Build error on Windows

[Maheshppibm]

Step 8/10 : RUN md5sum -c md5sums.txt # check file integrity ---> Running in 55153672e0a7 md5sum: 'assets/sentiment_BERT_base_uncased/vocab.txt'$'\r': No such file or directory md5sum: 'assets/sentiment_BERT_base_uncased/saved_model.pb'$'\r': No such file or directory md5sum: 'assets/sentiment_BERT_base_uncased/variables/variables.index'$'\r': No such file or directory md5sum: 'assets/sentiment_BERT_base_uncased/variables/variables.data-00000-of-00001'$'\r': No such file or directory md5sum: WARNING: 4 listed files could not be read

[bdwyer2]

Hi @Maheshppibm, the build is working for me locally and seems to be passing on Travis CI. Can you please provide more information about your environment?

[splovyt]

If it helps, you can avoid building the image yourself and directly run the following command in your terminal to start the service:

docker run -it -p 5000:5000 codait/max-text-sentiment-classifier

[xuhdev]

Can you make sure you haven't changed md5sums.txt to CRLF EOL? Some editors will automatically convert all line terminators to CRLF upon saving. You can revert back by running dos2unix md5sum.txt.

[Maheshppibm]

Hi Brendam,

I am also working locally, with -

Windows 10,

Docker 18.09.2.

Python 3.7.3

Let me know what else I need to do.

Thanks and Regards,

Mahesh Patil IBM Global Business Services - Global Delivery Ozone 2, SP Infocity, Saswad Road, Hadapsar, Pune 412 308, India Desk - +91 020 66722064 Email Id: [email protected]

From: Brendan Dwyer [email protected] To: IBM/MAX-Text-Sentiment-Classifier [email protected] Cc: Maheshppibm [email protected], Mention [email protected] Date: 30-05-2019 20:18 Subject: [EXTERNAL] Re: [IBM/MAX-Text-Sentiment-Classifier] Build error (#9)

Hi @Maheshppibm, the build is working for me locally and seems to be passing on Travis CI. Can you please provide more information about your environment?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

[ptitzler]

I was able to recreate the issue. The problem is not model specific.

[ptitzler]

Your git is most likely configured to automatically convert newlines in text files. When Docker copies the md5sums.txt file to the Docker file system that change is not reverted, triggering the file not found errors. You have several options to resolve the issue:

• As @xuhdev has suggested if you run dos2unix -n md5sums.txt md5sums.txt prior to starting the Docker build you should be good to go. (dos2unix is available in git bash)
• You can clone the repository git clone https://github.com/IBM/MAX-Text-Sentiment-Classifier.git --config core.autocrlf=input and override the config setting, which will preserve the original encoding.
• You can download the source code ZIP file from the GitHub repository, which avoids the text file encoding changes.
• You can change your git configuration.
• You can change the file encoding using your favorite editor.

[imv7]

This worked: git clone https://github.com/IBM/MAX-Text-Sentiment-Classifier.git --config core.autocrlf=input Should it be included as "good to know" workaround on Docker Hub?