With newspaper, you can do “news extraction, article extraction and content curation in python. Built with multithreading, 10+ languages, NLP, ML, and more!”
I haven’t tried it yet but if you need a corpus with news articles, this project can help.
Human Coders News est un service permettant de partager les meilleures ressources trouvées sur la toile à propos d’un thème précis. Vous pouvez consulter l’ensemble des news sur la page d’accueil, ou bien, cliquer sur un sujet pour filtrer.
In this post I describe how to watch news on a Hungarian site. Although the video that we want to play is in Hungarian, you might get some ideas that you can use in a different project.
Currently I live abroad and sometimes I want to watch news in my mother tongue. So, the Hungarian News Television (HírTV) collects its news programs at http://www.hirtv.hu/view/videoview/hirado . Here, a video has the following URL: http://www.hirtv.net/filmek/hirado21/hiradoYYYYMMDD.wmv , where
YYYYMMDD is the date (for instance http://www.hirtv.net/filmek/hirado21/hirado20101018.wmv). Instead of starting a web browser, visiting this page and clicking on a link, I want to launch the news video with a Python script.
When the script is executed, it may be possible that the news of the current day is not yet uploaded. So we need to verify if the URL exists. However, if we want to get a WMV file that doesn’t exist, the web server of HirTv will return an HTML page instead of indicating that the given URL is missing. So we will have to verify the Content-Type of the URLs. If it’s
text/html => error, if it’s
video/x-ms-wmv => OK.
#!/usr/bin/env python import datetime import urllib import os WMV = 'video/x-ms-wmv' base = 'http://www.hirtv.net/filmek/hirado21/hirado' ext = '.wmv' def get_content_type(url): d = urllib.urlopen(url) return d.info()['Content-Type'] def date_to_str(d): return "%d%02d%02d" % d def prettify(d): return "%d-%02d-%02d" % d def play_video(video_url): print "> " + video_url command = 'mplayer %s 1>/dev/null 2>&1' % video_url #command = 'vlc %s 1>/dev/null 2>&1' % video_url # if you prefer VLC os.system(command) today = datetime.date.today().timetuple()[:3] video_today = base + date_to_str(today) + ext if get_content_type(video_today) == WMV: play_video(video_today) else: yesterday = (datetime.date.today() - datetime.timedelta(days = 1)).timetuple()[:3] video_yesterday = base + date_to_str(yesterday) + ext print "The video for today (%s) is not available." % prettify(today) val = raw_input( "Do you want to watch the video of yesterday (%s) [y/n]? " % prettify(yesterday) ) if val == "y": if get_content_type(video_yesterday) == WMV: play_video(video_yesterday) else: print "Sorry. The video of yesterday (%s) is not available either." % prettify(yesterday)
First we determine the today’s date and using this information we create a URL for the video file. If it really exists (i.e. the Content-Type is correct), then we play it calling mplayer. If the Content-Type is incorrect, then the video of today was not yet uploaded. In this case we offer the user to play the video of yesterday.
Update (20101107): A bug in
prettify() was corrected. Months and days must be padded with
6 must become
06 for instance. VLC support is also added, it’s put in comment.