From 3db09ba8836322b7752963fd1efe0665861c4ba3 Mon Sep 17 00:00:00 2001
From: Katy Baulch <46493669+katybaulch@users.noreply.github.com>
Date: Thu, 31 Oct 2024 15:32:50 +0000
Subject: [PATCH] Feature/pdct 1514 replace passed/approved from backend (#71)

* Add FIXMEs

* Update logic for family published date

* Validate the taxonomy

* Bump to 3.8.24

* Try adding expression for published_date

* Try adding expression for published_date

* Fix metadata check

* Fix logic for checking datetime_event)name

* Return None if can't find datetime_event_name in event meta

* Remove type ignore

* Error handling
---
 db_client/functions/metadata.py | 40 ++++++++++++++++++++++++++-------
 db_client/models/dfce/family.py | 18 ++++++++++++++-
 pyproject.toml                  |  2 +-
 3 files changed, 50 insertions(+), 10 deletions(-)

diff --git a/db_client/functions/metadata.py b/db_client/functions/metadata.py
index 5a953a2..348f79a 100644
--- a/db_client/functions/metadata.py
+++ b/db_client/functions/metadata.py
@@ -1,5 +1,5 @@
 import logging
-from typing import Mapping, Optional, Sequence, Union
+from typing import Any, Mapping, Optional, Sequence, Union
 
 from sqlalchemy.orm import Session
 
@@ -134,7 +134,7 @@ def build_valid_taxonomy(
     Takes the taxonomy from the database and builds a dictionary of
     TaxonomyEntry objects, used for validation.
 
-    :param Sequence taxonomy: From the database model
+    :param Mapping taxonomy: From the database model
         CorpusType.valid_metadata and potentially filtered by entity key
     :param Optional[TaxonomyDataEntry] metadata: The metadata to
         validate.
@@ -152,14 +152,38 @@ def build_valid_taxonomy(
     taxonomy_entries: Mapping[str, TaxonomyEntry] = {}
 
     for key, values in taxonomy.items():
-        # TODO: Can we do extra validation of the taxonomy here - e.g., compare the
-        # metadata datetime_event_name value against the list of allowed event types in
-        # the taxonomy. In fact - split any conditionals under this for loop into a
-        # separate validate_taxonomy function that we can isolate to test against.
-        if not isinstance(values, dict):
-            raise TypeError(f"Taxonomy entry for '{key}' is not a dictionary")
+        _validate_taxonomy(taxonomy, key, values)
 
         # We rely on pydantic to validate the values here
         taxonomy_entries[key] = TaxonomyEntry(**values)
 
     return taxonomy_entries
+
+
+def _validate_taxonomy(taxonomy: Mapping, key: str, values: Any) -> None:
+    """Extra validation of the taxonomy.
+
+    :param Mapping taxonomy: From the database model
+        CorpusType.valid_metadata and potentially filtered by entity key
+    :param str key: A taxonomy key.
+    :param Any values: Values for a taxonomy key.
+    :raises TypeError: If values is not a dictionary.
+    :raises ValueError: If too many datetime_event_name values.
+    :raises ValueError: If datetime_event_name value is not in list of
+        allowed event_type values.
+    """
+    if not isinstance(values, dict):
+        raise TypeError(f"Taxonomy entry for '{key}' is not a dictionary")
+
+    if key == "datetime_event_name":
+        # Compare the metadata datetime_event_name value against the list of allowed
+        # event_types under _event in the taxonomy.
+        datetime_event_name_values = values["allowed_values"]
+        if len(datetime_event_name_values) > 1:
+            raise ValueError(f"Too many values for taxonomy '{key}'")
+
+        datetime_event_name_value = datetime_event_name_values[0]
+        if datetime_event_name_value not in taxonomy["event_type"]["allowed_values"]:
+            raise ValueError(
+                f"Invalid value '{datetime_event_name_value}' for taxonomy '{key}'"
+            )
diff --git a/db_client/models/dfce/family.py b/db_client/models/dfce/family.py
index d282775..c48666f 100644
--- a/db_client/models/dfce/family.py
+++ b/db_client/models/dfce/family.py
@@ -1,3 +1,4 @@
+import logging
 from datetime import datetime
 from typing import Literal, Optional, cast
 
@@ -12,6 +13,8 @@
 
 from .geography import Geography
 
+_LOGGER = logging.getLogger(__name__)
+
 
 class FamilyCategory(BaseModelEnum):
     """Family categories as understood in the context of law/policy."""
@@ -144,9 +147,22 @@ def published_date(self) -> Optional[datetime]:
         """A date to use for filtering by published date."""
         if not self.events:
             return None
+
         date = None
         for event in self.events:
-            if event.event_type_name == "Passed/Approved":
+            event_meta = cast(dict, event.valid_metadata)
+            if "datetime_event_name" not in event_meta:
+                _LOGGER.error(event_meta)
+                return None
+
+            if not isinstance(event_meta["datetime_event_name"], list):
+                _LOGGER.error(
+                    f"datetime_event_name is type {type(event_meta['datetime_event_name'])}"
+                )
+                return None
+
+            datetime_event_name = event_meta["datetime_event_name"][0]
+            if event.event_type_name == datetime_event_name:
                 return cast(datetime, event.date)
             if date is None:
                 date = cast(datetime, event.date)
diff --git a/pyproject.toml b/pyproject.toml
index c60c54f..8808c9d 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "db-client"
-version = "3.8.23"
+version = "3.8.24"
 description = "All things to do with the datamodel and its storage. Including alembic migrations and datamodel code."
 authors = ["CPR-dev-team <tech@climatepolicyradar.org>"]
 license = "Apache-2.0"