From 956bf30d5c6f9cba7efa2c84f09fc63ace744600 Mon Sep 17 00:00:00 2001
From: Thomas A Caswell <tcaswell@bnl.gov>
Date: Fri, 27 Aug 2021 20:03:35 -0400
Subject: [PATCH] ENH: add dtype_str and dtype_descr to descriptor schema

---
 event_model/schemas/event_descriptor.json | 24 +++++++++
 event_model/tests/test_em.py              | 62 +++++++++++++++--------
 2 files changed, 64 insertions(+), 22 deletions(-)

diff --git a/event_model/schemas/event_descriptor.json b/event_model/schemas/event_descriptor.json
index 50136867e..661b90bb9 100644
--- a/event_model/schemas/event_descriptor.json
+++ b/event_model/schemas/event_descriptor.json
@@ -1,5 +1,15 @@
     {
         "definitions": {
+            "np_dtype": {
+                "title": "nd_dtype",
+                "type": ["array", "string"],
+                "pattern": "[|<>][tbiufcmMOSUV][0-9]+",
+                "items": {
+                    "type": "array",
+                    "items": [{"type": "string"}, {"$ref": "#/definitions/np_dtype"}],
+                    "additionalItems": {"type": "array", "items": {"type":"integer"}}
+                }
+            },
             "data_key": {
                 "title": "data_key",
                 "description": "Describes the objects in the data property of Event documents",
@@ -15,6 +25,20 @@
                         "type": "string",
                         "description": "The type of the data in the event."
                     },
+                    "dtype_str" : {
+                        "type": "string",
+                        "description": "The __array_protocol__ typestring of the 'inner' data",
+                        "pattern": "[|<>][tbiufcmMOSUV][0-9]+"
+                    },
+                    "dtype_descr" : {
+                        "type": "array",
+                        "description": "The __array_protocol__ description of the 'inner' data.  This maybe nested.",
+                        "items": {
+                            "type": "array",
+                            "items": [{"type": "string"}, {"$ref": "#/definitions/np_dtype"}],
+                            "additionalItems": {"type": "array", "items": {"type":"integer"}}
+                        }
+                    },
                     "external": {
                         "pattern": "^[A-Z]+:?",
                         "type": "string",
diff --git a/event_model/tests/test_em.py b/event_model/tests/test_em.py
index ba3ec00a1..652b671c8 100644
--- a/event_model/tests/test_em.py
+++ b/event_model/tests/test_em.py
@@ -1,12 +1,14 @@
 from distutils.version import LooseVersion
 import json
 import pickle
+import numpy as np
+
 
 import event_model
 import jsonschema
 import numpy
 import pytest
-
+from event_model._numpy import infer_datakeys
 
 JSONSCHEMA_2 = LooseVersion(jsonschema.__version__) < LooseVersion("3.0.0")
 
@@ -46,39 +48,55 @@ def test_compose_run():
     assert bundle.compose_descriptor is compose_descriptor
     assert bundle.compose_resource is compose_resource
     assert bundle.compose_stop is compose_stop
+    motor_data = 5.0
+    counter_data = 10
+    image_data = np.zeros((512, 512))
     bundle = compose_descriptor(
-        data_keys={'motor': {'shape': [], 'dtype': 'number', 'source': '...'},
-                   'image': {'shape': [512, 512], 'dtype': 'array',
-                             'source': '...', 'external': 'FILESTORE:'}},
-        name='primary')
+        data_keys={
+            "motor": {"source": "...", **infer_datakeys(motor_data)},
+            "counter": {"source": "...", **infer_datakeys(counter_data)},
+            "image": {
+                **infer_datakeys(image_data),
+                "source": "...",
+                "external": "FILESTORE:",
+            },
+        },
+        name="primary",
+    )
     descriptor_doc, compose_event, compose_event_page = bundle
     assert bundle.descriptor_doc is descriptor_doc
     assert bundle.compose_event is compose_event
     assert bundle.compose_event_page is compose_event_page
     bundle = compose_resource(
-        spec='TIFF', root='/tmp', resource_path='stack.tiff',
-        resource_kwargs={})
+        spec="TIFF", root="/tmp", resource_path="stack.tiff", resource_kwargs={}
+    )
     resource_doc, compose_datum, compose_datum_page = bundle
     assert bundle.resource_doc is resource_doc
     assert bundle.compose_datum is compose_datum
     assert bundle.compose_datum_page is compose_datum_page
-    datum_doc = compose_datum(datum_kwargs={'slice': 5})
+    datum_doc = compose_datum(datum_kwargs={"slice": 5})
     event_doc = compose_event(
-        data={'motor': 0, 'image': datum_doc['datum_id']},
-        timestamps={'motor': 0, 'image': 0}, filled={'image': False})
-    datum_page = compose_datum_page(datum_kwargs={'slice': [10, 15]})
-    event_page = compose_event_page(data={'motor': [1, 2], 'image':
-                                          datum_page['datum_id']},
-                                    timestamps={'motor': [0, 0],
-                                                'image': [0, 0]},
-                                    filled={'image': [False, False]},
-                                    seq_num=[1, 2])
-    assert 'descriptor' in event_doc
-    assert 'descriptor' in event_page
-    assert event_doc['seq_num'] == 1
+        data={"motor": 0.0, "counter": 1, "image": datum_doc["datum_id"]},
+        timestamps={"motor": 0, "counter": 0, "image": 0},
+        filled={"image": False},
+    )
+    datum_page = compose_datum_page(datum_kwargs={"slice": [10, 15]})
+    event_page = compose_event_page(
+        data={
+            "motor": [1.0, 2.0],
+            "image": datum_page["datum_id"],
+            "counter": [10, 11],
+        },
+        timestamps={"motor": [0, 0], "image": [0, 0], "counter": [0, 0]},
+        filled={"image": [False, False]},
+        seq_num=[1, 2],
+    )
+    assert "descriptor" in event_doc
+    assert "descriptor" in event_page
+    assert event_doc["seq_num"] == 1
     stop_doc = compose_stop()
-    assert 'primary' in stop_doc['num_events']
-    assert stop_doc['num_events']['primary'] == 3
+    assert "primary" in stop_doc["num_events"]
+    assert stop_doc["num_events"]["primary"] == 3
 
 
 def test_round_trip_pagination():