Read Images on Azure Blob Storage

Krishna615
Level 2
Read Images on Azure Blob Storage

Hi,

We are trying to create images dataset from images stored on Azure blob storage. We have successfully established connection with the blob storage and able to view the list of images from a container while creating a dataset. 

Now when we select an image(png/jpg) from the list it shows that the format of the image is not supported - Failed to detect file format. Please manually fix. Is it that Dataiku doesn't support reading images from Azure blob storage?

If it does could someone please help on accessing the images from the blob?

If it doesn't could you please suggest an alternative to access images from Azure?

0 Kudos
3 Replies
fchataigner2
Dataiker

Hi,

if you're getting this error in the folder view, then this means DSS wasn't able to auto-detect the MIME type of the (image) file you selected. It can be that the files are corrupted, or the file extension isn't a simple .png or .jpg... In any case, this is only for viewing in the DSS UI, there should be nothing preventing you from accessing/retrieving the data from those files in your recipes or notebooks.

0 Kudos
Krishna615
Level 2
Author

Could you please elaborate on how to access the images stored in Azure blob storage in the recipes/notebooks?

I am looking for a way to create datasets based on the images imported from the blob storage and view them on UI as well(may not be important but required).

First let me ask you if there is a way to import images in Azure blob storage into the Dataiku.

0 Kudos
fchataigner2
Dataiker

the method I'm referring to is https://doc.dataiku.com/dss/latest/python-api/managed_folders.html#dataiku.Folder.get_download_strea... , and is usable from python recipes and notebooks.

Note that there are no particular facilities to work with images in the DSS UI, so you'll have to use notebooks to explore your data

0 Kudos