fixed corrupt provenance object + serialization; added CODEOWNERS

CODEOWNERS

@@ -0,0 +1 @@
+*    @jblom @Veldhoen

models.py

@@ -1,6 +1,6 @@
-from dataclasses import dataclass
+from dataclasses import dataclass, field
 from enum import Enum
-from typing import Optional, TypedDict
+from typing import Optional, TypedDict, Dict
 
 
 # These are the types of output this worker (possibly) provides (depending on configuration)
@@ -35,8 +35,8 @@ class Provenance:
     input_data: dict[str, str]
     output_data: dict[str, str]
     parameters: Optional[dict] = None
-    software_version: Optional[dict[str, str]] = None
-    steps: Optional[list["Provenance"]] = None  # a list of subactivity provenance items
+    software_version: Optional[Dict[str, str]] = None
+    steps: Optional[list["Provenance"]] = field(default_factory=list["Provenance"])
 
     def to_json(self):
         return {
@@ -48,7 +48,7 @@ def to_json(self):
             "software_version": self.software_version,  # .to_json
             "input_data": self.input_data,  # .to_json
             "output_data": self.output_data,  # .to_json
-            "steps": [step.to_json for step in self.steps],
+            "steps": [step.to_json() for step in self.steps],
         }
 
 

provenance.py

@@ -1,4 +1,5 @@
 from functools import reduce
+import json
 import logging
 import os
 from time import time
@@ -39,8 +40,11 @@ def generate_full_provenance_chain(
     )
 
     output_file = get_provenance_file(input_file_path)
-    with open(output_file, "w+") as f:
-        f.write(str(provenance.to_json()))
+    fdata = provenance.to_json()
+    logger.info("Going to write the following to disk:")
+    logger.info(fdata)
+    with open(output_file, "w", encoding="utf-8") as f:
+        json.dump(fdata, f, ensure_ascii=False, indent=4)
         logger.info(f"Wrote provenance info to file: {output_file}")
     return provenance
 

worker.py

@@ -228,7 +228,6 @@ def callback(self, task: Task, doc: Document) -> CallbackResponse:
             "message": "Successfully generated VisXP data for the next worker",
         }
 
-    # TODO adapt to VisXP
     def save_to_dane_index(
         self,
         doc: Document,