2

This is probably a obvious question, but i cant find a direct answer anywhere.

I installed, docker and docker machine, no problems so far. After that, i pulled this imaged based on jupyter/datascience-notebook docker. Supposedly, the only diffences is that it has Open CV installed.

Now, this is my first time using docker. I run the jupyter notebook using this command on bash:

docker run -d -p 8888:8888 -v $(pwd)/WD:/notebook dash00/datascience-notebook-opencv

This succeed in loading the jupyter notebook, it loads packages and works for every python command i try. For example, if a define functions and try them, they work.

The problem is that when i try to load data, for example: 

Import pandas as pd
data=pd.read_csv("/home/mario/WD/test.csv")

i get the following error: "OSError: File b'/home/mario/WD/train.csv' does not exist" the same happens when i try to load any kind of file using any kind of package, including images using cv2

Am i referencing the path wrong? do i have to refer to a host machine?

MarioG
  • 51
  • 4

2 Answers2

1

From your docker command, you mounted /home/mario/WD into the container as /notebook (as a host volume). Therefore inside your container, you would use:

Import pandas as pd
data=pd.read_csv("/notebook/test.csv")
BMitch
  • 231,797
  • 42
  • 475
  • 450
  • @abdoulsn `$(pwd)` expands to the current directory (assuming you are in a bash shell). The `pwd` command is likely short for print working directory. – BMitch Oct 01 '20 at 23:16
1

The -v $(pwd)/WD:/notebook argument is mounting the $(pwd)/WD directory on your host to /notebook in your container. If the directory that you issued docker run … from is /home/mario, then you would find /home/mario/WD/test.csv at /notebook/test.csv in the container.

Import pandas as pd
data=pd.read_csv("/notebook/test.csv")

You can also mount more directories from your host into the container as appropriate.

King Chung Huang
  • 5,026
  • 28
  • 24
  • No cigar. i tried 'docker run -d -p 8888:8888 -v $(pwd)/WD 05c3 ' and problem persist. Any suggestion? perhaps some specific documentation on jupyter docker you can point me? – MarioG Apr 24 '17 at 05:59