VERSION

-0.2.3-rc3
+0.2.3-rc4

arkindex_worker/reporting.py

@@ -10,11 +10,13 @@ from arkindex_worker import logger
 
 
 class Reporter(object):
-    def __init__(self, name):
-        # TODO: use real data from workers
+    def __init__(
+        self, name="Unknown worker", slug="unknown-slug", version=None, **kwargs
+    ):
         self.report_data = {
-            "slug": name,
-            "version": "0.0",
+            "name": name,
+            "slug": slug,
+            "version": version,
             "started": datetime.utcnow().isoformat(),
             "elements": {},
         }

arkindex_worker/worker/__init__.py

@@ -39,9 +39,6 @@ class ElementsWorker(
     def __init__(self, description="Arkindex Elements Worker", support_cache=False):
         super().__init__(description, support_cache)
 
-        # Add report concerning elements
-        self.report = Reporter("unknown worker")
-
         # Add mandatory argument to process elements
         self.parser.add_argument(
             "--elements-list",
@@ -93,6 +90,13 @@ class ElementsWorker(
         ), "Worker must be configured to access its process activity state"
         return self.process_information.get("activity_state") == "ready"
 
+    def configure(self):
+        super().configure()
+        # Add report concerning elements
+        self.report = Reporter(
+            **self.worker_details, version=getattr(self, "worker_version_id", None)
+        )
+
     def run(self):
         """
         Process every elements from the provided list

arkindex_worker/worker/base.py

@@ -145,16 +145,19 @@ class BaseWorker(object):
                 f"Loaded worker {worker_version['worker']['name']} revision {worker_version['revision']['hash'][0:7]} from API"
             )
             self.config = worker_version["configuration"]["configuration"]
+            self.worker_details = worker_version["worker"]
             required_secrets = worker_version["configuration"].get("secrets", [])
         elif self.args.config:
             # Load config from YAML file
             self.config = yaml.safe_load(self.args.config)
+            self.worker_details = {"name": "Local worker"}
             required_secrets = self.config.get("secrets", [])
             logger.info(
                 f"Running with local configuration from {self.args.config.name}"
             )
         else:
             self.config = {}
+            self.worker_details = {}
             required_secrets = []
             logger.warning("Running without any extra configuration")
 

requirements.txt

 arkindex-client==1.0.8
 peewee==3.14.4
-Pillow==8.4.0
+Pillow==9.0.1
 python-gitlab==2.7.1
 python-gnupg==0.4.7
 sh==1.14.2

tests/test_reporting.py

 # -*- coding: utf-8 -*-
+import json
+import uuid
+from datetime import datetime
+from tempfile import NamedTemporaryFile
+
 import pytest
 from apistar.exceptions import ErrorResponse
 
@@ -6,10 +11,16 @@ from arkindex_worker.reporting import Reporter
 
 
 def test_init():
-    reporter = Reporter("worker")
+    version_id = str(uuid.uuid4())
+    reporter = Reporter(name="Worker", slug="worker-slug", version=version_id)
     assert "started" in reporter.report_data
     del reporter.report_data["started"]
-    assert reporter.report_data == {"slug": "worker", "version": "0.0", "elements": {}}
+    assert reporter.report_data == {
+        "name": "Worker",
+        "slug": "worker-slug",
+        "version": version_id,
+        "elements": {},
+    }
 
 
 def test_process():
@@ -213,3 +224,32 @@ def test_error():
             "content": '{"coffee": "Can\'t touch this"}',
         },
     ]
+
+
+def test_reporter_save(mocker):
+    datetime_mock = mocker.MagicMock()
+    datetime_mock.utcnow.return_value = datetime(2000, 1, 1)
+    mocker.patch("arkindex_worker.reporting.datetime", datetime_mock)
+    version_id = str(uuid.uuid4())
+    reporter = Reporter(name="Worker", slug="worker-slug", version=version_id)
+    reporter.add_element("myelement", type="text_line", type_count=4)
+    with NamedTemporaryFile() as f:
+        reporter.save(f.name)
+        exported_data = json.load(f)
+    assert exported_data == {
+        "name": "Worker",
+        "slug": "worker-slug",
+        "started": "2000-01-01T00:00:00",
+        "version": version_id,
+        "elements": {
+            "myelement": {
+                "classifications": {},
+                "elements": {"text_line": 4},
+                "entities": [],
+                "errors": [],
+                "metadata": [],
+                "started": "2000-01-01T00:00:00",
+                "transcriptions": 0,
+            }
+        },
+    }