Add 'meta' field to BuildTaskArtifacts (#1082)

Fixes: AlmaLinux/build-system#395

Author: javihernandez

alws/alembic/versions/bf866d7aed7a_add_meta_field_in_buildtaskartifacts.py

@@ -0,0 +1,28 @@
+"""Add meta field in BuildTaskArtifacts
+
+Revision ID: bf866d7aed7a
+Revises: 63bc4e0b699f
+Create Date: 2024-12-03 00:40:12.606988
+
+"""
+from alembic import op
+import sqlalchemy as sa
+from sqlalchemy.dialects import postgresql
+
+# revision identifiers, used by Alembic.
+revision = 'bf866d7aed7a'
+down_revision = '63bc4e0b699f'
+branch_labels = None
+depends_on = None
+
+
+def upgrade():
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.add_column('build_artifacts', sa.Column('meta', postgresql.JSONB(astext_type=sa.Text()), nullable=True))
+    # ### end Alembic commands ###
+
+
+def downgrade():
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.drop_column('build_artifacts', 'meta')
+    # ### end Alembic commands ###

alws/crud/build_node.py

@@ -412,7 +412,28 @@ def append_errata_package(_, errata_package, artifact, rpm_info):
         )
         for href, _, artifact in processed_packages
     ]
+
     rpms_info = get_rpm_packages_info(rpms)
+    for rpm in rpms:
+        rpm_info = rpms_info[rpm.href]
+        meta = {
+            "name": rpm_info["name"],
+            "epoch": rpm_info["epoch"],
+            "version": rpm_info["version"],
+            "release": rpm_info["release"],
+            "arch": rpm_info["arch"],
+            "sha256": rpm_info["sha256"],
+        }
+        rpm.meta = meta
+
+    # TODO: Consider whether we really need or want to add
+    # ErrataToAlbsPackages proposals here, see:
+    # https://github.com/AlmaLinux/build-system/issues/402
+    #
+    # If so, I think that maybe we should add them if the errata that the
+    # packages match with is not released, and avoid adding unnecessary data
+    # to db and its processing here.
+    #
     errata_record_ids = set()
     for build_task_artifact in rpms:
         rpm_info = rpms_info[build_task_artifact.href]

alws/models.py

@@ -663,6 +663,9 @@ class BuildTaskArtifact(Base):
     name: Mapped[str] = mapped_column(sqlalchemy.Text, nullable=False)
     type: Mapped[str] = mapped_column(sqlalchemy.Text, nullable=False)
     href: Mapped[str] = mapped_column(sqlalchemy.Text, nullable=False)
+    meta: Mapped[Optional[Dict[str, Any]]] = mapped_column(
+        JSONB, nullable=True
+    )
     build_task: Mapped["BuildTask"] = relationship(
         "BuildTask", back_populates="artifacts"
     )

alws/schemas/build_schema.py

@@ -151,6 +151,7 @@ class BuildTaskArtifact(BaseModel):
     type: str
     href: str
     cas_hash: typing.Optional[str] = None
+    meta: typing.Optional[dict] = None
 
     class Config:
         from_attributes = True

alws/utils/rpm_package.py

@@ -34,6 +34,7 @@ def get_rpm_packages_info(
             "release": pkg.release,
             "arch": pkg.arch,
             "rpm_sourcerpm": pkg.rpm_sourcerpm,
+            "sha256": pkg.sha256,
         }
         for _, pkg in pulp_packages.items()
     }

tests/test_utils/pulp_utils.py

@@ -1,3 +1,4 @@
+import hashlib
 import uuid
 
 from alws.schemas.build_node_schema import BuildDoneArtifact
@@ -51,5 +52,6 @@ def get_rpm_pkg_info(artifact: BuildDoneArtifact):
         "release": nevra.release,
         "arch": nevra.arch,
         "rpm_sourcerpm": rpm_sourcerpm,
+        "sha256": hashlib.sha256().hexdigest(),
     }
     return pkg_info