def output_target_to_eir(df, exp, model):
"""
Interpolates and adjusts the output target to the expected EIR (Entomological Inoculation Rate) values
based on input data for different models and experimental conditions.
Args:
df (pd.DataFrame): DataFrame containing simulation job data, where each row represents a specific condition
(e.g., `target_output_values`, `cm_clinical`, `seasonality`).
exp (Experiment): Experiment object that contains attributes such as `interp_path`, `target_output_name`, `pop_size`,
and others required for interpolation.
model (str): The model name used in the interpolation (e.g., 'malariasimulation', 'OpenMalaria').
Returns:
list: A list of interpolated EIR values corresponding to each row of the input DataFrame.
Raises:
ValueError: If the interpolation file is not found at the specified path.
ValueError: If the `target_output_name` specified in the experiment object is not valid.
Notes:
- The function performs interpolation based on the specified output type (`eir`, `prevalence_2to10`, or
`clinical_incidence_U5`).
- It applies scaling adjustments for population sizes, ensuring that the values are consistent with the
`pop_size` defined in the experiment.
- Warnings are issued if the interpolated values are outside the valid range (`<= 0` or `>= 1000`), with
adjustments made to these values accordingly.
"""
interp_path = os.path.join(exp.interp_path, model, exp.interp_csv)
if not os.path.exists(interp_path):
raise ValueError(f'interp_file not found: {interp_path}')
data = pd.read_csv(interp_path)
output = []
if exp.target_output_name == 'eir':
output_type_col = 'eir'
elif exp.target_output_name == 'prevalence_2to10':
output_type_col = 'prevalence_2to10'
elif exp.target_output_name == 'clinical_incidence_U5':
output_type_col = 'clinical_incidence_U5'
else:
raise ValueError(f'Specified input_type is not valid {exp.target_output_name}')
if model == 'EMOD':
popsize = exp.emod_pop_size
elif model == 'malariasimulation':
popsize = exp.malariasimulation_pop_size
elif model == 'OpenMalaria':
popsize = exp.openmalaria_pop_size
else:
raise ValueError(f'Specified model is not valid {model}')
# NOTE: Workaround to allow smaller populations for test runs
if exp.test and popsize not in data.loc[data['models'] == model, 'pop_size'].unique():
popsize = data.loc[data['models'] == model, 'pop_size'].min()
for i, row in df.iterrows():
idata = data[((data['cm_clinical'] == row['cm_clinical'])
& (data['cm_severe'] == row['cm_severe'])
& (data['seasonality'] == row['seasonality'])
& (data['models'] == model)
& (data['pop_size'] == popsize))]
xt = idata[output_type_col]
yp = idata['input_target']
xnew = row['target_output_values']
interpolated_value = np.interp(xnew, xt, yp)
if interpolated_value <= 0:
interpolated_value = 0.01
print("Warning: eir < 0, value set to 0.01 (this will likely cause elimination in the simulations)")
if interpolated_value >= 1000:
interpolated_value = 1000
print("Warning: eir > 1000, value set to 1000 (higher values cannot be reasonably simulated)")
output.append(interpolated_value)
return output