Backing up Pocket metadata from a Kobo

[faceless007]

With the announced Pocket shutdown, I want to find a way to archive all my saved articles. The provided Export tool on Pocket's website only includes links, not the saved text of all the articles. But thanks to the Pocket integration with Kobo, I had previously downloaded all my thousands of saved articles (including text) onto my Kobo Libra itself. And now I want to try save as much as I can from the Pocket integration.

Plugging my Kobo into my computer, it looks like the articles themselves are stored at the path G:\.kobo\articles, each in a directory with a random number, that includes an index.html file with the text. But this doesn't include any of the metadata (article title, URL) so it's impossible to navigate. Does anyone know anything about how Pocket stores its metadata on the reader and is there a way to copy it?

[PeterT]

I think you'll find all the data in the KoboReader.sqlite database. Check the content table, especially the rows with MimeType=application/x-kobo-html+pocket

[faceless007]

Oh thanks for the tip! I cracked open the sqlite3 file and found 5119 rows matching that MimeType. But when I look at the schema for the content table, I don't see an obvious way to map a row to the numbered folders containing the text. There has to be some way the Kobo is looking up which folder to load for an article, is it somewhere else in the database? These are the columns on "content":

ContentID, ContentType, MimeType, BookID, BookTitle, ImageId, Title, Attribution, Description, DateCreated, ShortCoverKey, adobe_location, Publisher, IsEncrypted, DateLastRead, FirstTimeReading, ChapterIDBookmarked, ParagraphBookmarked, BookmarkWordOffset, NumShortcovers, VolumeIndex, ___NumPages, ReadStatus, ___SyncTime, ___UserID, PublicationId, ___FileOffset, ___FileSize, ___PercentRead, ___ExpirationStatus, FavouritesIndex, Accessibility, ContentURL, Language, BookshelfTags, IsDownloaded, FeedbackType, AverageRating, Depth, PageProgressDirection, InWishlist, ISBN, WishlistedDate, FeedbackTypeSynced, IsSocialEnabled, EpubType, Monetization, ExternalId, Series, SeriesNumber, Subtitle, WordCount, Fallback, RestOfBookEstimate, CurrentChapterEstimate, CurrentChapterProgress, PocketStatus, UnsyncedPocketChanges, ImageUrl, DateAdded, WorkId, Properties, RenditionSpread, RatingCount, ReviewsSyncDate, MediaOverlay, MediaOverlayType, RedirectPreviewUrl, PreviewFileSize, EntitlementId, CrossRevisionId, DownloadUrl, ReadStateSynced, TimesStartedReading, TimeSpentReading, LastTimeStartedReading, LastTimeFinishedReading, ApplicableSubscriptions, ExternalIds, PurchaseRevisionId, SeriesID, SeriesNumberFloat, AdobeLoanExpiration, HideFromHomePage, IsInternetArchive, titleKana, subtitleKana, seriesKana, attributionKana, publisherKana, IsPurchaseable, IsSupported, AnnotationsSyncToken, DateModified, StorePages, StoreWordCount, StoreTimeToReadLowerEstimate, StoreTimeToReadUpperEstimate, Duration, IsAbridged, SyncConflictType

[compurandom]

You can also download everying in your pocket account from https://support.mozilla.org/kb/expor...ur-pocket-list

[PeterT]

I believe you also need to select based on contenttype = 6.

ContentURL appears to be the original URL
ContentID is the name of the folder within .kobo/articles

[faceless007]

Originally Posted by PeterT

I believe you also need to select based on contenttype = 6.

ContentURL appears to be the original URL
ContentID is the name of the folder within .kobo/articles

Thanks for the tip! I was initially thrown because it seemed like a lot of rows had ContentId that did not match any of the folders in /articles, but then I realized I need to filter to articles I had still saved on my Kobo and not archived by including isDownloaded.

For anyone else wanting to do the same thing, this query run against KoboReader.sqlite will return metadata on any Pocket articles you have downloaded on your Kobo:

"SELECT ContentID, ContentURL, Title, Description, DateAdded, WordCount from content WHERE MimeType='application/x-kobo-html+pocket' AND ContentType = 6 AND isDownloaded = 'true';

[PeterT]

Glad it helped. Again how I wish davidfor was still here; he'd have been able to answer that without even thinking and could have explained the difference between type 6 and 9..