Princeton-CDH / ppa-nlp

Discovering patterns in poetry’s data with machine learning; software for use with Princeton Prosody Archive (PPA) full-text corpus
1 stars 0 forks source link

setup images for test set to be served locally #73

Closed mnaydan closed 1 month ago

mnaydan commented 1 month ago
rlskoeser commented 1 month ago

Here's where I documented the syntax I used before for running mogrify: https://github.com/Princeton-CDH/ppa-nlp/issues/39#issuecomment-2204495465 :

 mogrify -format jpg -resize 500 */*.TIF

This was for Gale content, hence the .TIF.

Maybe copy the image directories we want from tigerdata to local storage then run mogrify to resize them in place, replacing the existing file. Relative path and filenames need to be the same because they are referenced in the dataset loaded into prodigy. Mogrify isn't installed on cdh-prodigy1 yet, but let's add it.

On the cdh-ansible side of things:

laurejt commented 1 month ago
laurejt commented 1 month ago

The temporary fix has been pushed to cdh-ansible. That said we'll want to make a longer term fix that has ansible automatically copy (and correctly set permissions) the (possible transformed) images to the hosting server.