From 11012974b68c91e5f7c46b435a3665271b67a599 Mon Sep 17 00:00:00 2001
From: zxenia <ksenia1zaytseva@gmail.com>
Date: Wed, 28 Jul 2021 15:02:54 -0400
Subject: [PATCH 1/2] add schema validation in phenopacket ingestion

---
 chord_metadata_service/chord/ingest.py        | 22 ++++++++++++++++++-
 .../chord/tests/example_phenopacket.json      |  6 ++---
 .../phenopackets/schemas.py                   |  2 +-
 3 files changed, 24 insertions(+), 6 deletions(-)

diff --git a/chord_metadata_service/chord/ingest.py b/chord_metadata_service/chord/ingest.py
index 3f82af9cd..a43ed677a 100644
--- a/chord_metadata_service/chord/ingest.py
+++ b/chord_metadata_service/chord/ingest.py
@@ -8,6 +8,7 @@
 import shutil
 import tempfile
 import uuid
+import jsonschema
 
 from dateutil.parser import isoparse
 from typing import Callable
@@ -30,6 +31,7 @@
 )
 from chord_metadata_service.mcode.parse_fhir_mcode import parse_bundle
 from chord_metadata_service.mcode.mcode_ingest import ingest_mcodepacket
+from chord_metadata_service.phenopackets.schemas import PHENOPACKET_SCHEMA
 
 requests_unixsocket.monkeypatch()
 
@@ -344,9 +346,27 @@ def ingest_experiment(experiment_data, table_id) -> em.Experiment:
     return new_experiment
 
 
-def ingest_phenopacket(phenopacket_data, table_id) -> pm.Phenopacket:
+def ingest_phenopacket(phenopacket_data, table_id):
     """Ingests a single phenopacket."""
 
+    v = jsonschema.Draft7Validator(PHENOPACKET_SCHEMA, format_checker=jsonschema.FormatChecker())
+    try:
+        jsonschema.validate(
+            phenopacket_data,
+            PHENOPACKET_SCHEMA,
+            format_checker=jsonschema.FormatChecker(),
+        )
+        logger.info("JSON schema validation passed.")
+
+    except jsonschema.exceptions.ValidationError:
+        errors = [e for e in v.iter_errors(phenopacket_data)]
+        logger.info(f"JSON schema validation failed. Errors: {errors}")
+        for i, error in enumerate(errors, 1):
+            logger.error(
+                f"{i} Validation error in {'.'.join(str(v) for v in error.path)}: {error.message}",
+            )
+        return
+
     new_phenopacket_id = phenopacket_data.get("id", str(uuid.uuid4()))
 
     subject = phenopacket_data.get("subject")
diff --git a/chord_metadata_service/chord/tests/example_phenopacket.json b/chord_metadata_service/chord/tests/example_phenopacket.json
index 10d1d8f87..2da1a3dcc 100644
--- a/chord_metadata_service/chord/tests/example_phenopacket.json
+++ b/chord_metadata_service/chord/tests/example_phenopacket.json
@@ -13,8 +13,7 @@
                 "label": "Hematuria"
             },
             "negated": false,
-            "modifier": [],
-            "evidence": []
+            "modifier": []
         },
         {
             "description": "",
@@ -27,8 +26,7 @@
                 "id": "HP:0012828",
                 "label": "Severe"
             },
-            "modifier": [],
-            "evidence": []
+            "modifier": []
         }
     ],
     "diseases": [
diff --git a/chord_metadata_service/phenopackets/schemas.py b/chord_metadata_service/phenopackets/schemas.py
index 33817646b..7667891d2 100644
--- a/chord_metadata_service/phenopackets/schemas.py
+++ b/chord_metadata_service/phenopackets/schemas.py
@@ -360,5 +360,5 @@
         "meta_data": PHENOPACKET_META_DATA_SCHEMA,
         "extra_properties": EXTRA_PROPERTIES_SCHEMA
     },
-    "required": ["id", "meta_data"],
+    "required": ["meta_data"],
 }, descriptions.PHENOPACKET)

From 9d93e48e8391548f3a29abda04f024a481720414 Mon Sep 17 00:00:00 2001
From: zxenia <ksenia1zaytseva@gmail.com>
Date: Wed, 28 Jul 2021 15:16:55 -0400
Subject: [PATCH 2/2] update version in package.cfg

---
 chord_metadata_service/package.cfg | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/chord_metadata_service/package.cfg b/chord_metadata_service/package.cfg
index 49b50469d..b4b56f60e 100644
--- a/chord_metadata_service/package.cfg
+++ b/chord_metadata_service/package.cfg
@@ -1,4 +1,4 @@
 [package]
 name = katsu
-version = 2.2.0
+version = 2.2.1
 authors = Ksenia Zaytseva, David Lougheed, Simon Chénard, Romain Grégoire