Google Scholar™'s Terms of Service explicitly forbids to use robots to automatically retrieve data from its servers. We do not condone, promote or tolerate the use of our software for illegal purposes and cannot be held responsible for what you might do with it. We publish it for research purpose only and peer-review. Don't be evil.
This program allows one to automatically scrape publications from Google Scholar, starting from words to search ("osteoporosis", "we have never been modern") and following the "Cited by" links. You can then export the graph of the publications.
How to install scholarScape
The first two steps (starred ones) are designed for a Debian based distribution as they involve installing packages (MongoDB and Scrapy) with apt-get. However MongoDB has packages in other distributions and you can install Scrapy from the sources. See in the "Read More" section the links to the install pages of these softwares.
The last steps should not be specific to any version of Linux.
Tell me if you have or have resolved any problem and I will add your solution here.
*Add some repositories first
Execute this command to add Scrapy and MongoDB repositories to your /etc/apt/sources.list, be careful to change DISTRO to the name of your Ubuntu distribution
sudo echo -e "deb http://archive.scrapy.org/ubuntu **DISTRO** main\ndeb http://downloads-distro.mongodb.org/repo/ubuntu-upstart dist 10gen" >> /etc/apt/sources.list
You can get the name of your distribution by executing :
and the line beginning by
DISTRIB_CODENAME tells you the name of your version (ex :
You'll want to update apt after having added repositories
sudo apt-get update
*Install them all
You can now install all the dependencies scholarScape relies upon. It is recommended to install
virtualenv to set up a virtual environment in order not to disturb other programs.
sudo apt-get install python-dev bzr mongodb-10gen scrapyd-0.15 git-core sudo easy_install virtualenv sudo easy_install scrapy sudo easy_install pymongo sudo easy_install pystache sudo easy_install python-levenshtein bzr branch lp:txjsonrpc cd txjsonrpc sudo python setup.py install cd ../ rm -rf txjsonrpc/ curl http://pypi.python.org/packages/source/n/networkx/networkx-1.6.tar.gz#md5=a5e62b841b30118574b57d0eaf1917ca | tar zx cd networkx-1.6 sudo python setup.py install
Fork some code
The latest version of scholarScape is always available at github. To clone the repository
git clone https://github.com/medialab/scholarScape.git
You can put scholarScape anywhere you want but if you want to follow the Linux filesystem hierarchy (explained here <http://serverfault.com/questions/96416/should-i-install-linux-applications-in-var-or-opt>, you might want to put it in /usr/local/scholarScape/.
Connect to MongoDB and add the user configured in
config.json, in the MongoDB part (user :
scholarScape, password :
diabal by default) .
./mongo > use scholarScape > db.addUser("scholarScape", "diabal")
Setup a few directories
Create the directories where scholarScape will export the graphs. The adress can be relative
(to the directory where scholarScape.tac is) or absolute. By default, the export directory
data/ and graph files are exported under
gexf/ so you'll do :
mkdir data data/gexf data/zip data/json
Launch the server
Now you can launch the scholarScape server. Go to the source directory and execute this command
cd scholarScape twistd -ny scholarScape.py
You can omit the parameter "n" if you want to start the server as a daemon.
- Usually other ports than 80 are not available from outside. If you want your server to
- be available from outside, you can setup a reverse proxy in Apache.
People then will be allowed to access scholarScape on localhost/scholarScape instead of accessing on localhost
The file is named scholarScape-apache.conf.
You can use it by doing
sudo cp scholarScape-apache.conf /etc/apache2/sites-available/scholarScape sudo a2ensite scholarScape
You may want to change the port used in this file if you changed the default port in scholarScape's config.
After the installation you'll want to type in
localhost:TWISTED_PORT in your
favorite browser and you will find the scholarScape's WebUI (
TWISTED_PORT is configured in your
You can then follow the tutorial from there.