def batchRun(request, model):
"""
Sets up and executes model batch run.
Returns: ModelBatch object
"""
batchoutputmodule = importlib.import_module('.'+model+'_batchoutput', 'models.'+model)
batchOutputPageFunc = getattr(batchoutputmodule, model+'BatchOutputPage') # function name = 'model'BatchOutputPage (e.g. 'sipBatchOutputPage')
dataFrame = batchOutputPageFunc(request)
dataFrame = dataFrame.transpose()
logging.info("===== batch.batchRun")
logging.info(dataFrame)
# Convert DataFrame to JSON string
json_inputs = dataFrame.to_json()
# Add 'run_type' : 'batch' to the JSON string
json = '{"inputs":' + json_inputs + ',"run_type":"batch"}'
# logging.info(json)
# Send JSON to model_handler module
from models import model_handler
batch_output = model_handler.call_model_server(model, json)
ModelList = model_handler.generate_model_object_list(batch_output)
return ModelList
def qaqcRun(model):
"""
Sets up and executes the model QAQC run.
Returns: ModelQAQC object
"""
qaqcmodule = importlib.import_module('.'+model+'_qaqc', 'models.'+model)
#modelQAQC_obj = getattr(qaqcmodule, model+'_obj') # Calling model object, e.g. 'sip_obj'
csv_path = os.path.join(os.environ['PROJECT_PATH'], 'models', model, model+'_qaqc.csv')
modelQAQC_function = getattr(qaqcmodule, model+'Qaqc')
pandas_read_csv = modelQAQC_function(model, csv_path)
# Read CSV and create an DataFrame for inputs and expected outputs
pd_obj_inputs = pandas_read_csv[0]
pd_obj_exp_out = pandas_read_csv[1]
# Rename index column, renumber columns, and transpose the DataFrames
# pd_obj_inputs.index.name = None
# pd_obj_inputs.columns = pd_obj_inputs.columns - 1
# pd_obj_exp_out.index.name = None
# pd_obj_exp_out.columns = pd_obj_exp_out.columns - 1
pd_obj_in_out_transpose = pd_obj_inputs.transpose()
pd_obj_exp_out_transpose = pd_obj_exp_out.transpose()
# logging.info(pd_obj_inputs)
# logging.info(pd_obj_exp_out)
# logging.info(pd_obj_in_out_transpose)
# logging.info(pd_obj_exp_out_transpose)
"""
The DataFrame is now in correct format to be converted to JSON,
but the dtypes for all columns is 'object' (text) because of
the transpose(). When the DataFrame is recreated on the backend
the dtypes will be properly inferred from read_json() method.
"""
# Convert DataFrames to JSON strings
json_inputs = pd_obj_in_out_transpose.to_json()
json_exp_out = pd_obj_exp_out_transpose.to_json()
# logging.info(json_inputs)
# logging.info(json_exp_out)
# Concatenate JSON strings under keys: "inputs" & "out_exp", respectively,
# adding 'run_type' : 'qaqc' to the JSON string
json = '{"inputs":' + json_inputs + ',"out_exp":' + json_exp_out + ',"run_type":"qaqc"}'
# logging.info(json)
# Send JSON to model_handler module
from models import model_handler
# return model_handler.ModelQAQC(model, json)
qaqc_output = model_handler.call_model_server(model, json)
ModelList = model_handler.generate_model_object_list(qaqc_output)
return ModelList
