From c058bfd067035f0a050e5647065049efd43694c5 Mon Sep 17 00:00:00 2001
From: msramalho <19508417+msramalho@users.noreply.github.com>
Date: Wed, 20 Dec 2023 18:38:28 +0000
Subject: [PATCH] fixes unique constraint issues for archives containing the
 same url in archive_urls

---
 src/db/models.py                              |  2 +-
 ...45b_modify_archive_url_to_have_uuid_id_.py | 27 ++++++++++++++++++
 ...vacuum_database_if_there_s_enough_space.py | 28 +++++++++++++++++++
 3 files changed, 56 insertions(+), 1 deletion(-)
 create mode 100644 src/migrations/versions/9369a264945b_modify_archive_url_to_have_uuid_id_.py
 create mode 100644 src/migrations/versions/93a611e4c066_vacuum_database_if_there_s_enough_space.py

diff --git a/src/db/models.py b/src/db/models.py
index 9cddc7f..604bbcc 100644
--- a/src/db/models.py
+++ b/src/db/models.py
@@ -45,8 +45,8 @@ class ArchiveUrl(Base):
     __tablename__ = "archive_urls"
 
     url = Column(String, primary_key=True, index=True)
+    archive_id = Column(String, ForeignKey("archives.id"), primary_key=True)
     key = Column(String, default=None)
-    archive_id = Column(String, ForeignKey("archives.id"))
 
     archive = relationship("Archive", back_populates="urls")
 
diff --git a/src/migrations/versions/9369a264945b_modify_archive_url_to_have_uuid_id_.py b/src/migrations/versions/9369a264945b_modify_archive_url_to_have_uuid_id_.py
new file mode 100644
index 0000000..a2b708a
--- /dev/null
+++ b/src/migrations/versions/9369a264945b_modify_archive_url_to_have_uuid_id_.py
@@ -0,0 +1,27 @@
+"""modify archive url to have uuid id instead of url unique constraint
+
+Revision ID: 9369a264945b
+Revises: 
+Create Date: 2023-12-20 17:24:59.320691
+
+"""
+from alembic import op
+
+# revision identifiers, used by Alembic.
+revision = '9369a264945b'
+down_revision = None
+branch_labels = None
+depends_on = None
+
+
+def upgrade() -> None:
+    # since the primary key constraint is not named, we have to recreate it first
+    with op.batch_alter_table("archive_urls") as batch_op:
+        batch_op.create_primary_key("pk_url", ["url"])
+        batch_op.drop_constraint("pk_url", type_='primary')
+        batch_op.create_primary_key("pk_url_archive_id", ["url", "archive_id"])
+
+def downgrade() -> None:
+    with op.batch_alter_table("archive_urls") as batch_op:
+        batch_op.drop_constraint("pk_url_archive_id", type_='primary')
+        batch_op.create_primary_key("url", ["url"])
diff --git a/src/migrations/versions/93a611e4c066_vacuum_database_if_there_s_enough_space.py b/src/migrations/versions/93a611e4c066_vacuum_database_if_there_s_enough_space.py
new file mode 100644
index 0000000..6b8d8d2
--- /dev/null
+++ b/src/migrations/versions/93a611e4c066_vacuum_database_if_there_s_enough_space.py
@@ -0,0 +1,28 @@
+"""vacuum database (if there's enough space)
+
+Revision ID: 93a611e4c066
+Revises: 9369a264945b
+Create Date: 2023-12-20 18:33:27.132566
+
+"""
+from alembic import op
+
+
+# revision identifiers, used by Alembic.
+revision = '93a611e4c066'
+down_revision = '9369a264945b'
+branch_labels = None
+depends_on = None
+
+
+def upgrade() -> None:
+    try:
+        with op.get_context().autocommit_block():
+            op.execute("VACUUM")
+    except Exception as e:
+        print("Unable to run vacuum, maybe there's not enough disk space. it should be 2x the size of the database")
+        print(e)
+
+
+def downgrade() -> None:
+    pass