feat: add archive label and active status to ReviewModel

CHANGELOG.md

@@ -4,9 +4,11 @@
 
 ## [Unreleased]
 
-* Fix: Parse archive and JOSS links to handle markdown links and validate DOI links are valid. Added python-doi as a dependency to ensure archive/DOI URLs fully resolve (@banesullivan)
+* Add: new `active` status under `ReviewModel` which is set to `False` if the `"archived"` label is present on a review to mark the package as inactive (@banesullivan)
+
+[v1.4] - 2024-11-22
 
-## [v1.4] - 2024-11-22
+* Fix: Parse archive and JOSS links to handle markdown links and validate DOI links are valid. Added python-doi as a dependency to ensure archive/DOI URLs fully resolve (@banesullivan)
 
 Notes: it looks like i may have mistakenly bumped to 1.3.7 in august. rather than try to fix on pypi we will just go with it to ensure our release cycles are smooth given no one else uses this package except pyopensci.
 

src/pyosmeta/models/base.py

@@ -266,6 +266,7 @@ class ReviewModel(BaseModel):
     partners: Optional[list[Partnerships]] = None
     gh_meta: Optional[GhMeta] = None
     labels: list[str] = Field(default_factory=list)
+    active: bool = True  # To indicate if package is maintained or archived
 
     @field_validator(
         "date_accepted",

src/pyosmeta/models/github.py

@@ -17,9 +17,10 @@
 from __future__ import annotations
 
 from datetime import datetime
+from enum import Enum
 from typing import Any, List, Literal, Optional, Union
 
-from pydantic import AnyUrl, BaseModel, ConfigDict, Field
+from pydantic import AnyUrl, BaseModel, ConfigDict, Field, model_validator
 
 
 class User(BaseModel):
@@ -61,14 +62,45 @@ class ClosedBy(User): ...
 class Owner(User): ...
 
 
+class LabelType(str, Enum):
+    """Enum for the different types of labels that can be assigned to an issue.
+
+    This enum is not meant to be exhaustive, but rather capture a few important
+    labels for life cycle of approved reviews.
+
+    For now, this only includes the "archived" label, which is used to mark
+    packages that are no longer maintained ("inactive"). The "archived" label
+    corresponds to setting ``active=False`` on the ReviewModel
+    """
+
+    ARCHIVED = "archived"
+
+
 class Labels(BaseModel):
+    name: str
     id: Optional[int] = None
     node_id: Optional[str] = None
     url: Optional[AnyUrl] = None
-    name: Optional[str] = None
     description: Optional[str] = None
     color: Optional[str] = None
     default: Optional[bool] = None
+    type: Optional[LabelType] = None
+
+    @model_validator(mode="before")
+    def parse_label_type(cls, data):
+        """Parse the label type from the name before validation.
+
+        This will parse the label name into an available LabelType enum value.
+        Not all labels will have a corresponding LabelType, so this will
+        gracefully fail. This was implemented for assigning the LabelType.ARCHIVED
+        value to the "archived" label so that we can easily filter out archived
+        issues.
+        """
+        try:
+            data["type"] = LabelType(data["name"])
+        except ValueError:
+            pass
+        return data
 
 
 class Issue(BaseModel):

src/pyosmeta/parse_issues.py

@@ -7,7 +7,7 @@
 from pydantic import ValidationError
 
 from pyosmeta.models import ReviewModel, ReviewUser
-from pyosmeta.models.github import Issue
+from pyosmeta.models.github import Issue, Labels, LabelType
 
 from .github_api import GitHubAPI
 from .utils_clean import clean_date_accepted_key
@@ -224,6 +224,31 @@ def _postprocess_meta(self, meta: dict, body: List[str]) -> dict:
 
         return meta
 
+    def _postprocess_labels(self, meta: dict) -> dict:
+        """
+        Process specific labels for attributes in the review model.
+
+        Presently, this method only checks if the review has the "archived"
+        (LabelType.ARCHIVED) label and sets the active attribute to False
+        if it does. We may add more label processing in the future.
+
+        The intention behind this is to assign specific ReviewModel attributes
+        based on the presence of certain labels in the review issue.
+        """
+
+        def _is_archived(label: str | Labels) -> bool:
+            """Internal helper to check if a label is the "archived" label"""
+            if isinstance(label, Labels):
+                return label.type == LabelType.ARCHIVED
+            return "archived" in label.lower()
+
+        # Check if the review has the "archived" label
+        if "labels" in meta and [
+            label for label in meta["labels"] if _is_archived(label)
+        ]:
+            meta["active"] = False
+        return meta
+
     def _parse_field(self, key: str, val: str) -> Any:
         """
         Method dispatcher for parsing specific header fields.
@@ -280,6 +305,7 @@ def parse_issue(self, issue: Issue | str) -> ReviewModel:
 
         # Finalize review model before casting
         model = self._postprocess_meta(model, body)
+        model = self._postprocess_labels(model)
 
         return ReviewModel(**model)
 

tests/integration/test_parse_issues.py

@@ -96,3 +96,28 @@ def test_parse_labels(issue_list, process_issues):
         issue.labels = labels
         review = process_issues.parse_issue(issue)
         assert review.labels == ["6/pyOS-approved", "another_label"]
+        assert review.active
+
+    # Now add an archive label
+    label_inst = Labels(
+        id=1196238794,
+        node_id="MDU6TGFiZWwxMTk2MjM4Nzk0",
+        url="https://api.github.com/repos/pyOpenSci/software-submission/labels/archived",
+        name="archived",
+        description="",
+        color="006B75",
+        default=False,
+    )
+    labels = [label_inst, "another_label"]
+    for issue in issue_list:
+        issue.labels = labels
+        review = process_issues.parse_issue(issue)
+        assert not review.active
+
+    # Handle label with missing details
+    label_inst = Labels(name="test")
+    labels = [label_inst, "another_label"]
+    for issue in issue_list:
+        issue.labels = labels
+        review = process_issues.parse_issue(issue)
+        assert review.labels == ["test", "another_label"]