adobe/NLP-Cube

Getting stuck at "Configuring tzdata"

Spaskich opened this issue · 5 comments

Spaskich commented

Describe the bug
I'm trying to build a docker image and use the NLP as a web service but it gets stuck on the "Configuring tzdata" phase where I have to choose my geographic area. The problem is that the console is unresponsive and I can't choose any of the given options. I've tried the whole process on multiple Windows machines running Docker for Windows and I always get the same result.

To Reproduce
Steps to reproduce the behavior:

  1. Clone the repository
  2. Got to 'repo-dir/docker'
  3. Execute docker build --tag nlp-cube:1.0 .
  4. Wait until it reaches the 'Configuring tzdata' phase

Expected behavior
I want to run the NLP in a docker container and use it as a web API for sentence splitting, tokenization, lemmatization, etc. According to the documentation, I should be able to do this by starting the server and accessing container:port/nlp?lang=en&text=test.

Screenshots
image

Desktop (please complete the following information):

  • OS: Windows 10 Pro 1909
  • Docker for Windows v.3.0.0(50684)
tiberiu44 commented

This seems to be docker related image issue. I'm currently away from my laptop and canot fix this. However, I think it can be easily solved by adding ENV DEBIAN_FRONTEND noninteractive right on the second line, in the Dockerfile. Could you please check if this works?

Spaskich commented

Yes, this helped, however there is another problem down the line. At the 'Cloning into dynet' phase it throws a 404 error because the eigen repository is no longer on Bitbucket.
I tried changing it to the official repository on GitLab but then it says:
Cloning into 'dynet'... abort: HTTP Error 302: The HTTP server returned a redirect error that would lead to an infinite loop. The last 30x error message was: Found

I'm guessing the changeset -r b2e267d needs to be changed too, but I can't seem to find the equivalent on GitLab.

tiberiu44 commented

Sorry for the delayed response. You could try using the pip dynet package. Just replace the whole DyNet installation with RUN pip install dynet
I don't guarantee that the package will include support for Intel's MKL, but at least it will allow you to build the docker image.

Spaskich commented

Thank you very much for your help! I manage to build it and run it successfully. I added RUN pip3 install dynet and had to install Flask and bs4 too. This is the final version of my Dockerfile if somebody needs it for future reference:

FROM ubuntu
ENV DEBIAN_FRONTEND noninteractive

# Installing build dependencies
RUN apt-get update && apt-get install -y build-essential automake make cmake g++ wget git mercurial python3-pip curl

# Preparing Python build environment
RUN pip3 install cython future scipy nltk requests xmltodict nose2

# Installing MKL library
RUN wget https://apt.repos.intel.com/intel-gpg-keys/GPG-PUB-KEY-INTEL-SW-PRODUCTS-2019.PUB && \
    apt-key add GPG-PUB-KEY-INTEL-SW-PRODUCTS-2019.PUB && \
    wget https://apt.repos.intel.com/setup/intelproducts.list -O /etc/apt/sources.list.d/intelproducts.list && \
    apt-get update && \
    apt-get install -y intel-mkl-64bit-2018.2-046

# Installing DyNET
RUN pip3 install dynet

# Prepare environment UTF-8
RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y locales
RUN sed -i -e 's/# en_US.UTF-8 UTF-8/en_US.UTF-8 UTF-8/' /etc/locale.gen && \
    locale-gen
ENV LANG en_US.UTF-8
ENV LANGUAGE en_US:en
ENV LC_ALL en_US.UTF-8

# Prepare cube

RUN mkdir /work && \
    cd /work && \
    git clone https://github.com/adobe/NLP-Cube.git

# Prepare notebook
RUN pip3 install jupyter
RUN pip3 install Flask
RUN pip3 install bs4

# Start notebook
CMD cd /work/NLP-Cube/cube/ && python3 webserver.py --port 8080 --lang=en --lang=fr --lang=de --lang=sk
  • 👍2
tiberiu44 commented

@Spaskich - glad to hear you got it working and thank you for sharing your solution.
If it's not too much trouble, could you do a pull request with the fix? We can do the copy-paste from our side, but it's your contribution and if we do it instead of you, it will not get reflected on GIT.

Thanks again,
Tibi

  • ❤️1