Feature/cleaned document state

pyproject.toml

@@ -1,6 +1,7 @@
 [project]
 name = "welearn-database"
-version = "1.5.3.dev1"
+version = "1.5.4.dev1"
+
 description = "All stuff related to relationnal database from the WeLearn project"
 authors = [
     {name = "Théo",email = "theo.nardin@cri-paris.org"}

welearn_database/alembic/versions/0310ff0bda4f_track_document_last_state.py

@@ -0,0 +1,38 @@
+"""track_document_last_state
+
+Revision ID: 0310ff0bda4f
+Revises: d22d6324c256
+Create Date: 2026-09-23 11:48:30.473174
+
+"""
+
+from typing import Sequence, Union
+
+import sqlalchemy as sa
+from alembic import op
+
+# revision identifiers, used by Alembic.
+revision: str = "0310ff0bda4f"
+down_revision: Union[str, None] = "d22d6324c256"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    op.execute("""
+    CREATE OR REPLACE VIEW document_related.track_document_latest_state
+    AS SELECT DISTINCT ON (ps.document_id) ps.id,
+        ps.document_id,
+        wd.corpus_id,
+        wd.lang,
+        ps.title,
+        ps.created_at,
+        ps.operation_order
+       FROM document_related.process_state ps
+         JOIN document_related.welearn_document wd ON ps.document_id = wd.id
+      ORDER BY ps.document_id, ps.operation_order DESC;
+    """)
+
+
+def downgrade() -> None:
+    op.execute("DROP VIEW IF EXISTS document_related.track_document_latest_state;")

welearn_database/alembic/versions/d22d6324c256_cleaned_document_state.py

@@ -0,0 +1,30 @@
+"""cleaned_document_state
+
+Revision ID: d22d6324c256
+Revises: 5d46d7920342
+Create Date: 2026-09-23 11:18:46.784209
+
+"""
+
+from typing import Sequence, Union
+
+import sqlalchemy as sa
+from alembic import op
+
+from welearn_database.data.enumeration import Step
+
+# revision identifiers, used by Alembic.
+revision: str = "d22d6324c256"
+down_revision: Union[str, None] = "5d46d7920342"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    op.execute(
+        f"""ALTER TYPE document_related.step ADD VALUE '{Step.DOCUMENT_CLEANED.value}'"""
+    )
+
+
+def downgrade() -> None:
+    pass

welearn_database/data/enumeration.py

@@ -12,6 +12,7 @@ class Step(Enum):
     DOCUMENT_IS_INVALID = "document_is_invalid"
     KEPT_FOR_TRACE = "kept_for_trace"
     DOCUMENT_IS_IRRETRIEVABLE = "document_is_irretrievable"
+    DOCUMENT_CLEANED = "document_cleaned"
 
 
 class Counter(Enum):

welearn_database/data/models/document_related.py

@@ -563,3 +563,21 @@ class QtyDocumentPerCorpus(Base):
 
     source_name: Mapped[str] = mapped_column(primary_key=True)
     count: Mapped[int]
+
+
+class TrackDocumentLatestState(Base):
+    __tablename__ = "track_document_latest_state"
+    __table_args__ = {"schema": schema_name}
+    __read_only__ = True
+
+    id: Mapped[UUID] = mapped_column(
+        primary_key=True,
+        type_=types.Uuid,
+    )
+    document_id: Mapped[UUID] = mapped_column(
+        type_=types.Uuid,
+    )
+    lang: Mapped[str]
+    title: Mapped[str]
+    created_at: Mapped[datetime] = mapped_column()
+    operation_order: Mapped[int]