ArchivesPortalEuropeFoundation / Topic-Detection

Using machine learning approaches for automatic topic detection in a multilingual environment
6 stars 0 forks source link

Dockerise the webapp #58

Closed fedenanni closed 2 years ago

fedenanni commented 2 years ago

@kstamatis This is still a work in progress, but whenever you have a moment would you or someone from the team be able to give it a quick test? (tagging @kerstarno so you are also in the loop)

I think now if someone clones the repo and switches to the add-docker branch, by running:

docker build --tag ape .
docker run -p 5000:5000 ape

should start the web-app in the container and it will communicate through port 5000.

To test it:

curl -v 127.0.0.1:5000/query\?lang=fr\&type=concept\&text=et\&n_res=10\&broad_entity_search=False\&boolean_search=False

This should give you back a html table as a response with some entries in French.

If this works fine I just need to work a bit on the Volumes part to add the large datasets (for the moment the web-app works in test-mode using only samples which are already in the repo because they are very light).

Closes #55

fedenanni commented 2 years ago

@kstamatis (and @kerstarno for context) I have looked into using volumes, however I need the large files when starting the api (not added later, after the build). So, unless I change it, I think it's easier if during build time I copy the files in the image (which might take a few mins) but then the API will work directly. So at this point I think I am done with this, if you can give me access to the new VM I can set everything up. Alternatively you can just put me in touch with the people maintaining it and I'll explain how to deploy it

fedenanni commented 2 years ago

@kstamatis I think i am done with this PR. I have added:

Let me know if you spot something else we should do as part of this PR otherwise I would: 1) merge it to dev. Update the dev container / interface on the VM and check that all works properly. 2) then I'll start addressing the other bugs and make specific tests. When we are satisfied with those I would then maybe have a first master version of the interface which is based on our docker container (maybe we can discuss about this with @kerstarno )