
I obtain constantly the same error when executing all the cells of the notebook sar-calibration.ipynb (snapista/sar-calibration.ipynb at master ¡ snap-contrib/snapista ¡ GitHub) (locally or even when running the demo note on Binder) : Exception: Graph execution failed (exit code 1).
I have tried to read distinct files from several sources:

  1. Downloading the sar-calibration.xml file from think link
  2. Following the instructions detailed in this forum (g = read_file("…/graph.xml"))
  3. Local tiff file indicating formatName=“GeoTIFF-BigTIFF”

When executing the last cell containing the g.run() command (my interpretation is that we launch all specified operations of the graph), I obtain the Exception: Graph execution failed (exit code 1) error. Why does it happen? How can I solve it?
Thanks in advance for your time!

#33

Hello @iazpiroz
Did you update cell #3 ?
It sets the path to the Sentinel-1 GRD manifest:

path_to_manifest = 'path to manifest'

#34

sure!
This is what I have written:
path_to_manifest = ‘sar-calibration.xml’

g.add_node(
operator=Operator(
“Read”,
formatName=“SENTINEL-1”,
file=path_to_manifest,
),
node_id=“read”,
)

#35

Oh I see, you need to download a Sentinel-1 GRD product, unzip it, and provide the path to the manifest file, just like you would do with SNAP Desktop in the Graph Builder section

#36

I have downloaded a zip from Copernicus Open Access Hub indicating Product Type GRD, and I have launched the following code:

g = Graph()
path_to_manifest = ‘S1A_IW_GRDH_1SDV_20211103T061756_20211103T061821_040403_04CA10_E94D.SAFE/manifest.safe’

g.add_node(
operator=Operator(
“Read”,
formatName=“SENTINEL-1”,
file=path_to_manifest,
),
node_id=“read”,
)
g.add_node(
operator=Operator(“Apply-Orbit-File”, continueOnFail=“true”),
node_id=“apply-orbit-file”,
source=“read”,
)

g.add_node(
operator=Operator(
“Remove-GRD-Border-Noise”, borderLimit=“2000”, trimThreshold=“0.2”
),
node_id=“noise-removal”,
source=“apply-orbit-file”,
)

g.add_node(
operator=Operator(“Calibration”),
node_id=“calibration”,
source=“noise-removal”,
)

g.add_node(
operator=Operator(“LinearToFromdB”),
node_id=“linear”,
source=“calibration”,
)

g.add_node(
operator=Operator(
“Terrain-Correction”,
pixelSpacingInMeter=“20.0”,
demName=“SRTM 3Sec”,
mapProjection=“AUTO:42001”,
),
node_id=“terrain-correction”,
source=“linear”,
)

g.add_node(
operator=Operator(“Write”, file=‘result.tif’, formatName=“GeoTIFF-BigTIFF”),
node_id=“write”,
source=“terrain-correction”,
)
g.run(gpt_options=["-x", “-c”, “20000M”])

And still I obtain the error Graph execution failed (exit code 1). Any hints?
Thanks in advance!

#37

Can you share with us the resources you have to run the notebook?

#38

The launched code is summarized in the attached notebook file:
calibration.ipynb (3.9 KB)
And the folder containing the Sentinel 1 GRD data is available through this link .

Should I provide more resources?

#39

@fabricebrito What do you mean exactly with resources? Are you referring to my PC resources?
If not, could you give me a hint about how can I run the notebook demo_binning.ypynb on the Binder that is available in the repository snap-contrib/snapista with the Sentinel ! GRD image available in this link ?

Thank you very much in advance!

#40

This page is giving an error as 404: Not found

#41

Please read the full thread. The 404 error has already been mentioned, and that you can find snapista at anaconda.org.