issues: 128687346

This data as json

id node_id number title user state locked assignee milestone comments created_at updated_at closed_at author_association active_lock_reason draft pull_request body reactions performed_via_github_app state_reason repo type
128687346 MDU6SXNzdWUxMjg2ODczNDY= 723 Implement tensordot for xarray with dask support 15167171 closed 0     8 2016-01-26T00:57:23Z 2016-03-05T00:47:49Z 2016-03-05T00:47:49Z NONE      

I've started using X-ray to store responses from convolutional neural nets over different transformations of images (translation(x,y), rotation (radians), etc). So far its been very intuitive storing and transforming results, unfortunately much of my analysis requires the use of tensor dot products, where I can choose arbitrary dimensions over which to make a projection, or perform a correlation. While dask implements np.tensordot, xray does not.

One can implement a dot product manually by multiplying data arrays then summing over dimensions.

fitm = (da_response*da_model).sum('imageID').sum('x_translation').max('models')

but this ends up being very slow, as I imagine when dot products are implemented by numpy or dask, there is a fair amount of optimization going on.

I am relatively new to GitHub, and this project, would you have any advice on the best way to contribute this functionality? tensordot where in you can put in a list of dimension names in two dataarray over which to compute a sum product, using dasks implementation.

 
{
    "url": "https://api.github.com/repos/pydata/xarray/issues/723/reactions",
    "total_count": 0,
    "+1": 0,
    "-1": 0,
    "laugh": 0,
    "hooray": 0,
    "confused": 0,
    "heart": 0,
    "rocket": 0,
    "eyes": 0
}
   completed 13221727 issue

Links from other tables