Merge branch 'master' into allow-context-in-data

Bring in up-to-date changes from the base branch
dandi · Jan 20, 2025 · 5ca9e5e · 5ca9e5e
2 parents fd3df90 + 6fd04fb
commit 5ca9e5e
Show file tree

Hide file tree

Showing 8 changed files with 84 additions and 37 deletions.
diff --git a/.github/workflows/test-nonetwork.yml b/.github/workflows/test-nonetwork.yml
@@ -47,7 +47,7 @@ jobs:
           DANDI_TESTS_NONETWORK: 1
 
       - name: Upload coverage to Codecov
-        uses: codecov/codecov-action@v4
+        uses: codecov/codecov-action@v5
         with:
           file: ./coverage.xml
           flags: unittests

diff --git a/.github/workflows/test.yml b/.github/workflows/test.yml
@@ -46,7 +46,7 @@ jobs:
       run: tox -e py -- -s --cov-report=xml
 
     - name: Upload coverage to Codecov
-      uses: codecov/codecov-action@v4
+      uses: codecov/codecov-action@v5
       with:
         file: ./coverage.xml
         flags: unittests

diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -1,3 +1,36 @@
+# 0.11.0 (Tue Jan 07 2025)
+
+#### 🚀 Enhancement
+
+- Upgrade to datacite v4.5 serialization from inveniosoftware [#261](https://github.com/dandi/dandi-schema/pull/261) ([@yarikoptic](https://github.com/yarikoptic))
+- Add "version" to datacite record and bundle datacite json serializations [#260](https://github.com/dandi/dandi-schema/pull/260) ([@yarikoptic](https://github.com/yarikoptic))
+- Drop Python 3.8 support (remove typing_extensions from depends) [#262](https://github.com/dandi/dandi-schema/pull/262) ([@yarikoptic](https://github.com/yarikoptic))
+
+#### 🐛 Bug Fix
+
+- Use `title` in place of `examples` [#271](https://github.com/dandi/dandi-schema/pull/271) ([@waxlamp](https://github.com/waxlamp) [@satra](https://github.com/satra) [@pre-commit-ci[bot]](https://github.com/pre-commit-ci[bot]))
+- RF: detect BIDS based on having dataset_description.json [#263](https://github.com/dandi/dandi-schema/pull/263) ([@yarikoptic](https://github.com/yarikoptic))
+- Add support for detection and addition of ome/ngff "standard" into assets summary [#252](https://github.com/dandi/dandi-schema/pull/252) ([@yarikoptic](https://github.com/yarikoptic))
+
+#### 🏠 Internal
+
+- [gh-actions](deps): Bump codecov/codecov-action from 4 to 5 [#267](https://github.com/dandi/dandi-schema/pull/267) ([@dependabot[bot]](https://github.com/dependabot[bot]))
+- [pre-commit.ci] pre-commit autoupdate (black 24.8.0 → 24.10.0; pre-commit-hooks: v4.6.0 → v5.0.0) [#253](https://github.com/dandi/dandi-schema/pull/253) ([@pre-commit-ci[bot]](https://github.com/pre-commit-ci[bot]))
+
+#### 📝 Documentation
+
+- doc: A comment on current identification of URI and missing space into a description string [#265](https://github.com/dandi/dandi-schema/pull/265) ([@yarikoptic](https://github.com/yarikoptic))
+
+#### Authors: 5
+
+- [@dependabot[bot]](https://github.com/dependabot[bot])
+- [@pre-commit-ci[bot]](https://github.com/pre-commit-ci[bot])
+- Roni Choudhury ([@waxlamp](https://github.com/waxlamp))
+- Satrajit Ghosh ([@satra](https://github.com/satra))
+- Yaroslav Halchenko ([@yarikoptic](https://github.com/yarikoptic))
+
+---
+
 # 0.10.4 (Thu Sep 05 2024)
 
 #### 🐛 Bug Fix

diff --git a/README.md b/README.md
@@ -35,6 +35,6 @@ out of the Pydantic models.
 ## Resources
 
 * To learn how to interact with the DANDI archive,
-see [the handbook](https://www.dandiarchive.org/handbook/).
+see the [DANDI Docs](https://docs.dandiarchive.org).
 * To file a feature request or bug report, go to https://github.com/dandi/helpdesk/issues/new/choose.
 * For all other issues, contact the DANDI team: [email protected].
diff --git a/dandischema/consts.py b/dandischema/consts.py
@@ -1,4 +1,4 @@
-DANDI_SCHEMA_VERSION = "0.6.8"
+DANDI_SCHEMA_VERSION = "0.6.9"
 ALLOWED_INPUT_SCHEMAS = [
     "0.4.4",
     "0.5.1",
@@ -11,6 +11,7 @@
     "0.6.5",
     "0.6.6",
     "0.6.7",
+    "0.6.8",
 ]
 
 # ATM we allow only for a single target version which is current

diff --git a/dandischema/datacite/__init__.py b/dandischema/datacite/__init__.py
@@ -78,16 +78,14 @@ def to_datacite(
     if publish:
         attributes["event"] = "publish"
 
-    attributes["identifiers"] = [
-        # TODO: the first element is ignored, not sure how to fix it...
-        {"identifier": f"https://doi.org/{meta.doi}", "identifierType": "DOI"},
+    attributes["alternateIdentifiers"] = [
         {
-            "identifier": f"https://identifiers.org/{meta.id}",
-            "identifierType": "URL",
+            "alternateIdentifier": f"https://identifiers.org/{meta.id}",
+            "alternateIdentifierType": "URL",
         },
         {
-            "identifier": str(meta.url),
-            "identifierType": "URL",
+            "alternateIdentifier": str(meta.url),
+            "alternateIdentifierType": "URL",
         },
     ]
 
@@ -98,7 +96,13 @@ def to_datacite(
     attributes["descriptions"] = [
         {"description": meta.description, "descriptionType": "Abstract"}
     ]
-    attributes["publisher"] = "DANDI Archive"
+    attributes["publisher"] = {
+        "name": "DANDI Archive",
+        "schemeUri": "https://scicrunch.org/resolver/",
+        "publisherIdentifier": "https://scicrunch.org/resolver/RRID:SCR_017571",
+        "publisherIdentifierScheme": "RRID",
+        "lang": "en",
+    }
     attributes["publicationYear"] = str(meta.datePublished.year)
     # not sure about it dandi-api had "resourceTypeGeneral": "NWB"
     attributes["types"] = {
@@ -110,7 +114,7 @@ def to_datacite(
     # assuming that all licenses are from SPDX?
     attributes["rightsList"] = [
         {
-            "schemeURI": "https://spdx.org/licenses/",
+            "schemeUri": "https://spdx.org/licenses/",
             "rightsIdentifierScheme": "SPDX",
             "rightsIdentifier": el.name,
         }
@@ -147,7 +151,7 @@ def to_datacite(
         contr_dict: Dict[str, Any] = {
             "name": contr_el.name,
             "contributorName": contr_el.name,
-            "schemeURI": "orcid.org",
+            "schemeUri": "orcid.org",
         }
         if isinstance(contr_el, Person):
             contr_dict["nameType"] = "Personal"
@@ -163,7 +167,7 @@ def to_datacite(
                 contr_dict["affiliation"] = []
             if getattr(contr_el, "identifier"):
                 orcid_dict = {
-                    "nameIdentifier": contr_el.identifier,
+                    "nameIdentifier": f"https://orcid.org/{contr_el.identifier}",
                     "nameIdentifierScheme": "ORCID",
                     "schemeUri": "https://orcid.org/",
                 }
@@ -252,7 +256,7 @@ def to_datacite(
 
 
 @lru_cache()
-def _get_datacite_schema(version_id: str = "datacite-4.3-17-gaa5db56") -> Any:
+def _get_datacite_schema(version_id: str = "inveniosoftware-4.5-81-g160250d") -> Any:
     """Load datacite schema based on the version id provided."""
     schema_folder = Path(__file__).parent / "schema"
     return json.loads((schema_folder / f"{version_id}.json").read_text())

diff --git a/dandischema/datacite/tests/test_datacite.py b/dandischema/datacite/tests/test_datacite.py
@@ -150,7 +150,16 @@ def test_datacite(dandi_id: str, schema: Any) -> None:
                     1,
                     {"description": "testing", "descriptionType": "Abstract"},
                 ),
-                "publisher": (None, "DANDI Archive"),
+                "publisher": (
+                    None,
+                    {
+                        "name": "DANDI Archive",
+                        "publisherIdentifier": "https://scicrunch.org/resolver/RRID:SCR_017571",
+                        "publisherIdentifierScheme": "RRID",
+                        "schemeUri": "https://scicrunch.org/resolver/",
+                        "lang": "en",
+                    },
+                ),
                 "rightsList": (
                     1,
                     {"rightsIdentifierScheme": "SPDX", "rightsIdentifier": "CC_BY_40"},
@@ -302,7 +311,7 @@ def test_datacite(dandi_id: str, schema: Any) -> None:
                         "name": "A_last, A_first",
                         "nameIdentifiers": [
                             {
-                                "nameIdentifier": "0000-0001-0000-0000",
+                                "nameIdentifier": "https://orcid.org/0000-0001-0000-0000",
                                 "nameIdentifierScheme": "ORCID",
                                 "schemeUri": "https://orcid.org/",
                             }
@@ -316,7 +325,7 @@ def test_datacite(dandi_id: str, schema: Any) -> None:
                         "contributorType": "Other",
                         "nameIdentifiers": [
                             {
-                                "nameIdentifier": "0000-0001-0000-0000",
+                                "nameIdentifier": "https://orcid.org/0000-0001-0000-0000",
                                 "nameIdentifierScheme": "ORCID",
                                 "schemeUri": "https://orcid.org/",
                             }
@@ -423,7 +432,7 @@ def test_datacite_publish(metadata_basic: Dict[str, Any]) -> None:
                         "givenName": "A_first",
                         "name": "A_last, A_first",
                         "nameType": "Personal",
-                        "schemeURI": "orcid.org",
+                        "schemeUri": "orcid.org",
                     }
                 ],
                 "creators": [
@@ -434,40 +443,39 @@ def test_datacite_publish(metadata_basic: Dict[str, Any]) -> None:
                         "givenName": "A_first",
                         "name": "A_last, A_first",
                         "nameType": "Personal",
-                        "schemeURI": "orcid.org",
+                        "schemeUri": "orcid.org",
                     }
                 ],
                 "descriptions": [
                     {"description": "testing", "descriptionType": "Abstract"}
                 ],
                 "doi": f"10.80507/dandi.{dandi_id_noprefix}/{version}",
-                "identifiers": [
+                "alternateIdentifiers": [
                     {
-                        "identifier": (
-                            f"https://doi.org/10.80507"
-                            f"/dandi.{dandi_id_noprefix}/{version}"
-                        ),
-                        "identifierType": "DOI",
-                    },
-                    {
-                        "identifier": f"https://identifiers.org/{dandi_id}/{version}",
-                        "identifierType": "URL",
+                        "alternateIdentifier": f"https://identifiers.org/{dandi_id}/{version}",
+                        "alternateIdentifierType": "URL",
                     },
                     {
-                        "identifier": (
+                        "alternateIdentifier": (
                             f"https://dandiarchive.org/dandiset"
                             f"/{dandi_id_noprefix}/{version}"
                         ),
-                        "identifierType": "URL",
+                        "alternateIdentifierType": "URL",
                     },
                 ],
                 "publicationYear": "1970",
-                "publisher": "DANDI Archive",
+                "publisher": {
+                    "name": "DANDI Archive",
+                    "publisherIdentifier": "https://scicrunch.org/resolver/RRID:SCR_017571",
+                    "publisherIdentifierScheme": "RRID",
+                    "schemeUri": "https://scicrunch.org/resolver/",
+                    "lang": "en",
+                },
                 "rightsList": [
                     {
                         "rightsIdentifier": "CC_BY_40",
                         "rightsIdentifierScheme": "SPDX",
-                        "schemeURI": "https://spdx.org/licenses/",
+                        "schemeUri": "https://spdx.org/licenses/",
                     }
                 ],
                 "schemaVersion": "http://datacite.org/schema/kernel-4",

diff --git a/dandischema/models.py b/dandischema/models.py
@@ -613,6 +613,7 @@ def __get_pydantic_json_schema__(
             if value.get("title") is None or value["title"] == prop.title():
                 value["title"] = name2title(prop)
             if re.match("\\^https?://", value.get("pattern", "")):
+                # triggers only for ROR in identifier
                 value["format"] = "uri"
             if value.get("format", None) == "uri":
                 value["maxLength"] = 1000
@@ -671,7 +672,7 @@ class PropertyValue(DandiBaseModel):
     )  # Note: recursive (circular or not)
     propertyID: Optional[Union[IdentifierType, AnyHttpUrl]] = Field(
         None,
-        description="A commonly used identifier for"
+        description="A commonly used identifier for "
         "the characteristic represented by the property. "
         "For example, a known prefix like DOI or a full URL.",
         json_schema_extra={"nskey": "schema"},
@@ -993,10 +994,10 @@ class Person(Contributor):
         json_schema_extra={"nskey": "schema"},
     )
     name: str = Field(
+        title="Use Last, First. Example: Lovelace, Augusta Ada",
         description="Use the format: familyname, given names ...",
         pattern=NAME_PATTERN,
         json_schema_extra={"nskey": "schema"},
-        examples=["Lovelace, Augusta Ada", "Smith, John", "Chan, Kong-sang"],
     )
     affiliation: Optional[List[Affiliation]] = Field(
         None,