903 How to extract a ZIP archive on Pocketbook 903

[rkomar]

Originally Posted by review

Well, there is one thing what you can do: you can suggest ideas for simple apps. For example at the moment I'm working on an improved rss feed reader. Brainstorming ideas can be sometimes very helpful for innovations. As long as the official SDK is not published the complexity of the programs is certainly limited (like: I won't be able to write an app for pdf annotation etc ...) but still feel free just to suggest what ever comes to you mind.

I think a port of the original text version of sokoban would be nice (using the buttons instead of key strokes, of course).

Oh yes, and an app for viewing cbz/cbr files directly (rather than unpacking them and viewing the images separately).

[paola]

Originally Posted by review

Well, there is one thing what you can do: you can suggest ideas for simple apps. For example at the moment I'm working on an improved rss feed reader. ...

on this, I am very good
indeed, getting news was going to be my next homework (after pdf viewing), so will get back to you on this one.

Just one thing, though: it was mentioned in post #6 in this thread that the stylus does not do pen strokes - I am surely missing something, and I guess I should ask LoneTech what s/he means, but just wanted to jhighlight that when reading you can turn pages with a swipe of the stylus, and of course you can write free-hand in scribble: don't these things amount to the software being able to recognise penstrokes?

[review]

OK, I will try my best to explain, please feel free if still something doesn't seem to make sense to you.

Lonetech is working on getting an SDK up an running as the official one probably won't be released in the next weeks (the earliest being March 2011). An SDK is a so called SoftwareDevelopmentKit. So basically a way for the writer of a program to give some output on the screen etc. Those programs will be translated (technical term: compiled) from a human readable language (so called programming language like C or C++) to the language the processor of the device speaks. Now, I will probably wait until the official SDK is published before I do some serious programming. However, what I will do in the meantime is to use the programs which are already on the device and put them together in a meaningful way (this is called script). You don't need a SDK for writing a script as it just uses the programs which are already available on the device and makes them useful to the end user. The best illustration is my unzip.app. There is a build-in functionality of the 903 to unzip zip files. I used simply that functionality and just defined where the program should look for the input file and where it should put the content of the zip file. With scripts I'm fairly limited. E.g. I won't be able to have any direct output on the screen. The news reader would be designed in the way that you set the parameters for the RSS feed on your computer, you upload the app to the device and once you run it it will gather the news, process them and create an output file (as in a script we can't directly talk to the screen) like a pdf or an epub etc.

Coming to penstrokes. This is directly related to the SDK. I will try to translate what Lonetech said in less technical terms (although I do appreciate that he did use technical terms): There are different ways to talk to the screen and to interface with the touchscreen. Most programmers usually don't talk directly to the screen but use the functions which are provided by the company (with their SDK). Now, LoneTech discovered that there are two ways how the 903 talks to the screen. One is called gtk-directfb which is the easiest way to get the annotations software called Xournal to work on the 903. However, this particular way to talk to the screen might not allow penstrokes. You can see this in all programs which use this method. As far as I'm aware of the browser is the only program which does this. The other programs use a different method to get the touchscreen information and obviously, like you have pointed out, those programs support penstrokes.

Hope this clarifies it a bit.

[paola]

Originally Posted by review

OK, I will try my best to explain, please feel free if still something doesn't seem to make sense to you.
...
Hope this clarifies it a bit.

crystal clear, and most helpful!

[paola]

Originally Posted by review

For example at the moment I'm working on an improved rss feed reader.

well, on this one after playing around with the pocketbook news tonight I have to say I am not impressed. The main problems I see currently are:

1) it looks like it is not possible to change the name of the various feeds;

2) only a very limited subset of the feeds is downloaded - in spite of having put three of them for a try, only one newspaper feed (the first one) was downloaded. In fact, it was very poorly done, as there were only the titles, and no full corresponding news article. In addition, there was no table of contents.

3) I could not figure out how to input details for downloading from sites that require a subscription (e.g. i subscrive to the printed copy, and I am entitled to the vitual copy too - but I have to provide a password for that).

Of course, it may be me - but I am pretty sure that e.g. for the site which created my problem 2) above the feed address is correct, as it works perfectly in Calibre (where I get all the newspaper sections, all of the articles - not just the titles as in pocketbook - and a nice table of contents that I can easily navigate.

Just a thought, in case you are still working on it

[review]

Yeah, the RSS feed reader which comes with the reader is indeed really bad and not useful at all. It is literally just displaying the information which is provided in the feed itself.

Originally Posted by paolamanzini

Just a thought, in case you are still working on it

Sure, I'm still working on it. Just yesterday I managed to complete the "recipe" for a german newspage (Der Spiegel). It will download all articles provided in the RSS feed, removes the ads etc. This is a very similar approach like what calibre does. You basically have to define for each newspage a recipe so that the program knows how to get the articles etc.

Calibre does a great job and has a large user base which makes it possible that many skilled programmers can contribute recipes. Further you have much more processing power on your desktop PC compared to the 903.

Nonetheless, I'm still working on it and hope to get a stable version soon.

[paola]

Originally Posted by review

Nonetheless, I'm still working on it and hope to get a stable version soon.

of course did not meant to put any pressure: only very grateful to all the competent guys like you in this forum that help us hapless folks!

[review]

@paolamanzini: Don't worry, I didn't take it the wrong way.

Since I need to make some little adjustments for each RSS feed (this is what I called earlier recipe) please feel free to just post the newspages you would like to read on your 903. As time permits I will try to integrate them. Will keep you updated on the progress then.

[paola]

Originally Posted by review

@paolamanzini: Don't worry, I didn't take it the wrong way.

Since I need to make some little adjustments for each RSS feed (this is what I called earlier recipe) please feel free to just post the newspages you would like to read on your 903. As time permits I will try to integrate them. Will keep you updated on the progress then.

that is very kind - if I only had to pick one, it would be this one:
http://xml.corriereobjects.it/rss/homepage.xml
I am not too sure whether this is the Calibre recipe, but from what I can read in Calibre I get:

#!/usr/bin/env python
__license__ = 'GPL v3'
__author__ = 'Lorenzo Vigentini, based on Darko Miletic'
__copyright__ = '2009, Darko Miletic <darko.miletic at gmail.com>, Lorenzo Vigentini <l.vigentini at gmail.com>'
__version__ = 'v1.01'
__date__ = '10, January 2010'
__description__ = 'Italian daily newspaper'

'''
http://www.corriere.it/
'''
import time
from calibre.web.feeds.news import BasicNewsRecipe

class ilCorriere(BasicNewsRecipe):
__author__ = 'Lorenzo Vigentini, based on Darko Miletic, Gabriele Marini'
description = 'Italian daily newspaper'

# cover_url = 'http://images.corriereobjects.it/images/static/common/logo_home.gif?v=200709121520


title = u'Il Corriere della sera'
publisher = 'RCS Digital'
category = 'News, politics, culture, economy, general interest'

encoding = 'cp1252'
language = 'it'
timefmt = '[%a, %d %b, %Y]'

oldest_article = 10
max_articles_per_feed = 100
use_embedded_content = False
recursion = 10

remove_javascript = True
no_stylesheets = True

html2lrf_options = [
'--comment', description
, '--category', category
, '--publisher', publisher
, '--ignore-tables'
]

html2epub_options = 'publisher="' + publisher + '"\ncomments="' + description + '"\ntags="' + category + '"\nlinearize_tables=True'

keep_only_tags = [dict(name='div', attrs={'class':['news-dettaglio article','article']})]

remove_tags = [
dict(name=['base','object','link','embed']),
dict(name='div', attrs={'class':'news-goback'}),
dict(name='ul', attrs={'class':'toolbar'})
]

remove_tags_after = dict(name='p', attrs={'class':'footnotes'})

def get_cover_url(self):
cover = None
st = time.localtime()
year = str(st.tm_year)
month = "%.2d" % st.tm_mon
day = "%.2d" % st.tm_mday
#http://images.corriere.it/primapagin...ina_grande.png
cover='http://images.corriere.it/primapagina/storico/'+ year + '_' + month +'_' + day +'/images/prima_pagina_grande.png'
br = BasicNewsRecipe.get_browser()
try:
br.open(cover)
except:
self.log("\nCover unavailable")
cover ='http://images.corriereobjects.it/images/static/common/logo_home.gif?v=200709121520'
return cover

feeds = [
(u'Ultimora' , u'http://www.corriere.it/rss/ultimora.xml' ),
(u'Editoriali' , u'http://www.corriere.it/rss/editoriali.xml'),
(u'Cronache' , u'http://www.corriere.it/rss/cronache.xml' ),
(u'Politica' , u'http://www.corriere.it/rss/politica.xml' ),
(u'Esteri' , u'http://www.corriere.it/rss/esteri.xml' ),
(u'Economia' , u'http://www.corriere.it/rss/economia.xml' ),
(u'Cultura' , u'http://www.corriere.it/rss/cultura.xml' ),
(u'Scienze' , u'http://www.corriere.it/rss/scienze.xml' ),
(u'Salute' , u'http://www.corriere.it/rss/salute.xml' ),
(u'Spettacolo' , u'http://www.corriere.it/rss/spettacoli.xml'),
(u'Cinema e TV', u'http://www.corriere.it/rss/cinema.xml' ),
(u'Sport' , u'http://www.corriere.it/rss/sport.xml' ),
(u'Roma' , u'http://www.corriere.it/rss/homepage_roma.xml'),
(u'Milano' , u'http://www.corriere.it/rss/homepage_milano.xml')
]

But there is no rush, I imagine I could put this into your app once it sees the light

And in the meantime, good thing that Calibre exists!

[review]

well, the calibre recipes are not compatible to my app. You see, there is only a handful of programs on the 903.
The good news is that the page you've provided looks fairly easy to integrate. If you're lucky I can send you the first alpha test within the next hour or so. With alpha I mean, there might be a lot of improvements necessary before it becomes stable.
Talk to you later.

[review]

Here is the first version of my RSS app (adapted to your italien newspage).

1) download and unzip the attached file,
2) put it somewhere on your device (doesn't matter where). e.g. games
3) you need to connect now to the internet. So what I usually did was to just start the webbrowser and close it immediately once it has loaded.
4) then run my RSS feed gathering app.
5) This may take a while and you will just see nothing apart from the waiting clock symbol. On my network it took about one minute.
6) once it is complete (the screen will refresh) you can disconnect from your wireless connection
7) then go to the topmost folder of the device. There will be a new file called corriere.html


Known Bugs / Currently working on
1) The most annoying bug at the moment is that sometimes (maybe always) the text is treated as indiviual characters. This is really bad as the word wrapping will fail and fbreader will just wrap a word somewhere in the middle of it.
2) connected to the first bug, the TTS will just spell all letters individually instead of reading the text.
3) no images yet.


So you see, there is a lot of work left to do. The topmost priority now is the first bug.

[review]

Originally Posted by paolamanzini

that is very kind - if I only had to pick one, it would be this one...

well, you can post many more and others are very welcome to do the same. However, the next task for me will be to remove the first bug mentioned above, so that the app is actually useful. Once that is done it should be easily possible to adapt the app to other sources. Further, the current version has plenty of possibilities for improvements:
1) other output formats (like epub)
2) if you ran the app a second time it would take exactly the same time although most of the news might be already downloaded. So I might be able to speed up the update process to be far less than a minute.
3) probably plenty more.....

[paola]

gosh, isn't this beautiful!

Review, superthanks, you talk about bugs but this is way better than the Pocketbook app!
Presumably for a different feed I would simply have to change the webaddress in your file?

Originally Posted by review

Known Bugs / Currently working on
1) The most annoying bug at the moment is that sometimes (maybe always) the text is treated as indiviual characters. This is really bad as the word wrapping will fail and fbreader will just wrap a word somewhere in the middle of it.
2) connected to the first bug, the TTS will just spell all letters individually instead of reading the text.
3) no images yet.


So you see, there is a lot of work left to do. The topmost priority now is the first bug.

Well, this is good enough for me - but if you need feedback, I noticed that the table of content looks a bit weird - it is probably taking various headings and subheadings and treating them as different news - it does not really matter, though, as they all correctly refer to the same page. By all means I am not complaining, just thought you might want to know.

Many many thanks!!!

[review]

Thanks for the feedback. Yes, I noticed that the table of contents is indeed too full. Working on it. Thanks for mentioning. Feel free to post any further improvements or bugs you experience.

Unfortunately changing just the feed url is not enough. In order to filter out adverts etc. I don't follow the link provided in the feed but modify it to a link which would be used for print outs because most stylesheets don't put the adverts on the printout. But since each page is different it will always need a little finetuning to adapt the script to a given site.

But as I said, feel free to post more feeds.

[paola]

[QUOTE=review;1324935]
Unfortunately changing just the feed url is not enough...[/QUOTE
should have thought it was too easy
yep, as soon as I overcome my sense of guilt for free-riding, I will post more feeds!