Mobiperl Mobiperl - Perl tools for handling MobiPocket files

[HarryT]

Originally Posted by nrapallo

Sorry, it is in KBytes/s. I should have said 700KBytes/s and 100KBytes/s. I'll re-edit my previous posts to correct this!

In Mbit/s it is 6 Mbit/s download and 1 Mbit/s upload.

Here's a link to my Extreme internet connection service

That's more realistic . That's pretty similar to what most ADSL customers in the UK get - a nominal 8Mbit/sec download speed, although it's generally less due to contention. On a good day (eg early on a Saturday morning) I get about 6.5Mbit downloads - about 600 kBytes/sec. Right now (pretty much the busiest time of day) I'm downloading data which is coming in about 350 kBytes/sec, or about 2.7 Mbit/sec.

[Nick31]

I downloaded the MobiPerl tools today (version 0.0.36) and after a bit of playing around, got it to work with my 5.10 install of ActivePerl. Finding the Palm modules to install was the only difficult part, for those I had to use CPAN and compile them myself. In any case, mobi2mobi is working fine, and I expect to have no problem writing a quick perl script to wrap around this so I can convert the many .prc files I have into .mobi ones.


The first question I have (and this may have been answered elsewhere) is about getting the cover images to display in my converted .mobi files. mobi2mobi identifies the images and sets the offsets fine, but MobiPocket Reader doesn't show the covers. Is the problem that the images in the .prc files are bitmaps and not .jpeg images? Is there a way to get this to work (other than manually extracting each one from the many ebooks I have, converting the image, and then rebuilding the ebook)?


The other question involves using the --exthtype and --exthdata options. Is the type the number or the keyword? For example, the data from a newer .mobi shows:

EXTH item: 100 - Author - 11 - Wen Spencer
EXTH item: 101 - Publisher - 10 - Baen Books

What options would set the publisher?

--exthtype 101  --exthdata "Baen Books"

or

--exthtype Publisher  --exthdata "Baen Books"

Thanks again for putting the MobiPerl scripts together, they're very well done!

[tompe]

Originally Posted by Nick31

The first question I have (and this may have been answered elsewhere) is about getting the cover images to display in my converted .mobi files. mobi2mobi identifies the images and sets the offsets fine, but MobiPocket Reader doesn't show the covers. Is the problem that the images in the .prc files are bitmaps and not .jpeg images? Is there a way to get this to work (other than manually extracting each one from the many ebooks I have, converting the image, and then rebuilding the ebook)?

If you set the cover image offset correctly and the thumbnail offset then as I remember MobiPocket reader will display the images in "library" mode. But I do not think the image is displayed in the book if you do not have a link to the image in the actual book. But I might be mistaken since it was a couple of months ago I wrote and tested that code.

Originally Posted by Nick31

The other question involves using the --exthtype and --exthdata options. Is the type the number or the keyword? For example, the data from a newer .mobi shows:

EXTH item: 100 - Author - 11 - Wen Spencer
EXTH item: 101 - Publisher - 10 - Baen Books

What options would set the publisher?

--exthtype 101  --exthdata "Baen Books"

or

--exthtype Publisher  --exthdata "Baen Books"

There is a bug in one file concerning publisher. If you search for pubisher and then replace that with publisher then the flag --publisher will work. This fix will be in the next release.

--exthtype and --exthdata was added to be able to experiment if you knew what you do. It is the type name you have to use and for "Publisher" you should use "publisher". You have to look in MobiPerl/EXTH.pm and in the hash %typename_to_type to see what name to use.

Pleae tell me if you want to set something that is missing. The intention was to add specific flags for everything that people wanted to set.

[tompe]

Originally Posted by Nick31

I downloaded the MobiPerl tools today (version 0.0.36) and after a bit of playing around, got it to work with my 5.10 install of ActivePerl. Finding the Palm modules to install was the only difficult part, for those I had to use CPAN and compile them myself. In any case, mobi2mobi is working fine, and I expect to have no problem writing a quick perl script to wrap around this so I can convert the many .prc files I have into .mobi ones.

Which Palm module was it? And did the ActivePerl packages have the same name in 5.10 as in 5.8 (the names is on
http://www.ida.liu.se/~tompe/mobiperl/)?

[Nick31]

Let's see. I got Encode, Getopt-Mixed, Image-BMP, Image-Size, Palm-PalmDoc, and TimeDate from ActiveState's repository. GD and XML-Parser I got from Bribes. HTML-Format, P5-Palm, and Palm-Doc I got as source code off of CPAN and compiled. So yes, the names are the same for the most part.

Thanks for the info on the publisher fix and how to use the EXTH flags. (I suppose I could have looked in the code myself on those. ) I'll try some different PRC files and see if I can get the cover images to work. No luck so far, but I will figure this out one way or another!

[Jad]

Got a query regarding the coverimage function of Mobi2Mobi. When adding a coverimage to a PRC while converting to MOBI it appears when looking at the converted file in Mobipocket or the Cybook, that you get a double coverpage image. In the original there is a small thumbnail on the Page 1 TOC page, but after the convert I get the Coverimage on Page 0 and the thumnail becomes a full size image on the Page 1 TOC.

On another track and for purely selfish reasons, is it possible to add to your todo for Mobi2Mobi a way to extract the Thumbnail/Coverimage images as a flag option. I know the Mobi2HTML code does this and was wondering if the same code existed in M2M. I could leverage this to populate the Coverimage display in the Mobi2Mobi Windows GUI or add a thumbnail preview. I was looking at leveraging off of Mobi2HTML do to this, but if time permits.... =8)

[tompe]

Originally Posted by Jad

Got a query regarding the coverimage function of Mobi2Mobi. When adding a coverimage to a PRC while converting to MOBI it appears when looking at the converted file in Mobipocket or the Cybook, that you get a double coverpage image. In the original there is a small thumbnail on the Page 1 TOC page, but after the convert I get the Coverimage on Page 0 and the thumnail becomes a full size image on the Page 1 TOC.

Yes, this is expected behaviour, see the documentation. If a cover image is already there it will be replaced. That you see it on the TOC page is because there is a link to it in the HTML code. The Cybook displays the coverimage first and then the TOC page which has the link.

Are you really sure you get two images in Mobipocket Reader? I though the Mobipocket Reader did not display the cover image when opening the book.

On another track and for purely selfish reasons, is it possible to add to your todo for Mobi2Mobi a way to extract the Thumbnail/Coverimage images as a flag option. I know the Mobi2HTML code does this and was wondering if the same code existed in M2M. I could leverage this to populate the Coverimage display in the Mobi2Mobi Windows GUI or add a thumbnail preview. I was looking at leveraging off of Mobi2HTML do to this, but if time permits.... =8)

Maybe. Or maybe a flag to mobi2html to only explode the images? Is this enough? Or maybe it is better with a flag that saves the cover image or the thumb nail image in a specified file. That sound like a good functionality to have and it is very easy to add. I will add it tomorrow.

Let me know if you need some other flags that are easy to add...

[Jad]

I'm running Mobipocket 6.2 Beta and I definately get the initial coverpage then I next page to get the TOC. It looks like a double up, especially on the Cybook as the replaced thumbnail is oblivously the same size now as the coverimage. Its not a problem, JSWolf noticed while testing the GUI as did I, but I'm happy that its normal behaviour.

Adding an extraction flag would be great. If you could either extract the Thumnail/Coverimage to a supplied name each or just dump them out using the original filename with an appropriate identifier _TN / _CI that would be awesome.

[JSWolf]

But can that bug be fixed? Having two cover images the same size is not the way the cover is supposed to work.

[HarryT]

Originally Posted by JSWolf

But can that bug be fixed? Having two cover images the same size is not the way the cover is supposed to work.

It's not really a "bug", Jon, but the way the book is made.

With a Mobi book, the cover is specified as an entry in the "Guide" part of the OPF file (the "Guide" section is what controls navigation links).

You only get a double image if you also display the cover by having an image link within the HTML of the book itself. Don't do this, and you won't get the double image.

[tompe]

Originally Posted by HarryT

It's not really a "bug", Jon, but the way the book is made.

With a Mobi book, the cover is specified as an entry in the "Guide" part of the OPF file (the "Guide" section is what controls navigation links).

You only get a double image if you also display the cover by having an image link within the HTML of the book itself. Don't do this, and you won't get the double image.

Exactly. And I really do not want to manipulate the html code in mobi2mobi so it is a bit hard to solve this problem. Also to see the cover image in FBReader you need the link. So my feeling is that you have to live with two images since you want program that does not display the actual cover image to work.

[JSWolf]

Why do we end up with two cover images when we have no cover images to start with? So from what I am seeing, mobi2mobi puts in the extra link and therein lies the bug.

But I am not doing this for FBReader. I am doing this for Mobipocket. If you want, why not made a switch for FBReader. On Mobipocket, it just looks silly to have two cover images. And yes, it does make it look like a bug under mobi2mobi.

[HarryT]

I thought I'd explained this, Jon. Let me have another go.

A normal "linked" image in a book is NOT a cover image - even if it appears on the first page of the book. A cover image is especially marked as such in the file. The "Cover Image" menu item in the Windows mobi reader will display it, and the library manager on the Gen3 will automatically display it as the book's "cover" in the library.

When you use "mobi2mobi" to add a cover image to a book, you are adding that special flag to the file which says "this image is a cover image". What the tool CAN'T do is go through the HTML code of your book and remove any OTHER references to that same image that it might find - and nor would you want it to, in many cases.

It's down to the book creator NOT to "link" the image that they want as a cover image within the normal HTML, but simply to add that image via the "Guide" section of the OPF.

With the older creation tools such as BD, there was no way to properly designate an image as being a cover image; all you could do with such tools was simply have an image link for the picture on the first page of the book. If you use "mobi2mobi" to "re-add" that same image as a cover image in such a book, you will end up with it duplicated - it'll be there once as a "real" cover image, and again as an image link within the HTML. In such a case, the correct solution would be to "explode" the book to its OEB source, edit the HTML to remove the image reference, then rebuild the book as a Mobi file with its "real" cover image.

[wallcraft]

Originally Posted by tompe

Exactly. And I really do not want to manipulate the html code in mobi2mobi so it is a bit hard to solve this problem. Also to see the cover image in FBReader you need the link.

I agree about not manipulating the HTML. However, FBReader not recognizing a cover image is a bug in FBReader. I'll report it and see if this can be fixed. The iLiad's MobiPocket Reader has the same bug though (no way to show the cover image). I thought the "standard" behavior was to only show the cover image on request, which makes adding the same image to the start of the HTML a good thing, but apparently there is no standard behavior in this case - so someone is going to be unhappy.

[JeffElkins]

Is there such a function?