Wrong covers imported, covers do not update manually

[AIndoria]

Essentially the title. When a goodreads csv import is done, wrong covers get uploaded. Eg:

Some are hilariously wrong.

When you edit the book and upload the cover manually via manual file upload (not the URL), it never changes. the Upload progress bar does work, but old cover persists.

[bayang]

The covers are cached on the browser side to preserve performances. Can you try to wait a bit and force refresh of the page on the book detail page for example ?

If you're used to the browser dev console you can have a look at the book json data, find the image name and try to see in your covers folders if the file is the right one. The book data is after the log line that says "called userbook + UUID"

I just tested and the old cover persisted a few seconds and a page refresh.

[aslmx]

First of all, thanks for Jelu. So far I like it really much. I want to use it to retire the Excel sheet my wife is using to keep track of her huge book collection.

I came here to also report horribly wrong covers when using bulk import with Auto-metadata. It seems to be scientific book covers mostly.

I'm currently trying to use Jelu API to get the ISBNs of around 1300 Books from an excel sheet. Then I want to import these ISBN using the bulk import. (Btw: there is a 70% success rate of fetching something via the API somehow).

This is a screenshot of one page of imported books, see how i even have the same wrong Cover as AIndoria Parable of the Sower and Systems design and Analysis

Is there some way to improve the Import results? Can cover-fetching be disabled somehow? I'd rather have no cover at all, than wrong cover.

I'm using the docker version with the example compose file (just changed the volumes) version 0.34.4.

[bayang]

@aslmx It is not possible at the moment to disable cover fetching for the ISBN list import, I will add it (so it will be the same as the goodreads import)

Concerning the covers that are automatically imported, I can't do anything about it, but no issues have been spotted before, maybe it is a recent change in the tool that is used to grab the metadata, I'll investigate. If you need to discuss further please create a new issue.

[zacharyasmith]

I can confirm the same results in my recent import of Goodreads CSV file. I downloaded the docker container today and was surprised to see nursing textbooks and business books replacing many of the novels I have been reading.

[bayang]

@zacharyasmith thanks for the report, that's weird, it all started recently. I need to investigate the tool that Jelu uses to fetch metadata and covers to see what's going wrong.

[zacharyasmith]

Thank you for looking into it. I don't have any time equity into the service just yet. I like the idea of this project a lot and think the implementation looks clean. Just some 🐞 🐛 🐜 to squash looks like.

[bayang]

I think I'll make a new release soon with a first attempt to fix this cover issue. We'll see if it works.

[zacharyasmith]

The metadata is also wrong. For some books, the description is correlated with the erroneous cover. Again it is only 30% or so.

Zach Smith

On Mon, Jan 9, 2023 at 1:07 PM, bayang @.***> wrote:

I think I'll make a new release soon with a first attempt to fix this cover issue. We'll see if it works.

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: @.***>

[bayang]

Finally I can't fix this wrong covers or metadata problem, we are getting rate limited by google see #59 . I leave this issue open for the manual update of covers not working though.

I'll probably add the possibility to search covers and metadata directly from the book page to make editing wrong or missing data easier.