EasyModeler is a package for calibration and validation of Ordinary Differential Equations ODEs to sample data.
- Python 2.6 or 2.7
- SciPy and NumPy 2.6 or 2.7
- Matplotlib 2.6 or 2.7
- sas7bdat
- ODEINT Wrapper Intelligent non-invasive wrapper to SciPy's integrator
- ODE Calibration Auto-calibrate a series of ODEs to data
- TimeSeries Files Handling of dtInput
- Model Validation Validate using Goodness of Fit statistics
- Graphical Plotting Basic plotting via matplotlib
- Graphical Interface Coming in version 2.3
- https://dl.dropboxusercontent.com/u/66459905/site/index.html
- Supports comprehensive autodocs with example usage inside source!
- Looking for a permanent document home online please suggest ideas to me!
$ easy_install easymodeler
$ unzip easymodeler-x.x.x.zip $ cd easymodler-x.x.x $ python setup.py install
- bugfixes
- added RMSD GOF parameter
- bugfixes
- bugfixes
- bugfixes
- dtinput fixes
- example dataset inclusion
- SAS filetype support
- fixes to calibration
- autodocs continue to update
- Additions to Calibration object
- GraphOpt object creation
- Rollout of simple plotting interface
- autodocs continue to update
- trying yet again to fix the pypi readme
- autodocs continue to update
- rename functions to naming conventions
- autodocs continue to update
- README change
- Sample Example
- LICENSE
Support for this project was made possible by grant number NA11NOS0120024 from NOAA to support the Gulf of Mexico Coastal Ocean Observing System (GCOOS) via subcontract S120015 from the TAMU Research Foundation.
Here is a snippet of the userguide available at https://dl.dropboxusercontent.com/u/66459905/site/index.html
The Lotka Volterra system is a simple model of predator-prey dynamics and consists of two coupled differentials. http://en.wikipedia.org/wiki/Lotka%E2%80%93Volterra_equation
This is a simple example highlighting EasyModler's ability to integrate ODEs without complication! At a minimum to integrate we require:
- A defined ODE function
- A set of initial conditions as a list
- Number of times to run the integrator
Declare an ODE_INT function in your source code. This will be passed to the scipy.integrate.odeint integrator
def LV_int(t,initial,dtinput,coefficients): x = initial[0] y = initial[1] A = 1 B = 1 C = 1 D = 1 x_dot = (A * x) - (B * x *y) y_dot = (D * x * y) - (C * y) return [x_dot, y_dot]
Pass the ODE function to emlib.Model as
>>> import emlib >>> LV = emlib.Model(LV_int) INFO -512- New Model(1): LV_int INFO -524- No algorithm supplied assuming vode/bfd O12 Nsteps3000
Now lets integrate our LV function for 200 timesteps!
>>> LV.Integrate([1,1],maxdt=200) DEBUG -541- ODEINT Initials:11 DEBUG -579- Ending in 200 runs DEBUG -600- Integration dT:0 of 200 Remaining:200 DEBUG -612- Completed Integration, created np.array shape:(200, 2)
The model output is stored in the emlib.Model object as arrays computedT and computed
>>> print LV.computed [[ 0.37758677 2.93256414] [ 0.13075395 1.32273451] [ 0.14707288 0.55433421] [ 0.27406944 0.24884565]
EasyModeler is organized where time is stored separately from data. This is a design feature to aid processing timeseries data.