John's Blog

This week comes with some great new additions. It also comes with some great new challenges. Not to mention more work for next week.

eReader output is complete and working. The files produced are the same format as those produced by Dropbook (more on this in a bit). In addition to eReader output I’ve added metadata writing for eReader files.

There is one major issue I’ve come across concerning the eReader format. The files produced by Makebook are significantly different than those produced by Dropbook. The files I’ve been using for reverse engineering the format have all been produced by Dropbook. My implementation for eReader input only works with Dropbook format files. As such, Makebook produced files will not work and are currently not supported. They will be unsupported for the foreseeable future because of how different the format is to Dropbook, and because Makebook is not supported nor being developed. It was replaced by Dropbook some time ago.

The PDB container format also got a metadata writer. However, the PDB wrapper itself only supports setting the book title. So that’s all that gets written to PDB files that either don’t support metadata in their internal format or that don’t have their own specific metadata writer.

eReader wasn’t the only format to get some output work. FB2 output has been added as well. However, there is no metadata writer for it yet.

The final bit of work I did for this week was to add auto convert to sending by email in the GUI. It’s the same idea as auto convert for sending to a device. If the file that is being sent is not in a format that is accepted by the email address (this is configurable setting) the file will be auto converted to a suitable format before being sent.

Tags: calibre, eReader, fb2, GUI.

Device interfaces can now be configured in the GUI. Also, there is a simple framework for creating plugin configuration widgets.

I’ve added a metadata reader for the eReader format. However, eReader supports 3 ways to set the metadata in the file. 1) In the pdb header (only supports setting a short title). 2) In the metadata section of the file (supports the most information: title, author, publisher, copyright, isbn). 3) Embedded in the text as a comment. 2 and 3 are only accessible if the book does not contain DRM (or has been unlocked, but Calibre does not support this). 3 is not supported at all with this metadata reader. The reader first tires 2 then falls back to 1 if the book is DRMed or if the metadata section is non-existent.

Two new input and output formats have been added. ztxt and palmdoc. They are both pdb formats like eReader. For input the pdb input plugin will automatically determine the internal format and call the appropriate code path. For output the default is palmdoc but there is an option –format that can be used to change it to any other supported pdb output format (ztxt is the only other currently). The format option is also available in the conversion dialog in the GUI.

Speaking of conversion in the GUI. It now works. There are all new dialogs for single and bulk conversion. Pretty much anything that can be done using the command line ebook-convert can be done in the GUI. Bulk, single and auto conversion are all complete and working. Auto conversion will also honor a users preferences for formats set for the device interface plugin.

Tags: calibre, eReader, GUI, palmdoc, pdb, ztxt.

This weeks review of what I’ve been working on is a little late. Overall it wasn’t as productive as last week looking at what was accomplished but I spent just as much time coding as last. With projects like this you can’t judge output by the number of features add or bugs fixed.

The GUI received context sensitive treatment for the device menu. It will only have send to device, when a device is connected and send to card A and B will only enabled when they are available as well. A simple change but one that will reduce confusion.

I’ve spent a lot of time working with Lee Dolsen (ldolse from mobileread) on pdftohtml processing rules. They are nearly complete and the output is looking really good. I know I’ve been saying that for a while now but each week it just keeps getting better. However, PDF is still not an ebook format and should not be treated as such. This simply helps to get content out of the PDF format and into a more manageable one.

One big thing I spent most of my time this week on was eReader input. Yep, eReader pdb files can now be converted to any supported output format. Metadata reading of eReader files is not yet supported. That is on my todo list. The html it produces could probably use some work but that will come as people report issues once 0.6 is released.

The other big thing that has taken up my Calibre time is eReader output. Sadly, it does not work. Also, it will not be working for the foreseeable future. The issue I’ve run into is I don’t know enough about the format to produce a file that can be read by eReader’s reading software. The main problem I face is there are around 66 “sections” to the eReader format header (not the pdb header, this is record 0 of an eReader file). I know what 10 of those sections are and what values they should have as they are used for my reader. Around 40ish of the sections should have a value of 0. However, that leaves 26ish sections that I don’t know what they are, what they do or what value they should have and how it relates to the rest of the file. Suffice it to say until I know more about the format I won’t be able to complete the output plugin.

Oh, I did write an inspector script (it’s in the eReader directory in the Calibre source tree) to help understand the eReader format. If anyone is interested in analyzing the format they can use it to help them see what is in the header.

Tags: calibre, eReader, GUI, pdf.