What is the best way to obtain captions and descriptions for most/all images at Wikimedia commons (I do not need the images themselves). I know there is API, but I would prefer to obtain the information in bulk.
For example, when I look at https://commons.wikimedia.org/wiki/File:Schloss_Schönbühel_20180919.jpg, I see at least a German and English summary. I assume this is in some table in a structured way. I looked at http://dumps.wikimedia.org/commonswiki/latest/commonswiki-latest-image.sql.gz, but this file seems to contain only basic image metadata (size, date, ...)
I am aware of Get depictions from Wikimedia Commons, it addresses a related but different problem: depicts are a recently introduced structured data feature and the answer explains how to obtain the data one-by-one using Wikibase APIs not in bulk in a dump.
-- UPDATE --
It seems that the information is contained in commonswiki-latest-pages-articles.xml.bz2 (one of the dumps at https://dumps.wikimedia.org/commonswiki/latest/) but has to be extracted from the semi-structured text of each page.