docker build fails about opensartoolkit HOT 19 OPEN

Hi Mneuw,

sorry for running into this. You are doing nothing wrong. Just usual clash of dependencies, especially when you are working with gdal...there are some dependency issues between gdal and nodejs/npm.
Unfortunately I am not having too much time to look into this right now.
Could you check if removing the gdal related stuff in the 8th RUN solves this?
i.e. delete those 2 entries: libgdal-dev python3-gdal

gdal should later then be automatically installed via pip (so no additional entry needed).

Please let me know how that goes. And I hope that solves it.
Cheers

from opensartoolkit.

Finally got it to run:)

needed to modify the dockerfile quite a bit:

1. updated to OTB Version 7.2.0 because wget couldn't find the old one? you think this could cause problems?
2. removed libgdal-dev but needed python3-gdal
3. added a RUN to add nodejs >= v.10 (was required for installation)
4. added pip update > maybe this is already newest version and not necessary
5. removed last line jupyter nbextension enable --py widgetsnbextension because of an error. are these extensions important?

FROM ubuntu:18.04

LABEL maintainer="Petr Sevcik, EOX"
LABEL OpenSARToolkit='0.10.1'

set work directory to home and download snap
WORKDIR /home/ost

copy the snap installation config file into the container
COPY snap7.varfile $HOME

update variables
ENV OTB_VERSION="7.2.0"
TBX_VERSION="7"
TBX_SUBVERSION="0"
ENV TBX="esa-snap_sentinel_unix_${TBX_VERSION}_${TBX_SUBVERSION}.sh"
SNAP_URL="http://step.esa.int/downloads/${TBX_VERSION}.${TBX_SUBVERSION}/installers"
OTB=OTB-${OTB_VERSION}-Linux64.run
HOME=/home/ost
PATH=$PATH:/home/ost/programs/snap/bin:/home/ost/programs/OTB-${OTB_VERSION}-Linux64/bin

install all dependencies
RUN groupadd -r ost &&
useradd -r -g ost ost &&
apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -yq
python3
python3-pip
git
python3-gdal
libspatialindex-dev
libgfortran3
wget
unzip
imagemagick
npm
curl

RUN apt-get update &&
curl -sL https://deb.nodesource.com/setup_10.x | bash - &&
apt-get -y install nodejs &&
ln -s /usr/bin/nodejs /usr/local/bin/node

RUN alias python=python3 &&
rm -rf /var/lib/apt/lists/* &&
python3 -m pip install -U pip &&
python3 -m pip install -U setuptools &&
python3 -m pip install jupyterlab &&
mkdir /home/ost/programs &&
wget $SNAP_URL/$TBX && \
chmod +x $TBX &&
./$TBX -q -varfile snap7.varfile &&
rm $TBX &&
rm snap7.varfile &&
cd /home/ost/programs &&
wget https://www.orfeo-toolbox.org/packages/${OTB} && \
chmod +x $OTB &&
./${OTB} &&
rm -f OTB-${OTB_VERSION}-Linux64.run

update snap to latest version
RUN /home/ost/programs/snap/bin/snap --nosplash --nogui --modules --update-all 2>&1 | while read -r line; do
echo "$line" &&
[ "$line" = "updates=0" ] && sleep 2 && pkill -TERM -f "snap/jre/bin/java";
done; exit 0

set usable memory to 12G
RUN echo "-Xmx36G" > /home/ost/programs/snap/bin/gpt.vmoptions

get OST and tutorials
RUN python3 -m pip install git+https://github.com/ESA-PhiLab/OpenSarToolkit.git &&
git clone https://github.com/ESA-PhiLab/OST_Notebooks &&
jupyter labextension install @jupyter-widgets/jupyterlab-manager

EXPOSE 8888
CMD jupyter lab --ip='0.0.0.0' --port=8888 --no-browser --allow-root

best
M

from opensartoolkit.

Hi,

responding to the numbers above:

1. not at all. Orfeo is needed during mosaicking. So 7.2. is fine. it needs to ship with otbcli_Mosaic function (which it does)
   2.ok
2. ok
3. shouldn't do a harm
4. ok, you might have visualization issues of the progress bar during download. but the download itself should still work

from opensartoolkit.

Thanks a lot, all clear now for me!

from opensartoolkit.

great, let us know if there are further issues and have a great christmas time.

from opensartoolkit.

ok!! I just tried the S1GRD_Batch-processing and encountered that the grds_to_ards-module doesn't work for the RTC-Gamma0 product type. For some reason it produces only NA-layers. Contrary it runs perfectly for the GTC-Gamma product.
And I was wondering if and how it is possible to change to out_projection for the ard-products e.g. to WGS 1984 UTM Zone 33N?

Anyway it's a great toolkit! wish you a pleasent christmas 2020!!

Best Martin

from opensartoolkit.

hmm, in the docker file there is still the update of snap . uncomment following lines and build again. There are some issues with the latest version of snap that I still need to fix (when I have time)

RUN /home/ost/programs/snap/bin/snap --nosplash --nogui --modules --update-all 2>&1 | while read -r line; do
echo "$line" &&
[ "$line" = "updates=0" ] && sleep 2 && pkill -TERM -f "snap/jre/bin/java";
done; exit 0

from opensartoolkit.

then try to use this notebook as template:
https://github.com/ESA-PhiLab/OST_Notebooks/blob/master/4c%20-%20Sentinel-1%20GRD%20Batch%20-%20Timescan.ipynb

you need to change aoi, start,end and project_dir of course
for me this works.

Check out also the other notebooks. Should not take you more than a day to go through, and will hopefully help to get the concept. Most trickiest parts are the refine inventory and the ard parameters. the rest is just code execution

happy christmas to you, too!

from opensartoolkit.

and yes, you can set the out projection by using the EPSG code. so Lat/Lon is 4326, UTM 33N is 32633

e.g. s1.ard_parameters['single_ARD']['dem']['out_projection'] = '32633'

from opensartoolkit.

Now after getting back to SNAP version 7 I'm running into the following error while executing grds_to_ards in your above shared notebook:

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-26-b95cd587dd24> in <module>
      3     timeseries=True,
      4     timescan=True,
----> 5     overwrite=False
      6 )

/usr/local/lib/python3.6/dist-packages/ost/Project.py in grds_to_ards(self, inventory_df, timeseries, timescan, mosaic, overwrite, max_workers, executor_type)
    848         # time-series part
    849         if timeseries or timescan:
--> 850             grd_batch.ards_to_timeseries(inventory_df, self.config_file)
    851 
    852         if timescan:

/usr/local/lib/python3.6/dist-packages/ost/s1/grd_batch.py in ards_to_timeseries(inventory_df, config_file)
    146 
    147     # create all extents
--> 148     _create_extents(inventory_df, config_file)
    149 
    150     # update extents in case of ls_mask

/usr/local/lib/python3.6/dist-packages/ost/s1/grd_batch.py in _create_extents(inventory_df, config_file)
    186             fargs=([str(config_file), ])
    187     ):
--> 188         track, list_of_scenes, extent = task.result()
    189         out_dict['track'].append(track)
    190         out_dict['list_of_scenes'].append(list_of_scenes)

/usr/local/lib/python3.6/dist-packages/godale/_concurrent.py in result(self)
    154         """Return func result or raise Exception."""
    155         if self._exception:
--> 156             raise self._exception
    157         else:
    158             return self._result

AttributeError: 'DataFrame' object has no attribute 'crs'

I changed the crs of the inventory_pandas_dataframe o the same as the output projection with no effect.First results (ard products in snap-format) appear in the processing directory but timeseries and timescan products fails.

may you have an idea where this could come from? steps before while querring the S1-datasets I already encounter some crs-warnings:
` INFO (10:00:32): Created project directory at /home/ost/shared
INFO (10:00:32): Downloaded data will be stored in: /home/ost/shared/download.
INFO (10:00:32): Inventory files will be stored in: /home/ost/shared/inventory.
INFO (10:00:32): Processed data will be stored in: /home/ost/shared/processing.
INFO (10:00:32): Using /home/ost/shared/temp as directory for temporary files.

/usr/local/lib/python3.6/dist-packages/pyproj/crs/crs.py:53: FutureWarning: '+init=:' syntax is deprecated. ':' is the preferred initialization method. When making the change, be mindful of axis order changes: https://pyproj4.github.io/pyproj/stable/gotchas.html#axis-order-changes-in-proj-6
return _prepare_from_string(" ".join(pjargs))
/usr/local/lib/python3.6/dist-packages/pyproj/crs/crs.py:294: FutureWarning: '+init=:' syntax is deprecated. ':' is the preferred initialization method. When making the change, be mindful of axis order changes: https://pyproj4.github.io/pyproj/stable/gotchas.html#axis-order-changes-in-proj-6
projstring = _prepare_from_string(" ".join((projstring, projkwargs)))
/usr/local/lib/python3.6/dist-packages/geopandas/geodataframe.py:422: RuntimeWarning: Sequential read of iterator was interrupted. Resetting iterator. This can negatively impact the performance.
for feature in features_lst:

from opensartoolkit.

Hi,

it is a problem that we do not fix the version of certain python packages. There have been huge changes to the pyproj lib. Almost all of the geolibs in python use this for projection related stuff, and the syntax changed. You could downgrade to Proj 4 somehow. But it will take me some time to fix all this, as at the moment I do not really have time for it.

Best,
Andreas

from opensartoolkit.

ok thanks for your answer! just let me know when there is an update. meanwhile I'll try on my own start.
Best, M

from opensartoolkit.

Hm, all steps are working now except the creation of the image bounds for the time-series processing. So I think we do not need huge changes. What does this actually? align the scences to same extent? and which script/function is responsible for the job?may you have another hint where I can start digging. sorry for interrupting.

Best,
Martin

from opensartoolkit.

Hi, this step is needed to calculate the common minimum bound of all images in the time-series, as their extent may vary (slightly).
I think it uses geopandas for that, which is based on pyproj, which again had those huge changes. So I assume something related to this. Here you can find an example of this problem (as many have this) https://gis.stackexchange.com/questions/348997/constant-future-warnings-with-new-pyproj

from opensartoolkit.

from opensartoolkit.

Yes, the apihub URL changed https://scihub.copernicus.eu/news/News00868
I still did not have the time do update the toolkit. I am working on a release with the new SNAP 8 and will include this.

from opensartoolkit.

from opensartoolkit.

Hi Martin,
so there is a solution I think without having to change the code.
You can set the base url in the search function of your class instance.
Let's say you initialized your Sentinel1Batch to s1 then:
s1.search(base_url='https://scihub.copernicus.eu/dhus/')

However, setting the base_url for DOWNLOAD from scihub is NOT possible. But you can use ASF for download, if you haven't done so yet. You can download 10 files concurrently and it keeps all of the archive, as opposed to scihub where you only have the last year as it is a rolling archive. It is much faster.

Hope that helps,
Andreas

from opensartoolkit.

from opensartoolkit.