I am using DeepPavlov FAQ (http://docs.deeppavlov.ai/en/master/features/skills/faq.html) to answer FAQs.
I would like to use it as a multi-tenant service (rise), such that each tenant has its own set of FAQ, for example for all Apache Software Foundation projects:
How could I achieve this?
Than you very much for interest in our FAQ models!
Basically there are few steps to achieve your goals
- Make a copy of FAQ config file
https://github.com/deepmipt/DeepPavlov/blob/master/deeppavlov/configs/faq/tfidf_logreg_en_faq.json for each tenant.
- Alter them according to your needs. Provide training data file in the
- Train the models.
- Finally, you can run them as follows
python -m deeppavlov riseapi deeppavlov/configs/faq/tfidf_logreg_en_faq.json -d -p [port]
where [port] is a unique port number for each tenant.
More about REST API you can find here
Let me know if you need further assistance.
Thanks very much for your reply and instructions!
I was hoping that I could run it in one instance, using some unique ID for each tenant in the REST call.
But of course I can use a table to map the tenant ID onto a particular port.
All the best
Please share with us your progress Would love to try how the final system work.