Suggestions for improvement
The link added should be properly localized. As it is now, the link appears in one only language irrespectively of the interface language selected in the Users preferences. Compare with the link for exporting a PDF version, which changes if the user changes their language.
Parameters for flexibility
Hi! Thank you for this great tool. I have been considering how we can use it to the best effect at English Wikisource, and I have some suggestions that would allow us to integrate the book-generation process with the existing PediaPress PDF tool, as well as allowing more unconventional collections of pages.
Firstly, the PediaPress tool operates on a simple list of pages, like this: en:Wikisource:Books/Bull-dog Drummond. We could use ws-summary and metadata microformats on this page to generate the book, but it will include the raw pages list, as well as the formatted contents contained in the work itself. Thus, my first suggestion is to accept a parameter "ignorelistpage" which will direct WSexport to harvest the links in the list, but not include the page in the final document.
This manual (but with a page-collector helper) listing method allow us to have separate works group together (for example, a collection of related scientific papers), allows works that are not all linked to a single front page (like newspapers) and allows us to have a curated collection of verified (and editable) books.
Secondly, if the metadata is included on the list page, that allows flexibility in unusual works (such as a compilation of separate works by many authors). However, for the common case where the information on a single page is sufficient, this duplication of data invites mistakes to creep in when pages are updated, but not the related book. Therefore, I propose a "metadatalocation" parameter, which directs WSexport to harvest metadata from another page.
I am working on a way to integrate this into a system more targeted at WSexport than the PediaPress extension, but I would need to know exactly what format the script would take the parameters before it could be completed. Thank you for your great work! Inductiveload 03:06, 20 February 2012 (UTC)
- For the ignorelistpage, no problem, it's a good idea ! I'll implement it. For the metadatalocation it's a little bit more complicated but I think it's possible. Thanks for your interest. Tpt 20:32, 21 February 2012 (UTC)
- Wonderful! I feel the ignorelistpage is the most important, as that means any pagelist that is not part of the work itself can be used easily. I have talked to the mwlib/PediaPress list, and it is unlikely that we can get enough "care" to have the Wikisource-specific metadata thing implemented any time soon (sounds like they have a lot of problems, and cate ring to Wikisource is not one of them), so I might try to knock a client-side one up myself. If you can't get an external metadata collection scheme working, it's not a huge crisis, it just means we'll need to keep an eye on diverging metadata. Thanks for the quick reply, I'll keep you informed of any useful progress I make on a book generator.
- One thing that was mentioned in the list was the Dublin Core metadata system. Can you see any milage in the DC system, as compared to (or in addition to) the current WS microformat? Just food for thought. Cheers, Inductiveload 19:05, 22 February 2012 (UTC)
- The current WS microformat is inspired by w:en:Dublin Core that is not adapted to html content. The xhtml ouput of WSexport add dublin core metadata from the microformat. I'm working on a metadata system build with proofreadPage that will provide an api and add DublinCore in header of Wikisource pages. Here is a quick description of the the keys problems of the project. Tpt 21:50, 26 February 2012 (UTC)
Implmenting in Wikibooks
I'm very interested in using this tool on wikibooks. At the moment the pdf exporter there is pretty poor and it would be great to get an epub exporter in place. Do you have any advice as to how this could be done. The book I want to convert is here Pluke (talk) 11:39, 20 January 2013 (UTC)
- Wsexport is currently design to work on Wikisource, so this tool is adapted to Wikisource specificity and some changes in its code is needed to make it work well for Wikibooks. It's not an impossible thing but I'll do it only if there is a strong request of the Wikibooks community. Tpt (talk) 20:01, 20 January 2013 (UTC)
Running from command line
I've just submitted a pull request for some little variation to the way in which paths are coded for the CLI bit of the WSexport tool. I'm working on a little thing that will let me keep my local ePub library up-to-date with Wikisource (updating books as they cange on WS). Does such a thing exist anywhere, does anyone know? Would it be of interest to anyone? And many thanks to tpt for a great tool! :-) Samwilson (talk) 05:12, 19 March 2013 (UTC)