Help with Docker Medcat service

MedCAT
1

Hey - trying to run MedCAT service in Docker but throwing an error.

My process so far:
(1) Image downloaded from cogstacksystems/medcat-service
(2) Run container directly using this image
(installed curl as command not available)
(3) curl -XGET http://localhost:5000/api/info
→ get an internal server error

Log shows the following:
File “/cat/medcat_service/nlp_processor/medcat_processor.py”, line 216, in _create_cat
raise ValueError(“Vocabulary (env: APP_MODEL_VOCAB_PATH) not specified”)
ValueError: Vocabulary (env: APP_MODEL_VOCAB_PATH) not specified

Is there a step I’m missing to define these?
Many thanks for help!

2

Hello,

I’ve pushed some fixes that have addressed this issue after the 1.4.0 upgrade.
Re-pull the latest image or the release image tagged “0.3.2”

3

Thanks for this update!

I’m now using 0.3.2.

Unfortunately I’m still getting the error with the steps above.

This is the full log: Starting up Flask app using gunicorn server ...[2022-10-21 21:33:28 +0000] [8] - Pastebin.com

I get this prior to the error “MedCatProcessor: APP_MEDCAT_MODEL_PACK not set, skipping…”
Is there a step I’m missing here, outside of loading the Docker image?

Let me know if I can help with any other information

4

Hello,

It seems like no models are being loaded in this case, did you execute the " ./models/download_medmen.sh" script (mentioned in the readme) ?

5

Figured out. Cloned the git and ran docker/run_example_medmen.sh which handled the model download and image creation.
Thanks for help!
BW

Apologies - I may need an “explain like I’m 5” on this as not that experienced with Docker.

I’ve pulled and loaded latest from Dockerhub and tried to execute this script from within the container - but the script isn’t present in the container.

Presumably then, I need to execute the script through creating a custom image?
At what point do I execute the script?
And do I need to edit any of the Docker files?

Thanks again in advance.

1 Like