GDAL processing output to memory in Pyqgis

Question

How do I save GDAL processing output rasters to a temporary layer? Using 'memory:' for vector processes works fine, but I couldn't get it to work for GDAL processes.

output_rast = self.parameterAsOutputLayer(parameters,self.OUTPUTR,context)

if True:
    magnitude_rast = processing.run("gdal:rasterize", {
        'INPUT':buffer_extent,
        'BURN': magnitude,
        'UNITS': 1,
        'WIDTH': 1.0,
        'HEIGHT': 1.0,
        'EXTENT': buffer_extent,
        'OUTPUT': 'memory:'
    }, context=context, feedback=feedback)['OUTPUT']

if True:
    source_rast = processing.run("gdal:rasterize", {
        'INPUT':parameters['SOURCE'],
        'BURN': 100,
        'UNITS': 1,
        'WIDTH': 1.0,
        'HEIGHT': 1.0,
        'EXTENT': buffer_extent,
        'OUTPUT': 'memory:'
    }, context=context, feedback=feedback)['OUTPUT']

if True:
    proximity_rast = processing.run("gdal:proximity", {
        'INPUT':source_rast,
        'BAND': 1,
        'VALUES': 100,
        'UNITS': 0,
        'DATA_TYPE': 5,
        'OUTPUT': output_rast
    }, context=context, feedback=feedback)['OUTPUT']
return {self.OUTPUTR: output_rast}

I've also tried using None for the output parameter as recommended here but I get an incorrect parameter value error.

Answer 1

You can use "OUTPUT": QgsProcessing.TEMPORARY_OUTPUT but this will "just" write the data to a file in your operating system's temporary directory. This is not necessarily stored in volatile memory (RAM).

I am not aware of an easy way to make GDAL Processing algorithms store their output in a RAM-backed location. Maybe one could use https://gdal.org/drivers/raster/mem.html ?