@hbkunn Have you tried just using the Vagrant environment (see https://github.com/DigitalSlideArchive/HistomicsTK/blob/master/ansible/README.rst)? Usually that is the easiest way to get everything installed.

@manthey still I got this error

$ sudo python deploy_docker.py start
Pulling rabbitmq:management
Creating rabbitmq:management - histomicstk_rmq
Starting rabbitmq:management - histomicstk_rmq
Pulling mongo:latest
Creating mongo:latest - histomicstk_mongodb
Starting mongo:latest - histomicstk_mongodb
Pulling dsarchive/girder_worker:v0.1.3
Creating dsarchive/girder_worker:v0.1.3 - histomicstk_girder_worker
Starting dsarchive/girder_worker:v0.1.3 - histomicstk_girder_worker
Pulling dsarchive/histomicstk_main:v0.1.3
Creating dsarchive/histomicstk_main:v0.1.3 - histomicstk_histomicstk
Starting dsarchive/histomicstk_main:v0.1.3 - histomicstk_histomicstk
Traceback (most recent call last):
File "deploy_docker.py", line 774, in
containers_start(**vars(args))
File "deploy_docker.py", line 153, in containers_start
mongo=mongo, provision=provision, **kwargs):
File "deploy_docker.py", line 224, in container_start_histomicstk
client.start(container=ctn.get('Id'))
File "/usr/local/lib/python2.7/dist-packages/docker/utils/decorators.py", line 21, in wrapped
return f(self, resource_id, *args, **kwargs)
File "/usr/local/lib/python2.7/dist-packages/docker/api/container.py", line 383, in start
self._raise_for_status(res)
File "/usr/local/lib/python2.7/dist-packages/docker/client.py", line 174, in _raise_for_status
raise errors.APIError(e, response, explanation=explanation)
docker.errors.APIError: 500 Server Error: Internal Server Error ("{"message":"driver failed programming external connectivity on endpoint histomicstk_histomicstk (52556e15c1e4e3c33947e362925efd8a4e630aed865671d35a0c34bc16e8b098): Error starting userland proxy: listen tcp 0.0.0.0:8080: bind: address already in use"}")

@manthey I just wanna use it as a Python toolkit. Is there an alternative way to avoid messing around with Docker?

https://stackoverflow.com/questions/42102656/cython-compile-error-is-not-a-valid-module-name

according to this, I delete __ init__.py then it's installed successfully.

thank you very much!
It's very useful and easy to use!

I got this same issue.
I had to deactivate and then reactivate my virtualenv, to get the paths of cmake and cython correctly (the ones from my virtualenv). Then I had to clone the repository again (maybe to get rid of some cmake cache files?), remove the init.py file, and everything worked fine.

Is this workaround still up to date? I want to try it HistomicsTK but struggling to get it running at the moment. I tried to install in editable and non-editable mode. The installation in editable mode leads to some errors with girder and the latter has problems with Cython or CMake whatsoever.

Removing the init.py file and/or re-activating my virtualenv didn't help.

@cklat Did you try this in the Vagrant or outside of it? Did you use girder-install to install HistomicsTK (which installs histomicstk in editable mode)?

@manthey I tried to install it as a Python-toolkit, so not in Vagrant. I didn't use girder-install, I strictly followed the installation instructions on the doc page (https://digitalslidearchive.github.io/HistomicsTK/installation.html)

I first encountered the problem with girder for which I opened an issue in the large_image git as I assumed there was some problem with large_image and girder (see girder/large_image#300). But it turns out the installation of large_image is fine and that the installation of HistomicsTK seems to introduce conflicts between all those packages or so. Anyway, installing all the packages in editable or non-editable mode doesn't help and the workaround mentioned here doesn't help either, unfortunately.

Assuming that you installed large_image in non-editable mode, the instructions at https://digitalslidearchive.github.io/HistomicsTK/installation.html will work (I just tried them in a clean Ubuntu 18.04 docker).

If you see a cmake error (such as cmake not found), it is because cmake is installed via pip (if it isn't present), and gets placed at ~/.local/bin. The default ubuntu .profile bash script only adds this path to the PATH if it exists when you log in. If you create this directory, log out and log in, then it will work.

See this example: https://gist.github.com/manthey/3fd17ce76ef01243b976137570cd6628

Yes, I installed large_image in non-editable mode but I'm still not able to get it running. Installing HistomicsTK in editable mode importing will fail because girder could not be found. Installing HistomicsTK in non-editable mode gives me an cmake-related error something about Ninja, like so: #362 (the longer error output).

If it makes any difference, I'm trying to run it on osx. I tried to install cmake via pip and via brew. No difference.

I'm actually familiar with how things get installed on osx. Do you have a ~/.local/bin directory and is in your path? If not, can you create such a directory and then add it to your path. After doing so, does installing HistomicsTK in editable mode work?

No, I didn't have a ~/.local/bin directory. I tried to create one and add it to PATH. Re-Installed histomicstk in editable mode but still no success. Gives me an ModuleNotFoundError: No module named 'girder' error. Re-Installing it in non-editable mode results again in a cmake-related error, like before.

I also tried to give it try from beginning with a clean virtual env, that is installing openslide, then large_image and then again histomicstk. Still, no changes. In fact, nothing gets placed into the ~/.local/bin directory during any of those installations

I'm not familiar enough with OSX to help further. I think if you installed cmake and its dependencies, it would work. We have relatively short term plans to make HistomicsTK directly pip installable. If you can't get it working, can you use a docker deployment?

A colleague of mine reported similar error outputs on his windows machine, also something about girder is missing and other. He cannot get it run either, unfortunately.

I haven't worked with docker yet, so I'm not sure if I know how to use it and can give you a reasonable feedback.

If you are planning to release it via pip the next couple days (maybe weeks), I will definitely try it then.