The current thumb extractor returns the first image that appears on the page, which may not be the lead one.
Acceptance criteria
- implement a better solution to extract the article's lead image by leveraging Extension:PageImages and their API. e.g. https://en.wikipedia.org/w/api.php?action=query&prop=pageimages&titles=Berlin&pithumbsize=100&piprop=name|original|thumbnail
- fallback for no image, we set minimum criteria resolution for sharing. If no version matches that criteria, then we display no image. Same as for articles with no images.
- ensure that the displayed thumbnail has a correct size. Refer to previous investigation [1, 2] in T417099: [Share Highlights] Client-side image generation
Note: We will need Elena to help QA



