most unit tests are passing now and the use of bpmn_json is almost gone in src

spiffworkflow-backend/src/spiffworkflow_backend/routes/process_api_blueprint.py

@@ -25,6 +25,7 @@ from spiffworkflow_backend.models.process_instance import ProcessInstanceModelSc
 from spiffworkflow_backend.models.process_instance import (
     ProcessInstanceTaskDataCannotBeUpdatedError,
 )
+from spiffworkflow_backend.models.process_instance_data import ProcessInstanceDataModel
 from spiffworkflow_backend.models.process_model import ProcessModelInfo
 from spiffworkflow_backend.models.spec_reference import SpecReferenceCache
 from spiffworkflow_backend.models.spec_reference import SpecReferenceSchema
@@ -194,16 +195,25 @@ def task_data_update(
                 f" It is currently: {process_instance.status}"
             )
 
-        process_instance_bpmn_json_dict = json.loads(process_instance.bpmn_json)
+        process_instance_data = ProcessInstanceDataModel.query.filter_by(process_instance_id=process_instance.id).first()
+        if process_instance_data is None:
+            raise ApiError(
+                error_code="process_instance_data_not_found",
+                message=(
+                    f"Could not find task data related to process instance: {process_instance.id}"
+                ),
+            )
+        process_instance_data_dict = json.loads(process_instance_data.runtime_json)
+
         if "new_task_data" in body:
             new_task_data_str: str = body["new_task_data"]
             new_task_data_dict = json.loads(new_task_data_str)
-            if task_id in process_instance_bpmn_json_dict["tasks"]:
+            if task_id in process_instance_data_dict["tasks"]:
-                process_instance_bpmn_json_dict["tasks"][task_id][
+                process_instance_data_dict["tasks"][task_id][
                     "data"
                 ] = new_task_data_dict
-                process_instance.bpmn_json = json.dumps(process_instance_bpmn_json_dict)
+                process_instance_data.runtime_json = json.dumps(process_instance_data_dict)
-                db.session.add(process_instance)
+                db.session.add(process_instance_data)
                 try:
                     db.session.commit()
                 except Exception as e:

spiffworkflow-backend/src/spiffworkflow_backend/routes/process_instances_controller.py

@@ -1,5 +1,6 @@
 """APIs for dealing with process groups, process models, and process instances."""
 import json
+from spiffworkflow_backend.models.process_instance_data import ProcessInstanceDataModel
 from typing import Any
 from typing import Dict
 from typing import Optional
@@ -550,9 +551,12 @@ def process_instance_task_list(
         )
 
     step_details = step_detail_query.all()
-    bpmn_json = json.loads(process_instance.bpmn_json or "{}")
+
-    tasks = bpmn_json["tasks"]
+    process_instance_data = ProcessInstanceDataModel.query.filter_by(process_instance_id=process_instance.id).first()
-    subprocesses = bpmn_json["subprocesses"]
+    process_instance_data_json = "{}" if process_instance_data is None else process_instance_data.runtime_json
+    process_instance_data_dict = json.loads(process_instance_data_json)
+    tasks = process_instance_data_dict["tasks"]
+    subprocesses = process_instance_data_dict["subprocesses"]
 
     steps_by_id = {step_detail.task_id: step_detail for step_detail in step_details}
 
@@ -584,7 +588,7 @@ def process_instance_task_list(
                 spiff_task_id, TaskState.FUTURE
             )
 
-    process_instance.bpmn_json = json.dumps(bpmn_json)
+    process_instance_data.runtime_json = json.dumps(process_instance_data_dict)
 
     processor = ProcessInstanceProcessor(process_instance)
     spiff_task = processor.__class__.get_task_by_bpmn_identifier(

spiffworkflow-backend/src/spiffworkflow_backend/services/process_instance_processor.py

@@ -520,9 +520,9 @@ class ProcessInstanceProcessor:
         )
 
     @classmethod
-    def get_full_bpmn_json(self, process_instance_model: ProcessInstanceModel) -> Optional[dict]:
+    def get_full_bpmn_json(cls, process_instance_model: ProcessInstanceModel) -> dict:
         if process_instance_model.serialized_bpmn_definition_id is None:
-            return None
+            return {}
         serialized_bpmn_definition = process_instance_model.serialized_bpmn_definition
         process_instance_data = ProcessInstanceDataModel.query.filter_by(process_instance_id=process_instance_model.id).first()
         # if process_instance_data is not None:
@@ -848,19 +848,32 @@ class ProcessInstanceProcessor:
                 )
         return subprocesses_by_child_task_ids
 
-    def save(self) -> None:
+    def add_bpmn_json_records(self) -> None:
-        """Saves the current state of this processor to the database."""
+        bpmn_dict = json.loads(self.serialize())
-        # self.process_instance_model.bpmn_json = self.serialize()
+        bpmn_dict_keys = ('spec', 'subprocess_specs', 'serializer_version')
-        bpmn_json = self.serialize()
+        bpmn_spec_dict = {}
+        process_instance_data_dict = {}
+        for bpmn_key in bpmn_dict.keys():
+            if bpmn_key in bpmn_dict_keys:
+                bpmn_spec_dict[bpmn_key] = bpmn_dict[bpmn_key]
+            else:
+                process_instance_data_dict[bpmn_key] = bpmn_dict[bpmn_key]
 
         if self.process_instance_model.serialized_bpmn_definition_id is None:
-            new_hash = {k: bpmn_json[k] for k in ('spec', 'subprocess_spec', 'serializer_version')}
+            new_hash_digest = sha256(json.dumps(bpmn_spec_dict, sort_keys=True).encode('utf8')).hexdigest()
-            new_hash_digest = sha256(json.dumps(new_hash, sort_keys=True).encode('utf8')).hexdigest()
+            serialized_bpmn_definition = SerializedBpmnDefinitionModel.query.filter_by(hash=new_hash_digest).first()
-            serialized_bpmn_definition = SerializedBpmnDefinitionModel(hash=new_hash_digest).first()
             if serialized_bpmn_definition is None:
-                serialized_bpmn_definition = SerializedBpmnDefinitionModel(hash=new_hash_digest, static_json=json.dumps(new_hash))
+                serialized_bpmn_definition = SerializedBpmnDefinitionModel(hash=new_hash_digest, static_json=json.dumps(bpmn_spec_dict))
                 db.session.add(serialized_bpmn_definition)
 
+        process_instance_data = ProcessInstanceDataModel.query.filter_by(process_instance_id=self.process_instance_model.id)
+        if process_instance_data is None:
+            process_instance_data = ProcessInstanceDataModel(process_instance_id=self.process_instance_model.id)
+        process_instance_data.runtime_json = json.dumps(process_instance_data_dict)
+
+    def save(self) -> None:
+        """Saves the current state of this processor to the database."""
+        self.add_bpmn_json_records()
 
         complete_states = [TaskState.CANCELLED, TaskState.COMPLETED]
         user_tasks = list(self.get_all_user_tasks())