Merge branch 'main' into feature/mx-1679_ingest_wikidata_primary_source

robert-koch-institut · Feb 27, 2025 · 2d45085 · 2d45085
2 parents c365a55 + 952c1e8
commit 2d45085
Show file tree

Hide file tree

Showing 15 changed files with 93 additions and 66 deletions.
diff --git a/.cruft.json b/.cruft.json
@@ -1,14 +1,14 @@
 {
   "template": "https://github.com/robert-koch-institut/mex-template",
-  "commit": "5446dae8486bd2412e75f97790c6bd800fa26a34",
+  "commit": "8bdad06a5eca52607f046b75cb0cdd9d7ab4912e",
   "checkout": null,
   "context": {
     "cookiecutter": {
       "project_name": "backend",
       "short_summary": "Backend server for the RKI metadata exchange.",
       "long_summary": "The `mex-backend` package is a multi-purpose backend application with an HTTP-API. It provides endpoints to ingest data from ETL-pipelines, for a metadata editor application, and for publishing pipelines to extract standardized data for use in upstream frontend applications.",
       "_template": "https://github.com/robert-koch-institut/mex-template",
-      "_commit": "5446dae8486bd2412e75f97790c6bd800fa26a34"
+      "_commit": "8bdad06a5eca52607f046b75cb0cdd9d7ab4912e"
     }
   },
   "directory": null

diff --git a/.dockerignore b/.dockerignore
@@ -117,6 +117,11 @@ dmypy.json
 # SQLite databases
 *.db
 
+# Reflex
+.states
+assets/external/
+.web
+
 # Default exports
 *.ndjson
 data/

diff --git a/.gitignore b/.gitignore
@@ -121,6 +121,11 @@ dmypy.json
 # SQLite databases
 *.db
 
+# Reflex
+.states
+assets/external/
+.web
+
 # Default exports
 *.ndjson
 data/

diff --git a/.pre-commit-config.yaml b/.pre-commit-config.yaml
@@ -3,7 +3,7 @@ default_language_version:
   python: python3.11
 repos:
   - repo: https://github.com/astral-sh/ruff-pre-commit
-    rev: v0.9.1
+    rev: v0.9.6
     hooks:
       - id: ruff
         args: [--fix, --exit-non-zero-on-fix]
@@ -25,7 +25,7 @@ repos:
       - id: fix-byte-order-marker
         name: byte-order
   - repo: https://github.com/pdm-project/pdm
-    rev: 2.22.2
+    rev: 2.22.3
     hooks:
       - id: pdm-lock-check
         name: pdm

diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -21,6 +21,11 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ### Security
 
+## [0.30.2] - 2025-02-27
+
+### Changes
+- preview and merged endpoint filter by multiple primary sources (was: one)
+
 ## [0.30.1] - 2025-02-17
 
 ### Added

diff --git a/mex/backend/graph/connector.py b/mex/backend/graph/connector.py
@@ -291,7 +291,7 @@ def fetch_merged_items(  # noqa: PLR0913
         query_string: str | None,
         identifier: str | None,
         entity_type: Sequence[str] | None,
-        had_primary_source: str | None,
+        had_primary_source: Sequence[str] | None,
         skip: int,
         limit: int,
     ) -> Result:
@@ -320,7 +320,7 @@ def fetch_merged_items(  # noqa: PLR0913
             query_string=query_string,
             identifier=identifier,
             labels=entity_type or list(MERGED_MODEL_CLASSES_BY_NAME),
-            referenced_identifier=had_primary_source,
+            referenced_identifiers=had_primary_source,
             skip=skip,
             limit=limit,
         )

diff --git a/mex/backend/graph/cypher/fetch_extracted_or_rule_items.cql b/mex/backend/graph/cypher/fetch_extracted_or_rule_items.cql
@@ -8,7 +8,7 @@ Globals:
 
 Args:
     filter_by_query_string: Whether the final query should accept a full text query string
-    filter_by_stable_target_id: Whether the final query should filter by stableTargetId
+    filter_by_stable_target_id: Whether the final query should filter by stableTargetIds
 
 Returns:
     total: Count of all items found by this query

diff --git a/mex/backend/graph/cypher/fetch_merged_items.cql b/mex/backend/graph/cypher/fetch_merged_items.cql
@@ -50,7 +50,7 @@ CALL () {
         AND merged_node.identifier = $identifier
         <%- endif %>
         <%- if filter_by_reference_to_merged_item %>
-        AND referenced_merged_node_to_filter_by.identifier = $referenced_identifier
+        AND referenced_merged_node_to_filter_by.identifier IN $referenced_identifiers
         <%- endif %>
 <%- endblock %>
     RETURN COUNT(merged_node) AS total

diff --git a/mex/backend/merged/helpers.py b/mex/backend/merged/helpers.py
@@ -80,7 +80,7 @@ def search_merged_items_in_graph(
     query_string: str | None = None,
     identifier: str | None = None,
     entity_type: list[str] | None = None,
-    had_primary_source: str | None = None,
+    had_primary_source: list[str] | None = None,
     skip: int = 0,
     limit: int = 100,
     validate_cardinality: Literal[False] = False,
@@ -92,7 +92,7 @@ def search_merged_items_in_graph(
     query_string: str | None = None,
     identifier: str | None = None,
     entity_type: list[str] | None = None,
-    had_primary_source: str | None = None,
+    had_primary_source: list[str] | None = None,
     skip: int = 0,
     limit: int = 100,
     validate_cardinality: Literal[True] = True,
@@ -103,7 +103,7 @@ def search_merged_items_in_graph(  # noqa: PLR0913
     query_string: str | None = None,
     identifier: str | None = None,
     entity_type: list[str] | None = None,
-    had_primary_source: str | None = None,
+    had_primary_source: list[str] | None = None,
     skip: int = 0,
     limit: int = 100,
     validate_cardinality: Literal[True, False] = True,

diff --git a/mex/backend/merged/main.py b/mex/backend/merged/main.py
@@ -16,7 +16,7 @@ def search_merged_items(  # noqa: PLR0913
     q: Annotated[str, Query(max_length=100)] = "",
     identifier: Identifier | None = None,
     entityType: Annotated[Sequence[MergedType], Query(max_length=len(MergedType))] = [],
-    hadPrimarySource: Identifier | None = None,
+    hadPrimarySource: Annotated[Sequence[Identifier] | None, Query()] = None,
     skip: Annotated[int, Query(ge=0, le=10e10)] = 0,
     limit: Annotated[int, Query(ge=1, le=100)] = 10,
 ) -> MergedItemSearch:
@@ -25,7 +25,7 @@ def search_merged_items(  # noqa: PLR0913
         q,
         identifier,
         [str(t.value) for t in entityType or MergedType],
-        hadPrimarySource,
+        [str(s) for s in hadPrimarySource] if hadPrimarySource else None,
         skip,
         limit,
         validate_cardinality=True,

diff --git a/mex/backend/preview/main.py b/mex/backend/preview/main.py
@@ -39,7 +39,7 @@ def preview_items(  # noqa: PLR0913
     q: Annotated[str, Query(max_length=100)] = "",
     identifier: Identifier | None = None,
     entityType: Annotated[Sequence[MergedType], Query(max_length=len(MergedType))] = [],
-    hadPrimarySource: Identifier | None = None,
+    hadPrimarySource: Annotated[Sequence[Identifier] | None, Query()] = None,
     skip: Annotated[int, Query(ge=0, le=10e10)] = 0,
     limit: Annotated[int, Query(ge=1, le=100)] = 10,
 ) -> PreviewItemSearch:
@@ -52,7 +52,7 @@ def preview_items(  # noqa: PLR0913
         q,
         identifier,
         [str(t.value) for t in entityType or MergedType],
-        hadPrimarySource,
+        [str(s) for s in hadPrimarySource] if hadPrimarySource else None,
         skip,
         limit,
         validate_cardinality=False,

diff --git a/pdm.lock b/pdm.lock
diff --git a/pyproject.toml b/pyproject.toml
@@ -1,6 +1,6 @@
 [project]
 name = "mex-backend"
-version = "0.30.1"
+version = "0.30.2"
 description = "Backend server for the RKI metadata exchange."
 authors = [{ name = "MEx Team", email = "[email protected]" }]
 readme = { file = "README.md", content-type = "text/markdown" }
@@ -13,7 +13,7 @@ dependencies = [
     "fastapi>=0.115,<1",
     "httpx>=0.27,<1",
     "jinja2>=3,<4",
-    "mex-common @ git+https://github.com/robert-koch-institut/mex-common.git@0.51.1",
+    "mex-common @ git+https://github.com/robert-koch-institut/mex-common.git@0.52.2",
     "neo4j>=5,<6",
     "pydantic>=2,<3",
     "starlette>=0.41,<1",