To save students local installation of TE, I added a dockerized version which can be installed nicely behind a web proxy as Nginx or Apache. Separation of users is done by sessions, and the actual calculations for the topic models runs in a queue to avoid stalling of output.
It is not very elaborated yet, but works fine as a low-level introduction to Topic Modelling, as intended. Of course, its deployment should still be restricted to a small user base, maybe a closed teaching course or a similar context where one can expect a reasonable (< 30) user count.
If there is wider interest in this feature addition, I could submit the code by pull request.
To save students local installation of TE, I added a dockerized version which can be installed nicely behind a web proxy as Nginx or Apache. Separation of users is done by sessions, and the actual calculations for the topic models runs in a queue to avoid stalling of output.
It is not very elaborated yet, but works fine as a low-level introduction to Topic Modelling, as intended. Of course, its deployment should still be restricted to a small user base, maybe a closed teaching course or a similar context where one can expect a reasonable (< 30) user count.
If there is wider interest in this feature addition, I could submit the code by pull request.