Sort inputs of jobs for static identifier

Author: janten

iptk/utils/job.py

@@ -36,7 +36,8 @@ def add_input_dataset(self, dataset_id, path="/input"):
             "path": path
         }
         self.inputs.append(input)
-
+        self.sort_inputs()
+        
     def to_json(self):
         return json_pretty(self.spec)
 
@@ -52,6 +53,9 @@ def request_resource(self, resource_type, quantity):
         self.resource_requests[resource_type] = quantity
         return self.resource_requests
 
+    def sort_inputs(self):
+        self.inputs = sorted(self.inputs, key=lambda x: x.get("path", ""))
+    
     def enqueue(self, dataset_store):
         """
         Enqueues this job by saving it into the given DatasetStore.
@@ -82,6 +86,7 @@ def minimal_spec(self):
         not include optional fields that are only relevant to the job scheduler
         (e.g. resource requests)
         """
+        self.sort_inputs()
         keys = ["version", "image", "command", "inputs"]
         spec = {x: self.spec[x] for x in keys}
         return spec

setup.py

@@ -2,7 +2,7 @@
 setup(
   name = 'iptk',
   packages = ['iptk', 'iptk.metadata', 'iptk.utils'],
-  version = '0.5.1',
+  version = '0.6',
   description = 'Python interface to the Imaging Pipeline Toolkit',
   author = 'Jan-Gerd Tenberge',
   author_email = 'jan-gerd.tenberge@uni-muenster.de',