carabistouille online

Introducing wikipedia2BibTeX

Stéphane Péchard — Wed, 16 Feb 2011 20:10:46 +0000

I know, it’s been a real long time I didn’t blog… again. But hey, it’s not as if you were a lot of readers of mine, right

Anyway, I come back for a quickie, introducing a new tool I fastly hacked: wikipedia2BibTex. Behind this cryptic name is a very basic Python tool which creates a BibTeX citation from a Wikipedia page. Simple, isn’t it? For example, if you want to cite the wonderful BibTeX page, just type:

wikipedia2BibTeX BibTeX --lang en

in your favorite terminal and it will show the result as:

@misc{wiki-bibtex, author = "Wikipedia", title = "BibTeX --- {W}ikipedia{,} The Free Encyclopedia", year = "2011", url = "http://en.wikipedia.org/w/index.php?title=BibTeX&oldid=406037891", note = "[Online; accessed 27-January-2011]" }

As default, the --lang parameter is put to French (--lang fr). Please enjoy and share!

Download

As usual, the code is on **gitorious**, get it with:
git clone git@gitorious.org:wikipedia2bibtex/wikipedia2bibtex-gitorious-wiki.git

Install

If you want a system-wide install, type: sudo python setup.py install.

If you to keep it just for you or if you don’t have enough rights, type: python setup.py install --user.

Future

Please, don’t hesitate to share your thoughts, remarks, critics and/or patches on wikipedia2BibTeX, it’s still a young baby who needs love ;–)

I’ve identified some problem to be fixed in the future:

unicode handling (damn you Python 2.x!)

I hope I’ll find some time to fix them soon!

OsmUserStats 0.3.2 is out

Stéphane Péchard — Thu, 12 Aug 2010 21:14:47 +0000

As I said in my previous post, I was working on this small Python script which allows you to plot some of your activities statistics as a OpenStreetMap user. The least I can say is that the code was very bad written, therefore very difficult to maintain. That’s the reason – the only reason – of this release, as no new feature are available.
Now adding some new ones would be easier, thanks to this work of rewriting and refactoring of the code. As a helper to get a cleaner code, I used Pylint. That’s a very nice (and nitpicking) tool that tells you are crappy your code is. In fact, it is quite difficult to get the maximum score (10), especially with HTML code, as OsmUserStats deals with. I used no pylintrc for this evaluation, I probably will use one in the future, in which I’ll disable the convention messages caused by HTML code. This 0.3.2 version gets a fairly good 9.81/10, with 5 warnings messages, 3 convention messages and 1 refactor messages. Warnings are caused by unused variables that have to be created by matplotlib, the plotting library. The refactor message is caused by a function that has 13 branches, one more than the maximum allowed by Pylint (nitpicking as I said).
As usual, the code is available at my gitorious account: http://gitorious.org/osmuserstats/osmuserstats
And the webpage of the project is: http://spechard.dgplug.org/OsmUserStats/

Enjoy!

Not completely dead…

Stéphane Péchard — Mon, 02 Aug 2010 23:36:05 +0000

Ok, it looks like this blog is dead, but here is the sound of a revival! Some news first: after several months of professional inactivity (which conducted me doing Mediaspy), I found a job last February in a small company in Nantes, called Ripple Motion. It is specialized into mobile application, and very specialized into iPhone apps. I work on a R&D project dedicated to Augmented Reality. My everyday work is a mix of Objective C/OpenCV/OpenGL code samples

The sad news is that I don’t think I will continue Mediaspy from now on. The good news is that the code is free so anyone can take it to play along! In fact, I heard there maybe are some people interested in this project! That would be great, and gladly encourage these retakers to do so. More to come on this topic, I hope.

My (not very) new pet project is called OsmUserStats. You can learn more about it at this address : http://spechard.dgplug.org/OsmUserStats/. It is a stats generator for user-based data of the OSM project. I have been an OSM contributor for about a year now, and found the project very stimulating and interesting. I mapped about 90% of my hometown and have a lot more to do. As a scientist, I like to see my work going on, as well as making summary, especially is such a big and complex crowdsourcing project. That’s why I decided to make OsmUserStats. The code is surely not very proper, but that’s the goal of the next release, expecting for the weeks to come. Again, I’ll tell more about this in the future.

Better requests with MediaSpy 0.7.4

Stéphane Péchard — Fri, 25 Sep 2009 09:37:44 +0000

MediaSpy 0.7.4 is out! With this release, you can now edit the title of every media, in order to make it as close as possible to the actual title. Of course, you may want to add data to know what kind of media it is (like dvdrip, or the number of a disk file), so the title is divided into two parts. The first one is the title as it is requested in online databases. The second part is whatever you want and not inserted in the request.
I had to choose a character to separate both parts, it will be ‘_’ (underscore), as I think it is not very common in media file names (at least less common than . ou -). Here is a screenshot where you can see a title being edited and the tooltip explaining how to do it.

The main window with editable title and tooltip

As you may have noticed, correctly found medias are no longer colored in green, they are just transparent as usual table view now. Enjoy!

Download

https://fedorahosted.org/releases/m/e/mediaspy/mediaspy-0.7.4.tar.bz2

MediaSpy 0.7 ready for service!

Stéphane Péchard — Tue, 01 Sep 2009 13:32:46 +0000

I just released a new major version of MediaSpy. A lot of new stuff is here:

operationnal filters for title, tags and seen/unseen movies
easy tags edit
option and media edition dialogs
medias are playable directly from MediaSpy and player is at your choice!
MediaSpy is available in French, English and (almost completely) Tamil
and some bugfixes of course

As usual, you can the last version on the fedora repo : https://fedorahosted.org/releases/m/e/mediaspy/

First release!

Stéphane Péchard — Fri, 14 Aug 2009 09:06:12 +0000

Thanks to Kushal Das, the first release of MediaSpy is done! He even said me he was planning to package it for Fedora!! Great news!
So, from now on, I will release every version as a tarball. You can find them here.

DISCLAIMER: this “release” is far from being usable on a daily basis! Use it at your own risk. Please, feel free to warn me of any bug you found. If you already used an old version of MediaSpy, you may need to delete the .MediaSpy directory first.

Use case 1 OK!

Stéphane Péchard — Sat, 01 Aug 2009 15:50:09 +0000

For those of you who didn’t notice, the yesterday’s release closed the first use case I presented at the beginning of the project. The only differences are about automation. Step 6 and 10 don’t need the user to interact with MediaSpy, it is done automatically. Enjoy!

MediaSpy 0.6 is out!

Stéphane Péchard — Fri, 31 Jul 2009 16:26:50 +0000

Yes, I know, it’s been a long time since version 0.5, but this one is made of several nice small steps. Together, these steps allow me to release an « almost usable » version of MediaSpy. Now, when you add directories to the collection, imdb website is requested for every media of the list. Information from the site are displayed on the right side. The nice plus is the ability to see which media has been found on imdb via a very simple color code. When the media name is on a green background, information has been found. It is red otherwise. When no color appears, it means that the media was already in the database when you launched MediaSpy, therefore research has already been done. Here is a screenshot of these features:

A major issue is the sensitivity of Mediaspy to the media names. In fact, if your medias aren’t well named, imdb queries will be hazardous and result in erroneous movie information…. The good news is I got some ideas to change this

Far from perfect and still a lot of things to do, but still going on!

[Tip] Handling an HTTP redirection with QNetworkAccessManager

Stéphane Péchard — Thu, 23 Jul 2009 14:55:35 +0000

While writing a (small) IMDb API for MediaSpy, I encountered the following problem. When you make an IMDb research for a movie title without ambiguity, IMDb makes an automatic redirection from the research page to the full page of the movie. The question is: how to handle this in Qt? Happily, Nokia provides some nice snippets like this one. Some adaptation later, here is my code:

QNetworkReply* networkReply;
QVariant possibleRedirectUrl =
    networkReply->attribute(QNetworkRequest::RedirectionTargetAttribute);

// We'll deduct if the redirection is valid in the redirectUrl function
QUrl urlRedirectedTo = redirectUrl(possibleRedirectUrl.toUrl(), urlRedirectedTo);

// If the URL is not empty, we're being redirected.
if(!urlRedirectedTo.isEmpty())
    this->networkManager_->get(QNetworkRequest(urlRedirectedTo));
else
    urlRedirectedTo.clear();

The function redirectUrl being:

QUrl InfoManager::redirectUrl(const QUrl& possibleRedirectUrl, const QUrl& oldRedirectUrl) const {
    QUrl redirectUrl;
    if(!possibleRedirectUrl.isEmpty() && possibleRedirectUrl != oldRedirectUrl)
        redirectUrl = possibleRedirectUrl;
    return redirectUrl;
}

Easy, don’t you think?

MediaSpy in your language!

Stéphane Péchard — Sun, 19 Jul 2009 15:32:00 +0000

Last week, I was on holiday in the Gulf of Morbihan. Had just some time to work on internationalization. Now, MediaSpy can speak your language!! For this, just ask me to provide you a .ts file for your language. Translate it with the use of linguist, and send me the .ts file back. I’ll integrate it in the git repo and we’ll all be happy!!