ralsina / aranduka Goto Github PK

We could create a basic core module to do some UI stuff that is repeated in 
many places. I'm currently thinking of displaying modals with messages for the 
user or asking for confirmation (QMessageBox), but it could be extended to 
other stuff.

For example, some plugins that don't have a UI (like the OPDS publisher) may 
need to display some messages to the user in case of error, or to show some 
information (like the URL where the catalog is published). With this module we 
could have all these type of messages centralized and let the plugin forget 
about importing Qt's modules. 

Needless to say, it would also help us repeat less code.

What do you think? What would you call it?

There is an API for bibliographic data:

http://www.hathitrust.org/bib_api

And one for actual pages of books:

http://www.hathitrust.org/data_api

It's non-standard and I have no idea how hard to use it could be.

The store plugins are all very similar, but they are not implemented in the 
same way.

Also, the look of the output is ugly.

So, I will take the feedbooks plugin polish the heck out of it and use it as a 
prototype. Then, I will take the others and make them work and look the same.

When you open the plugin configuration dialog, the items (plugins) don't show 
properly (see screenshot). If you enlarge the dialog, it fixes it.

Well,that. It throws exception,the UI is lacking, etc etc.

If you are viewing the list of titles, and double-click on a book to edit, then 
click on "titles" again, you are placed at the top of the list.

You should stay at the same position you were in.

The main window is a little small for large displays. Moreover, the annoying 
1-pixel tolerance for the resize cursor in Ubuntu 
(http://brainstorm.ubuntu.com/item/8473/) makes it hard to enlarge it. 

Suggestions:
- Make the main window larger
- Have the main window resize proportionally to the size of the screen.
- Have Aranduka remember the size of the main window from the last time it was 
closed.

Any (or all) of the above will do.

Currently the Manybooks.net book store plugin only allows you to download books 
in ePub format (which is the one published on the OPDS catalog). We could hack 
the plugin using the list of formats on the converter plugin, so that you could 
download the book directly in the format you prefer.

Write a plugin that scans barcodes and creates the book in the database.

I have done it before, should be semi-trivial.

Currently the tools menu is populated on startup.
That means you don't get the plugins you enable until you restart the app.

Also, disabled plugins should probably be displayed as disabled.

Pasos para reproducir la falla:

1. Ingresar el ISBN 9871187599 ("El Idiota", de Dostoievski)

Se esperaría ver la imagen de la tapa. Actualmente, está leyendola desde esta 
web:

'http://covers.openlibrary.org'

Pero, hay casos (como este) en el que la tapa no aparece en covers.openlibrary, 
pero si en google.books (http://bks3.books.google.com/books?id=zqKGPwAACAAJ)

Right now, when you obtain a book, if the bookstore plugin knows how to get the 
cover, it will do it by passing an explicit URL.

That's good.

OTOH, if the URL is not provided, or fails, it will try to guess a cover and 
get it. That should move from the Book model into the guesser plugins.

Design an "About this book" page, which will be displayed when a user 
double-clicks on a book.

It should be an HTML page which shows the book's information (title, author/s, 
cover, tags, identifiers and associated files) and the possibility to access 
the book editor.

The page should also display dynamic gadgets (chunks of HTML that allow you to 
perform different operations related to the book in question) that would be 
associated to plugins.

Guess with author "Cory Doctorow" and you get the same 4 books repeated 3 times.

That was done in an old branch: 

http://code.google.com/p/aranduka/source/browse/?r=smashwords#hg%2Fsmashwords

I implemented the functionality of the "delete" option in the context menu in 
r20ec584eb1. Could you check it out? I'm not sure if the file handling should 
be in the main module, but... 

When guessing a book with Google Guesser I get this error:

EP: set([u'Alibris Guesser', u'Books By Author', u'Folder Device', u'Books By 
Tag', u'Barcode Scanner', u'FeedBooks Converter', u'OPDS Publisher', u'Books By 
Title', u'manybooks_net', u'Google Books Guesser', u'FeedBooks', 
u'Manybooks.net Converter', u'archive_org', u'RSS'])
CIC
Traceback (most recent call last):
  File "/home/elbarto/aranduka/src/book_editor.py", line 73, in on_guessers_currentIndexChanged
    self.guesser = self.guesser_dict[unicode(text)]
KeyError: u'Google Books'
EP: set([u'Alibris Guesser', u'Books By Author', u'Folder Device', u'Books By 
Tag', u'Barcode Scanner', u'FeedBooks Converter', u'OPDS Publisher', u'Books By 
Title', u'manybooks_net', u'Google Books Guesser', u'FeedBooks', 
u'Manybooks.net Converter', u'archive_org', u'RSS'])
<GoogleGuesser object at 0x94035ec> <class 'GoogleGuesser'>

I'm suppose the problem is with the name of the plugin, but since I don't know 
really what happens there, I don't want to mess things up by trying to fix it 
:-)

On the book editor all the available Guessers are being listed even if they're 
not enabled on the plugins configuration.

We should add the posibility of stopping the OPDS catalog server without the 
need closing Aranduka.
One possibility would be to change the menu's action from "Publish catalog" to 
"Stop catalog server" or something like that.
But that raises the issue of how to implement that within the plugin structure.

Roberto Alsina mentioned he had done something like this and it would be nice 
to have it integrated to Aranduka, so that you can convert your feeds into one 
or more .epub files and read it on your favourite ebook-reader.

It doesn't do anything now. It should ask for confirmation and delete the file.

What's the point of the google_book folder at the root of the integrate branch? 
It should probably be deleted.

ODPS publisher plugin listents on localhost:8080 (this is hardcoded). If the 
port is being used, it fails miserably without displaying any error to the user.

1 - The user should be able to configure address and port on the plugin's 
configuration.

2 - The plugin should show a proper error to the user if it can't bind to the 
configured address/port.

Add support for comics file formats.
More information about this formats at Wikipedia:
http://en.wikipedia.org/wiki/Comic_Book_Archive_file

There's an inconsistency with the usage of the name "tags" or "shelves" for 
book tags.

On the database models we're calling them tags. The same in the menu at the 
left of the list. However, when you edit a book you define shelves.

We should pick one and stick to it.

The format looks simple enough.

The menu shows the book's list of files.
It should only show the files if the actual file is there. 

That way it won't display empty files and files that are currently downloading.

What steps will reproduce the problem?
1. Add an invalid epub file to your Aranduka
2. Attempt to open it

What is the expected output? What do you see instead?
The program should display an error message saying that the file has an 
incorrect format. Instead it doesn't do anything.

Please use labels and text to provide additional information.

Traceback information:

Traceback (most recent call last):
  File "main.py", line 177, in <lambda>
    lambda f = f: self.openEpub(f.file_name))
  File "main.py", line 155, in openEpub
    viewer = EpubViewer(fname)
  File "/home/elbarto/aranduka/src/epubviewer.py", line 11, in __init__
    self.epub = EpubDocument(fname)
  File "/home/elbarto/aranduka/src/epubparser.py", line 17, in __init__
    self.book = zipfile.ZipFile(fname, "r")
  File "/usr/lib/python2.6/zipfile.py", line 696, in __init__
    self._GetContents()
  File "/usr/lib/python2.6/zipfile.py", line 716, in _GetContents
    self._RealGetContents()
  File "/usr/lib/python2.6/zipfile.py", line 728, in _RealGetContents
    raise BadZipfile, "File is not a zip file"
zipfile.BadZipfile: File is not a zip file

I suggest we add support for managing quotes. It could be a good step in the 
way of making Aranduka more "sociable".

Aranduka could help the user keep record of his favourite quotes, save them, 
tag them, search them, share them on social networks (Twitter, Facebook, 
GoodReads -this last one also lets you save quotes), etc. As I suggested on 
issue 16, we could even have our file viewers let you select and save a quote 
in two clicks.

I'm thinking we could add the following entity:

Quote
=====
id
text
book_id (optional)
author_id (optional)
page (optional)
tags (many-to-many)

What do you think? In favour? Against? Suggesstions/corrections/questions?

What steps will reproduce the problem?
1. Edit a book's info
2. Change the author
3. Save
4. Go to the authors view

What is the expected output? What do you see instead?
Authors without any books shouldn't be listed.


What version of the product are you using? On what operating system?
Integrate branch on Ubuntu 10.10

So, when you add another book and it gets the same ID, it inherits the old 
(wrong) cover.

Well, that. In some cases it makes the guesser dialog fail. 

As it's been discussed on the mailing list, the current importer is mostly a 
proof of concept. We should define further how the importer should operate and 
maybe design a GUI. 

This elements have been suggested for the importer interface:

- Option to skip the search (guessing)
- Adding a progress bar
- Adding a special tag for imported books

manybooks.net provides the ebooks in lots of formats. 

Implement a converter like the feedbooks one, that works by re-downloading in 
the desired format.

Currently there are two viewers for ebooks: cbzviewer and epubviewer

There is a lot of code repeated between them. 

Now that we have two examples, it's a good moment to refactor the common code 
into a base class and create a decent architecture for them.

It cannot load the XML library

There are a bunch of places where aranduka blocks while a download is in 
progress (for example, books). That's not good.

It would be good to have some non-blocking download manager thing that can give 
feedback and do things in the background.

When a book has more than one file associated to it, the context menu should 
show a submenu with the files to delete. In this case the delete should be of 
the file only and not the whole book.

I got this error while trying to open the epub file I'm attaching.

Opening: 
/home/elbarto/.aranduka/ebooks/17-the-hawaiian-islands-the-department-of-foreign
-affairs-.epub
BASEPATH: 
Traceback (most recent call last):
  File "main.py", line 249, in <lambda>
    lambda f = f: self.openEpub(f.file_name))
  File "main.py", line 222, in openEpub
    viewer = EpubViewer(fname)
  File "/home/elbarto/issue47/src/epubviewer.py", line 14, in __init__
    self.epub = EpubDocument(fname)
  File "/home/elbarto/issue47/src/epubparser.py", line 45, in __init__
    self.toc_id = self.spine.attrib['toc']
KeyError: 'toc'

The current format (a list of text) is ugly, confusing, and doesn't show the 
covers.

When you start Aranduka without a ~/.aranduka folder, all plugins start as 
disabled.

So, when you click on the tools menu, it's empty.

However, when you go to the plugins config dialog, all plugins are set to 
enabled without user intervention, and now tools is full.

That's very confusing.

Provide some way to download plugins. It has the added bonus of a chance to 
ship updated plugins to users.

Since plugins are one-file-and-one-folder, we could just download zipped files, 
unzip-em, and stick them in ~/.aranduka/plugins.

Add to that a RSS feed for updates (we already have feedparser).

Sometimes when you change plugins settings (currently, when you enable or 
disable a plugin), it's necessary to refresh the current window.

Examples:

- If you enable or disable a shelf view, you need to refresh the list of views 
to show or hide that plugin.
- If you are editing a book and you enable/disable a Guesser, you need to 
reload the editor to display the new guesser on the list.

We should figure out a way to hook plugins (and maybe plugin categories) to UI 
elements, so that they are refreshed when a plugin/plugin category changes. 

I guess some witchcraft with slots/signals could help here.

This should probably be fixed once issue 6 is completed, but I'll post it so 
that we can check it again afterwards.

The Manybooks.net plugin lets you download books in ePub format. However it's 
not working properly. To the problems described in issue 6, we should add that 
the download doesn't work and instead creates an empty file.

Maybe the plugin could do a simple format-check after downloading (this could 
be part of the epubparser module).

It is supposed to open the editor,of course.

Currently it displays checkboxes so you can either use the current title as a 
guess hint or not.

However, if you want to try different variations on titles or authors, you need 
to cancel, change in the editor, save, and guess again.

That's not a great workflow.

If there is a cover image that Qt can't read or is corrupted, the item looks 
bad.

It should be fixed so it shows the "no cover" image.

A book can be associated to many files. If you delete a book, all its files are 
deleted. We need to figure out a UI to delete single files from a book, maybe 
in the book editor.

Do as the summary says, and bad things happen.

I think this is related to the Issue 22 changes?

ePub files (and probably other formats), have useful metadata in them.

For example, ePub usually has cover, author and title.
In addition, it's fast and cheap. So, it would be useful to implement a guesser 
that uses that data, specially since we already have an ePub parser in the 
project ;-)