calibre message: Failed to read metadate from the following ...

[porphyry5]

Is it possible to have calibre add such metadata-less ebooks to its database and then manually create metadata for each such file?

The full story: I originally created these txt files with calibre from epub originals. I no longer have either the epubs or the opf files that calibre created. The ereader I used then, an iRiver StoryHD, used the standard directory tree structure to organize its files and made no use of the metadata at all. Unfortunately it died and I now have a Sony PRS300 which creates an elaborate database for file access from their metadata. For files that have no such metadata, the result is an indigestible mess.

[theducks]

Originally Posted by porphyry5

Is it possible to have calibre add such metadata-less ebooks to its database and then manually create metadata for each such file?

The full story: I originally created these txt files with calibre from epub originals. I no longer have either the epubs or the opf files that calibre created. The ereader I used then, an iRiver StoryHD, used the standard directory tree structure to organize its files and made no use of the metadata at all. Unfortunately it died and I now have a Sony PRS300 which creates an elaborate database for file access from their metadata. For files that have no such metadata, the result is an indigestible mess.

Adding a 'Jacket' by conversion OR using Modify EPUB Plugin will update to what is the current Library View metadata

[porphyry5]

Originally Posted by theducks

Adding a 'Jacket' by conversion OR using Modify EPUB Plugin will update to what is the current Library View metadata

I take it both those options require the original epub, right? I no longer have them.

I was thinking more along the lines of creating an opf file from scratch, using an existing opf as a template, but attempts thus far have been ineffectual. Apparently I need to do more than just modify the filename and author name entries in the opf file.

[theducks]

So far, no one has created a plugin that reads metadata values from a full path. Only the file name is parsed when Get Metadata from Filename mode is used.

[BetterRed]

Try this

Add a few of your text files to calibre, fix up any titles and author names that need fixing, convert the txt files into whatever format you want to use on the Sony (e.g. ePub) and send those files to the Sony.

I think the conversion will create an opf in the book folder, which is actually irrelevant in this context.

The Calibre Send should embed an opf in the ePub from the data it has in the metadata database, even if it is sparse

Hopefully your Sony will find that more to its taste.

BR

[porphyry5]

Originally Posted by theducks

So far, no one has created a plugin that reads metadata values from a full path. Only the file name is parsed when Get Metadata from Filename mode is used.

Are you saying that the metadata is available in the file itself? Its not in the body of the file then, but in the file header? That this is where calibre obtains the data to create the opf file?

[porphyry5]

Originally Posted by BetterRed

Try this

Add a few of your text files to calibre, fix up any titles and author names that need fixing, convert the txt files into whatever format you want to use on the Sony (e.g. ePub) and send those files to the Sony.

That is what I tried to do, but calibre will not accept them, issuing the message I quoted in my first post. I assumed this was because I had disposed of the opf files created by calibre at the same time that it created these txt files

Originally Posted by BetterRed

I think the conversion will create an opf in the book folder, which is actually irrelevant in this context.

Apparently those opf files are what the Sony uses to construct its database. If I send the Sony the entire file and directory structure that calibre creates, its perfectly happy. But if I send it what sufficed for the iRiver reader, a single directory containing only the calibre generated txt files, then the Sony uses only the filenames and dumps everything into category "Other". The Sony doesn't reference directory names at all.

Unfortunately, when I originally created these txt files with calibre, I kept only the txt files themselves, and dumped all the directories and opf files that calibre also generated, along with then source epubs and pdfs that I had calibre convert to txt. Apparently calibre includes no metadata at all in the actual txt files it creates (it puts that in the generated opf), and now calibre refuses to work with those bare txt files, won't load them into its library. If I could get it to do that I could use calibre's Metadata Edit functions to create what the Sony needs.

[kovidgoyal]

You can add txt files to calibre that dont have any metadata perfectly fine. That message is telling you calibre could not read metadata, not that adding books failed.

[JSWolf]

Yes, you can edit the metadata in Calibre after adding the text files. The you can convert to ePub and the metadata will be included in the ePub. You may want to use Sigil to edit the ePub if need be. If the ePub is Ok as is except for the ToC, there is now a ToC editor you can use.

[porphyry5]

Originally Posted by kovidgoyal

You can add txt files to calibre that dont have any metadata perfectly fine. That message is telling you calibre could not read metadata, not that adding books failed.

When calibre (version 0.9.12) produced the message "could not read metadata", I clicked OK, and the gui returned to showing the library, but none of the files had been added to it. However, the problem is solved, I opened the files in vim and rewrote them, and calibre accepted the rewritten files without comment.

And thank you very much for writing calibre.

[porphyry5]

Thank you all for your help, it turns out all that was needed was to rewrite the txt files. At least, this worked rewriting them from vim. I guess this recreated enough metadata in the file header to satisfy calibre's requirements. Now I can use calibre's Metadata Edit to make them good for my reader.