send to device without conversion?

[jacktanner]

I've got a bunch of PDFs that I've imported them into Calibre and cleaned up the metadata.

Now, I want to read them on my device, which supports PDFs natively, no conversion necessary.

If I use Calibre to send the PDFs to the device, the sending is very slow, much slower than just copying the files. Presumably this is because Calibre does some conversion on the files. How can I get Calibre to not do this conversion? I'd be happy if Calibre just renamed the files according to my customized renaming prefs.

The conversion is actually worse then slow... I have a 200 page PDF is presumably a combination of multiple smaller PDFs. The PDF that Calibre sends to the device over only contains the first of these smaller PDFs, i.e., 15 pages. Seems like a dataloss bug to me.

[kovidgoyal]

go to preferences->plugins->metadata writer plugins and disable the pdf metadata writer

[jacktanner]

thanks for the quick response. i had assumed that disabling that plugin would also leave the PDF sent to the device with no metadata. is there any way to keep calibre's author and title meatadata when sending to the device, but not perform any other conversion?

[kovidgoyal]

there is no other conversion. setting metadata in pdf is slow.

[jacktanner]

Ah, then there is definitely a dataloss bug. From my orig. message: I have a 200 page PDF that is a combination of multiple smaller PDFs. The PDF that Calibre sends to the device only contains the first of these smaller PDFs, i.e., 15 pages, not the full 200 page ebook. I can send the full ebook somewhere if you want to try to reproduce the bug; it's 7MB.

I wonder what makes setting PDF metadata so slow. Does Calibre first send the copy to the device, and then modify the metadata on the device? It might be faster to modify a local copy, then to send that.

[kovidgoyal]

WHat makes setting metadata so slow is the design of PDF. You have to read the entire PDF into memory, set the metadata and then write it out.

[jacktanner]

oh, that's a really regretful design decision

that pdf truncation bug also occurs when i "save to disk", not only when i "send to device". interestingly, the PDF size is still about 7MBs; those pages are still in it.

does Calibre use a third-party library to read in / write out PDFs?

exactly what metadata does Calibre write out in addition to author and title? does it actually try to set page numbers?

[Worldwalker]

As I understand it (note: I'm a user, not a dev) calibre is at the mercy of a third-party library when it comes to PDF handling, which gives us calibre users even more reasons than we already had to hate PDFs. One great thing about calibre, though, is that it has the fastest release cycle of any software I have ever used, including when I've been an alpha and beta tester (the real kind that gets credited in the manual, not the "public demo" kind) for several programs. Things that annoy me don't stay annoying for long. If the improvements continue at this rate, by the end of the year it'll wash the dishes, feed the fish, and change the kitty litter.

[Elfwreck]

Originally Posted by jacktanner

Ah, then there is definitely a dataloss bug. From my orig. message: I have a 200 page PDF that is a combination of multiple smaller PDFs. The PDF that Calibre sends to the device only contains the first of these smaller PDFs, i.e., 15 pages, not the full 200 page ebook. I can send the full ebook somewhere if you want to try to reproduce the bug; it's 7MB.

If the PDF is a "portfolio" (an option available in Acrobat 9) or has attachments, those are not likely to be viewable third-party PDF viewers. The solution may involve extracting the PDFs and then combining them into a single actual PDF, instead of one with attachments or portfolio sections.

Acrobat Pro can do this; I have no idea if there's free or third-party software that can. At worst, the separate docs should be printable to separate PDFs, but that's definitely the slow, troublesome way to get them out of the combined package, and print-to-PDF may result in image files instead of reflowable text.

[jacktanner]

Originally Posted by Elfwreck

If the PDF is a "portfolio" (an option available in Acrobat 9) or has attachments, those are not likely to be viewable third-party PDF viewers. The solution may involve extracting the PDFs and then combining them into a single actual PDF, instead of one with attachments or portfolio sections.

Thanks for your thoughts. If I open the original 200-page PDF in Acrobat Reader 9 or with Evince, it shows 200 pages. If I import it into Calibre, add metadata, and then save it to disk with metadata, it's truncated (in the number of pages, not in file size), even if I open it with Acrobat Reader 9.

[Starson17]

Originally Posted by jacktanner

Thanks for your thoughts. If I open the original 200-page PDF in Acrobat Reader 9 or with Evince, it shows 200 pages. If I import it into Calibre, add metadata, and then save it to disk with metadata, it's truncated (in the number of pages, not in file size), even if I open it with Acrobat Reader 9.

Reader 9 supports pdf portfolio format. A lot of 3rd party pdf software does not. It looks like the library Calibre uses for its pdf handling does not support pdf portfolio. Why not check if you have a pdf portfolio format document , and if so, convert it to a standard pdf? It sure sounds like that's your problem. Portfolio is a bunch of different formats stuck together into a single pdf container. You're losing all but the first one. You'd be doing people a favor if you can confirm that's your problem and report back.

[DoctorOhh]

Originally Posted by Starson17

You're losing all but the first one. You'd be doing people a favor if you can confirm that's your problem and report back.

He's not disagreeing with this theory. He is however saying that the process calibre uses to add metadata causes this pdf file to be changed in such a way that even Adobe acrobat is no longer able to open the full file.

[jacktanner]

Originally Posted by dwanthny

He's not disagreeing with this theory. He is however saying that the process calibre uses to add metadata causes this pdf file to be changed in such a way that even Adobe acrobat is no longer able to open the full file.

Yes, that's exactly right.

I'm happy to test the portfolio theory and report back. How can I tell if my PDF is a portfolio, and how can I convert the portfolio to a flat PDF? I'm on Windows and Linux. Alternatively, I can send the possibly-portfolio PDF to whoever wants to investigate it.

[DoctorOhh]

Originally Posted by jacktanner

Yes, that's exactly right.

I'm happy to test the portfolio theory and report back. How can I tell if my PDF is a portfolio, and how can I convert the portfolio to a flat PDF? I'm on Windows and Linux. Alternatively, I can send the possibly-portfolio PDF to whoever wants to investigate it.

If I were you I would investigate other tools to add the metadata to the pdf prior to adding them to calibre. I'm not a pdf fan when it comes to ebooks but I have seen other tools for pdf metadata editing mentioned in some of the threads. Google might shed some light.

Good Luck.

[Starson17]

Originally Posted by dwanthny

He's not disagreeing with this theory. He is however saying that the process calibre uses to add metadata causes this pdf file to be changed in such a way that even Adobe acrobat is no longer able to open the full file.

I realized that. He was reporting that Reader 9 (software that supports Portfolio) would not read it, so I thought it was worthwhile to confirm that it's a portfolio format pdf problem, and not something else. A portfolio problem seemed the most likely, but not the only possible problem. He's got the original file and I'd start by looking at properties of that file in Reader 9. Maybe it will tell him.

As to how to flatten it, Acrobat Pro is the only method I'm aware of. There's a 30 day free trial available (Windows).