install.rst

Installation

For the moment the simulator has two user interfaces:

• Web. A user runs a simulation server and uses a web interface to set up model parameters and run the simulation.
• Veusz. A user edits script files describing a simulation, run a simulator and the results are rendered using Veusz software (in principle, it quite easy to plug in another means of visualization).

The first approach is easier for new users while the latter one gives more control over simulation.

• Installing Scala
• Steps common for Veusz and Web interfaces
• Running a Web-server with online interface
• Offline usage with Veusz

Installing Scala

The simulator uses a code generated by a translator from Strategy Definition Language. This translator is implemented in Scala progamming language. This language was chosen since it contains functional programming features like pattern-matching and variant types that facilitate the compiler developement but on other hand it is the most "mainstream" language of this kind (which implies existance of a wide developer community and mature tools).

The easiest way to install Scala is to install Simple Build Tool (SBT) from http://www.scala-sbt.org/release/docs/Getting-Started/Setup.html

At Ubuntu SBT may be installed in the following way:

wget http://apt.typesafe.com/repo-deb-build-0002.deb
sudo dpkg -i repo-deb-build-0002.deb
sudo apt-get update
sudo apt-get install sbt

After having installed SBT you may check does it work by typing sbt run in $(MARKETSIM_PATH)/marketsim/gen directory. This will download Scala 2.10 distribution and additional packages. If everything is fine sbt run repeated one more time will print something like:

parsing...done
building name tables...
    removing anonymous packages
    injecting base packages
    removing abstract packages
    applying before typing annotations
done
typing...done
generating python code...done

Note that installing Scala is required only if strategy descriptions are going to be edited.

Steps common for Veusz and Web interfaces

Latest public version of the simulator can be downloaded from https://github.com/fiquant/marketsimulator/releases

This version of the simulator requires Python 2.7 installed: http://www.python.org/download/

Following Python packages need to be installed:

• blist: https://pypi.python.org/pypi/blist/
• docutils: https://pypi.python.org/pypi/docutils (only needed to run Web server)
• numpy: https://pypi.python.org/pypi/numpy (only needed for strategy.MultiArmedBandit)
• pandas: http://pandas.pydata.org/ (only needed for Quotes at the moment)

Normally all these packages can be downloaded and installed using pip (http://www.pip-installer.org/) or easy_install (https://pypi.python.org/pypi/setuptools)

Running a Web-server with online interface

In order to run the web server you need Flask package installed: http://flask.pocoo.org/

wsgi/flaskapp.py launches the Web server at default port 80:

> cd wsgi
> python flaskapp.py
 * Running on http://127.0.0.1:80/

After that you may open localhost page in your Web browser:

Press Run button to launch the simulation:

Offline usage with Veusz

For running scripts and offline visualisation Veusz package is needed: http://home.gna.org/veusz/. After having it installed VEUSZ_EXE environment variable should be assigned to the path to the Veusz executable. Afterwards, sample simulations may be launched:

> cd marketsim/samples
> python try_dependency.py
. . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
75509 events processed in 18.849016855s with rate 4005.99143079 event/s
>