Python Generic Modelling Engine (PyGME): A simple python package to design, run and calibrate models used in environmental sciences.
- pygme is a set of tools to create simple models and calibrate them via automatic optimizer
- pygme provides interface to classical hydrological models and allows you to create your own models.
- Create a suitable python environment. We recommend using miniconda combined with the environment specification provided in the [env_mini.yml] (env_mini.yml) file in this repository.
- Git clone this repository and run
pip install .
To setup a model, change its parameters and run it:
import numpy as np
from pygme.models.gr2m import GR2M
import matplotlib.pyplot as plt
# Get an instance of the GR2M monthly rainfall-runoff model
gr = GR2M()
# Generate random inputs (to be replaced by real data)
inputs = np.random.uniform(0, 10, size=(300, 2))
# Allocate model
# This step allocates all internal variables
# used for any runs of the model.
#
# The number of outputs is set to the maximum
# to generate all GR2M outputs. Default is 1
# which reduces the output variables to
# streamflow only.
gr.allocate(inputs, noutputs=gr.noutputsmax)
# Set parameters
gr.X1 = 500
gr.X2 = 0.8
# Initialise model
# Here we initialise both GR2M stores
gr.initialise([450, 55])
# Run model
gr.run()
# Plot results
# S: production store
# R: routing store
# AE: Actual evapotranspiration
# F: Inter-basin exchange
# Q: Streamflow
df = gr.to_dataframe()
fig = plt.figure(layout="tight")
mosaic = [[s] for s in ["S", "R", "AE", "F", "Q"]]
axs = fig.subplot_mosaic(mosaic)
for varname, ax in axs.items():
df.loc[:, varname].plot(ax=ax)
ax.set_title(varname)
plt.show()A set of examples is provided in the folder examples.
The source code and documentation of the pygme package is licensed under the BSD license.