Input string value in QGIS 3 Processing Scripts

Question

I am trying to port my old QGIS 2.x processing scripts to QGIS 3.x. One of my scripts just needs two variables to initiate several processes to set-up a working environment in my QGIS map document (downloading layers, subsetting data, etc). Using the template script (QGIS 3.2.0) I have changed the static variables, 'INPUT' and 'OUTPUT' to 'START' and 'END'. I then changed the initAlogrithm function to:

 def initAlgorithm(self, config=None):
 """
 Here we define the inputs and output of the algorithm, along
 with some other properties.
 """
 self.addParameter(
 QgsProcessingParameterString(
 self.START,
 self.tr('Model Start Time'),
 'YYYYMMDDHH'
 )
 )
 self.addParameter(
 QgsProcessingParameterString(
 self.END,
 self.tr('Model End Time'),
 'YYYYMMDDHH'
 )
 )

Then in the processAlgorithm() function, I tried calling the inputs as:

start_time = self.parameterAsString(
 parameters,
 self.START,
 context)

This yields the following error:

TypeError: invalid result from AssimStartUp.processAlgorithm(), NoneType cannot be converted to a C/C++ QMap<QString,QVariant> in this context

So, what am I doing wrong that is prohibiting the string from the input window being passed into the processAlgorithm() function?

Answer 1

The error is caused because your processAlgorithm implementation isn't returning any values. If a python function has no return ... statement, it's treated as returning a None value, yet processAlgorithm must return a dictionary of results. Hence the error:

TypeError: invalid result from AssimStartUp.processAlgorithm(), NoneType cannot be converted to a C/C++ QMap<QString,QVariant> in this context

The easiest way to fix this is to ensure that your processingAlgorithm implementation ends with a return statement, e.g. return {}. But the correct approach is to ensure that this returned dictionary is populated with all the results of your algorithm, for instance paths to created layers, values for numeric outputs, etc. What to include would depend entirely on the logic of your particular algorithm.