Recent Update

pylori_dl is currently out-of-date (as of Oct 22nd, 2013), please use videolectures-videourl to extract the video URL instead.

PYLORI: PYthon videoLecture Offline RevIew

Pylori comes with a set of command-line tools to enjoy videolectures.net when disconnected:

• pylori_dl - to download video and slides from videolectures.net
• pylori - to play the video and slides downloaded by pylori_dl
• hdslides - to convert PDF slides downloaded by pylori_dl to high resolution images for pylori

Installation

Pylori requires Python, wxPython, MPlayer, ImageMagick, and Ghostscript.

On Mac OS X, you could install the above packages through MacPorts:

$ sudo port install python27 py27-wxpython-devel mplayer-devel ImageMagick ghostscript

On Linux, you could install required packages with package management tools on your system. For example, on Ubuntu:

$ sudo apt-get install python python-wxgtk2.8 mplayer imagemagick ghostscript

Usage

pylori_dl [-v] http://videolectures.net/XXX/[video/NNN/]

pylori_dl downloads video and all the slides of a lecture specified by the given URL, and put them in a directory named after the title of the lecture. If -v or --skip_video is specified, pylori_dl will download only slides and skip the video.

pylori DIRECTORY

pylori plays the video lecture and shows slides syncing with the video on a separate window. DIRECTORY is the directory created by pylori_dl that contains lecture materials.

hdslides [-w WIDTH] [-m] DIRECTORY [PAGES [PAGES ...]]

hdslides will find PDF slides in DIRECTORY, the directory created by pylori_dl, and convert it to high resolution PNG images. The generated images are put into DIRECTORY.

The option PAGES specifies page numbers of PDF to be used for the video in the exact order. PAGES could be a single page number or a range of page numbers in the format a-b without spaces in between. For example, 1-3 and 3-1 are equivalent to 1 2 3 and 3 2 1 respectively. If PAGES is omitted, all pages of PDF are generated.

WIDTH is the width of outputting images, which defaults to 840.

When -m or --match is specified, original low resolution images are automatically matched with PDF pages. PAGES are ignored when this option is used.

hdslides will also make pylori to display new images afterward.

Examples

Supposing you are interested in this lecture, you can run the following to download the materials of that lecture:

$ pylori_dl http://videolectures.net/nipsworkshops2011_ghahramani_nonparametrics/

When pylori_dl is done, a directory named after the title of the lecture will be created, in this case, "Why Bayesian nonparametrics?". You can then watch the lecture by:

$ pylori Why\ Bayesian\ nonparametrics\?/

When a PDF version of the slides is available, pylori_dl will download and place it in the created directory as well. If you are not satisfy with the quality of the original slides snapshots, you can use hdslides to generate a higher resolution copy from PDF slides:

$ hdslides Why\ Bayesian\ nonparametrics\?/

pylori can also be used for lectures with multiple sections such as this one. In this example, if you want to download, say, the fourth part, you can specify the URL like the following format:

$ pylori_dl http://videolectures.net/mlss07_rasmussen_bigp/video/4/

In the multi-part cases, a single section might just involve part of the pages in PDF slides. To generate the slide images for that particular section, you can specify the corresponding page numbers (or ranges) manually. For the same example, it is:

$ hdslides Bayesian\ inference\ and\ Gaussian\ processes-4/ 38-43 42-44 42 41 44 41 44 43

hdslides also provides an auto matching option -m to take care of everything:

$ hdslides -m Bayesian\ inference\ and\ Gaussian\ processes-4/

When -m is used, hdslides will output the matched page numbers so you can use them in the future to save time.