Feat: clean the workdirs at each iteration (#47)

Modify the `SelfConsistentHubbardWorkChain`
to clean the work directories at each iteration instead
of cleaning everything at the end, eventually saving
lots of storage space. This may be crucial for large scale
calculations or for users with low disk availability.

src/aiida_quantumespresso_hp/workflows/hubbard.py

@@ -172,6 +172,7 @@ def define(cls, spec):
         spec.inputs.validator = validate_inputs
         spec.inputs['hubbard']['hp'].validator = None
 
+        # yapf: disable
         spec.outline(
             cls.setup,
             while_(cls.should_run_iteration)(
@@ -188,10 +189,16 @@ def define(cls, spec):
                 ),
                 cls.run_hp,
                 cls.inspect_hp,
-                if_(cls.should_check_convergence)(cls.check_convergence,),
+                if_(cls.should_check_convergence)(
+                    cls.check_convergence,
+                ),
+                if_(cls.should_clean_workdir)(
+                    cls.clean_iteration,
+                )
             ),
             cls.run_results,
         )
+        # yapf: enable
 
         spec.output(
             'hubbard_structure',
@@ -673,14 +680,12 @@ def run_results(self):
         self.report(f'Hubbard parameters self-consistently converged in {self.ctx.iteration} iterations')
         self.out('hubbard_structure', self.ctx.current_hubbard_structure)
 
-    def on_terminated(self):
-        """Clean the working directories of all child calculations if `clean_workdir=True` in the inputs."""
-        super().on_terminated()
-
-        if self.inputs.clean_workdir.value is False:
-            self.report('remote folders will not be cleaned')
-            return
+    def should_clean_workdir(self):
+        """Whether to clean the work directories at each iteration."""
+        return self.inputs.clean_workdir.value
 
+    def clean_iteration(self):
+        """Clean all work directiories of the current iteration."""
         cleaned_calcs = []
 
         for called_descendant in self.node.called_descendants: