PyVLE

PyVLE description

PyVLE is a Python package for using VLE framework into Python. It provides features such as loading VPZ files, modifying conditions values of models, simulating VLE models and recovering simulation results. It provides other features regarding the experiment such as modifying duration or seed, obtaining information on models and views.

PyVLE is one component of a generic solution for developing web applications for VLE framework.

Installation

git clone https://github.com/vle-forge/pyvle.git

Usage

Once VLE and PyVLE are installed, you can launch python and import pyvle.

$ python
Python 2.7.2+ (default, Oct  4 2011, 20:06:09)
[GCC 4.6.1] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import pyvle

or

>>> from pyvle import Vle

The package is mostly based on a Python class Vle. In order to know what functions are available into this package.

>>> dir(pyvle.Vle)

In the following example: * a vpz file is loaded * initial conditions are modified * simulation is launched

Outputs are not recovered since the output plugin of the view is not storage.

>>> from pyvle import Vle
>>> f = Vle("scale1.vpz","glue")
>>> print(f)
<pyvle.Vle instance at 0x7fe231515e60>
>>> f.listConditions()
['condA', 'condB', 'condScale']
>>> f.listConditionPorts('condScale')
['InputTimeStep', 'OutputTimeStep']
>>> f.getConditionPortValues('condScale', 'OutputTimeStep')
1.0
>>> f.clearConditionPort('condScale', 'OutputTimeStep')
>>> f.addRealCondition('condScale', 'OutputTimeStep', 0.5)
>>> f.getConditionPortValues('condScale', 'OutputTimeStep')
0.5
>>> f.run()

Notes: * this example requires that the VLE package glue is installed. * outputs are not recovered since the output plugin of the view is not storage.

>>> from pyvle import Vle
>>> f=Vle('/tmp/myModel.vpz', 'glue')
>>> ...

It can be useful to recover the results of a simulation. To do this, the output plugin of views of the vpz (or at least the view that is expected) should be set to storage. Call to function run will return dictionaries (one per view) odf results.

Here is an example:

>>> from pyvle import Vle
>>> f = Vle("scale1.vpz","glue")
>>> f.setOutputPlugin('view', '', 'local', 'storage')
>>> result = f.run()
>>> result['view']['Top model:B4.b'][5]
20.0

To access to the ith value of observables obs into the view view, the syntax to use is: result[view][obs][i]:

  • view the name of the view
  • obs the path to the model concatenated with the port name. For example if the top model, called C1 is a coupled model containing the atomic model C2 which is observed through the observable port p, so obs = C1:C2.p
  • i index of the matrix. In a view configured as a timed-view, i represents the time step value.

Examples above show how to launch one simulation. Nevertheless it is possible to launch a bag of simulations with one call of command

In the case only one simulation is required, method run() can be used. In the other case, methods runManager or runManagerMatrix are available to simulate one bag of simulations.

For one simulation (see above for interpreting results):

>>> import pyvle
>>> v=pyvle.Vle('weed1.vpz','weed')
>>> result=v.run()
# result['view']['path:model.port'][index]

For a bag of simulations, results is more complex since they are as many view dictionaries as couples replica, combination.

>>> import pyvle
>>> v=pyvle.Vle('weed1.vpz','weed')
>>> result=v.runManager()
# result[replicat_index][combinaison_index]['view']['path:model.port'][index]

One can find the connection between combinaison_index and the initial conditions values it corresponds to.

>>> import pyvle
>>> v=pyvle.Vle('weed1.vpz','weed')
>>> v.getCombinations()
# n tuples containing the list of initial conditions of the n combinations
>>> v.getCombinations()[0]
#  the list of initial conditions of the first combination
>>> v.listConditions()
#the list of combinations
>>> listConditionPorts('xxx')
# the list of names of conditions port of condition 'xxx' where 'xxx' is the first condition