Currently the GoogleOCRFormatter formatter fails if antx fails to transfer annotations for a particular page. What should happen in that case instead of a failure is that the page should just be taken as Google says it is, and low confidence, languages, confidence indexes, etc. are not computed for that page. This shouldn't prevent the next pages to be imported normally. One example where the bug happens is W1KG6164 (I'm not sure what page though)
Currently the
GoogleOCRFormatter
formatter fails if antx fails to transfer annotations for a particular page. What should happen in that case instead of a failure is that the page should just be taken as Google says it is, and low confidence, languages, confidence indexes, etc. are not computed for that page. This shouldn't prevent the next pages to be imported normally. One example where the bug happens is W1KG6164 (I'm not sure what page though)