apsimNGpy: The Next Generation Agroecosytem Simulation Library
Our cutting-edge open-source framework, apsimNGpy, empowers advanced agroecosystem modeling through the utilization of object-oriented principles. It features fast batch file simulation, model prediction, evaluation, apsimx file editing, seamless weather data retrieval, and efficient soil profile development
- Dotnet, install from https://learn.microsoft.com/en-us/dotnet/core/install/
- Python3
- APSIM: Add the directory containing the models executable to the system's PATH or python path (to locate the required .dll files). This can be achieved in either of the following ways:
- Utilize the APSIM installer provided for this purpose.
- Build APSIM from its source code. This is comming soon
- Minimum; 8GM RAM, CPU Core i7
Installation
All versions are currently in development, phase and they can be installed as follows:
- Method 1. install from PyPI
pip install apsimNGpy- Method 1. clone the current development repositry
git clone https://github.com/MAGALA-RICHARD/apsimNGpy.git@dev
cd apsimNGpy
pip install .- Method 2. Use pip straight away and install from github
pip install git+https://github.com/MAGALA-RICHARD/apsimNGpy.git@devIf you have apsim installed and the program refuses to load run the following code at the top of your python script before importing any apsimNGpy class, especially class from ApsimNGpy.core modules The classes are CamelCased.
# search for the program binary installation path and add to os.environ as follows
import os
# A more intuitive way is to use apsimNGpy config Module
from apsimNGpy.config import Config
# now set the path to ASPIMX binaries
Config.set_aPSim_bin_path(path = r'path/toyourapsimbinaryfolder/bin)
# in the pythonnet_config module, priority is first given to the user supplied binary path, we also search through the python global env using the os module,
# if that fail it searches through other sources such as the user program installation folders.
# Not sure whether this can work all the time but you can try changing through os.environ as follows:
os.environ['APSIM'] =r'path/toyourapsimbinaryfolder/bin
# or
os.environ['Models'] =r'path/toyourapsimbinaryfolder/bin
# alternatively, you can add the path to the system environmental variables. if this is the case the shutil.which method is used to retrieve that path
# if all approaches have been tried and nothing has been returned, I assure you that a value errors will be raised
# now we are than we can import any module attached to pythonnet
# try importing SoilModel class
from apsimNGpy.core.apsim import ApsimModelThe above code is also applicable for running different versions of APSIM models. Please note that if your APSIM installation hasn't been added to the system path, this script line should always be placed at the beginning of your simulation script.
- numpy
- pandas
- pythonnet
- xmltodict
- tqdm
- requests
Please note that apsimNGpy is tested on Python 3. We are not aware of its performance in Python 2 because it utilizes some of the new libraries like pathlib and f-strings.
Usage
import apsimNGpy
from apsimNGpy.core.base_data import LoadExampleFiles
from apsimNGpy.core.apsim import ApsimModel as SoilModel
from pathlib import Path
import os
from apsimNGpy.validation.visual import plot_data
cwd = Path.cwd().home() # sending this to your home folder
wd = cwd.joinpath("apsimNGpy_demo")
if not wd.exists():
os.mkdir(wd)
# change directory
os.chdir(wd)
# Create the data
data = LoadExampleFiles(wd)
# Get maize model
maize = data.get_maize
# Alternatively, you can laod from the factory default modules
soybean_model = load_default_simulations(crop = 'soybean') # don't worry it is not case senstive
#the load_default_simulation returns a prelloaded model ready to run the existing module
# Initialize the simulation methods
apsim = SoilModel(maize, copy=True)
# Run the file
apsim.run() # use run to print time taken to excute or run the model
# print the results
print(apsim.results) # prints all data frames in the storage domain subset usign report names
# check the manager modules in the apsim simulation file
# first get the simualtion names
sim_name = apsim.simulation_names
apsim.examine_management_info(simulations=sim_name)
# show current simulation in apsim GUI
# plot the data
res = apsim.results['MaizeR']
plot_data(res.Year, res.Yield, xlabel='Years', ylabel=" Maize Yield (kg/ha)")A graph should be able to appear like the ones below. Note that plot_data function just wraps matplotlib plot function for quick visualisation
import apsimNGpy
from apsimNGpy.core.base_data import LoadExampleFiles
from apsimNGpy.core.apsim import ApsimModel as SoilModel
from pathlib import Path
import os
from apsimNGpy.validation.visual import plot_data
cwd = Path.cwd().home() # sending this to your home folder
wd = cwd.joinpath("apsimNGpy_demo")
if not wd.exists():
os.mkdir(wd)
# change directory
os.chdir(wd)
# Create the data
data = LoadExampleFiles(wd)
# Get maize model
maize = data.get_maize
# Initialize the simulation methods
apsim = SoilModel(maize, copy=True)
apsim.change_simulation_dates(start_date='01/01/1998', end_date='12/31/2010')# First, examine the manager scripts in the simulation node
apsim.examine_management_info()
# now create dictionary holding the parameters. the key to this is that the name of the script manage must be
passed in the dictionary.
# in this node we have a script named the Simple Rotation,we want to change the rotation to maybe Maize, Wheat or
something else
rotation = {'Name': "Simple Rotation", "Crops": 'Maize, Wheat, Soybean'}, # the crops must be seperated my commas
apsim.update_mgt(management = rotation, reload=True)
# now you cans see we passed rotation as aturple. That means you can add other scripts as your needs suggest. They will all be changed at the
same timefrom apsimNGpy.core.weather import daymet_bylocation_nocsv
lonlat = -93.08, 42.014
start_year, end_year = 2000, 2002
wf = daymet_bylocation_nocsv(lonlat, startyear, endyear, filename="mymet.met")
# you may need to first see what file currently exists in the model
mis = apsim.show_met_file_in_simulation()
print(mis)
# change
apsim.replace_met_file(weather_file=wf)
# check again if you want to
mis = apsim.show_met_file_in_simulation()
print(mis)The apsimNGpy Python package provides a convenient way to validate model simulations against measured data. Below is a step-by-step guide on how to use the validation.evaluator module from apsimNGpy.
# Start by importing the required libraries
from apsimNGpy.validation.evaluator import validate
import pandas as pd
# Load the data if external. Replace with your own data
df = pd.read_csv('evaluation.csv')
apsim_results = apsim.results # Assuming 'apsim' is a predefined object from aopsimNGpy.core.core.APSIMN class and contains your simualted results
# Preparing Data for Validation
# Extract the relevant columns from your DataFrame for comparison. In this example, we use
# 'Measured' for observed values and compare them with different model outputs:
measured = df['Measured']
predicted = apsim_results['MaizeR'].Yield
# Now we need to pass both the measured and the observed in the validate class
val = validate(measured, predicted)
# Both variables should be the same length, and here we are assuming that they are sorted in the corresponding order
# There are two options:
# 1. Evaluate all
metrics = val.evaluate_all(verbose=True)
# Setting verbose=True prints all the results on the go; otherwise, a dictionary is returned with the value for each metric
# 2. Select or pass your desired metric
RMSE = val.evaluate("RMSE")
print(RMSE)
# If you want to see the available metrics, use the code below
available_metrics = metrics.keys()
print(available_metrics)
# Then select your choice from the listWe welcome contributions from the community, whether they are bug fixes, enhancements, documentation updates, or new features. Here's how you can contribute to apsimNGpy:
If you find a bug or have a suggestion for improving apsimNGpy, please first check the Issue Tracker to see if it has already been reported. If it hasn't, feel free to submit a new issue. Please provide as much detail as possible, including steps to reproduce the issue, the expected outcome, and the actual outcome.
We accept code contributions via Pull Requests (PRs). Here are the steps to contribute:
Start by forking the apsimNGpy repository on GitHub. This creates a copy of the repo under your GitHub account.
Clone your fork to your local machine:
git clone https://github.com/MAGALA-RICHARD/apsimNGpy.git cd apsimNGpy
- Create a New Branch
Create a new branch for your changes:
git checkout -b your-branch-name
- Make Your Changes
- Make the necessary changes or additions to the codebase. Please try to adhere to the coding style already in place.
- Test Your Changes
- Run any existing tests, and add new ones if necessary, to ensure your changes do not break existing functionality.
- Commit Your Changes
Commit your changes with a clear commit message that explains what you've done:
git commit -m "A brief explanation of your changes"- Push to GitHub
Push your changes to your fork on GitHub:
git push origin your-branch-name
- Submit a Pull Request
- Go to the
apsimNGpyrepository on GitHub, and you'll see a prompt to submit a pull request based on your branch. Click on "Compare & pull request" and describe the changes you've made. Finally, submit the pull request.
Improvements or updates to documentation are greatly appreciated. You can submit changes to documentation with the same process used for code contributions.
Feel free to join in discussions on issues or pull requests. Your feedback and insights are valuable to the community!
Dynamic handling of simulations and their properties
replacements made easier
object oriented factorial experiment set ups and simulations
This project, ApsimNGpy, greatly appreciates the support and contributions from various organizations and initiatives that have made this research possible. We extend our gratitude to Iowa State University's C-CHANGE Presidential Interdisciplinary Research Initiative, which has played a pivotal role in the development of this project. Additionally, our work has been significantly supported by a generous grant from the USDA-NIFA Sustainable Agricultural Systems program (Grant ID: 2020-68012-31824), underscoring the importance of sustainable agricultural practices and innovations.
We would also like to express our sincere thanks to the APSIM Initiative. Their commitment to quality assurance and the structured innovation program for APSIM's modelling software has been invaluable. APSIM's software, which is available for free for research and development use, represents a cornerstone for agricultural modeling and simulation. For further details on APSIM and its capabilities, please visit www.apsim.info.
Our project stands on the shoulders of these partnerships and support systems, and we are deeply thankful for their contribution to advancing agricultural research and development. Please not that that this library is designed as a bridge to APSIM software, and we hope that by using this library, you have the appropriate APSIM license to do so whether free or commercial.
Lastly but not least, ApsimNGpy is not created in isolation but draws inspiration from apsimx, an R package (https://cran.r-project.org/web/packages/apsimx/vignettes/apsimx.html). We acknowledge and appreciate the writers and contributors of apsimx for their foundational work. ApsimNGpy is designed to complement apsimx by offering similar functionalities and capabilities in the Python ecosystem.