There are four different docker images here. I am new to RASA and I want to install it as a docker container (on my Mac). In particular I want to install rasa-core, rasa-nlu + spaCy + sklearn, and work inside the container. Which image should I pull?
I tried to run docker run -it --rm 7c /bin/bash after pulling rasa-core. But gives me
Available options:
start commands (Rasa Core cmdline arguments) - Start Rasa Core server
train - Train a dialogue model
start -h - Print Rasa Core help
help - Print this help
run - Run an arbitrary command inside the container
Isn’t pulling a fresh ubuntu image and installing all necessary packages (core, nlu, …) more flexible?
i recommend you build your image. Look into the dockerfile of rasa_core and rasa-nlu and build your image with your necessities, because your chatbot gonna have specific configurations that the officials images dont have. I dont know if i help, but this is my experience with rasa.
Yeah. I built my own image starting from anaconda3 image.
Just in case someone faces the same problem, after you pull anaconda3 image and run it, don’t forget to apt install build-essential, otherwise pip may fail to install some rasa-nlu or rasa-core dependencies.
I’m new to python and Rasa, so I’m totally lost… I was hoping I could just use a docker container so I could concentrate on Rasa stories, intents, etc. and avoid learning all the environment details for now.
I prefer to create my docker image from scratch. I simply pull Ubuntu, Ruby/Python image and install everything manually, all necessary requirements, packages, gems, etc, for my programming purpose. Then I always take backup of the image I have created, since Docker is terribly buggy, and every time it crashes, your images/containers simply get lost. I simply restore (from my backup) everything within a couple of minutes. Also keep your files on your host machine, and mount your working directory when you create a docker container.
I don’t recommend RASA’s official docker image. Create your own.
If you are not familiar with Docker I strongly recommend you devote some time to start with Docker and use it to work with RASA. Thus you will keep clean you host machine and avoid from package/library/gem conflicts if you’re like me working on a lot of different projects, and experimenting with different environments and libraries.