From 8bbf2c07023b7370e1edfffe52b31d4a7ab21901 Mon Sep 17 00:00:00 2001
From: Hongbo Wu <hongbo@omnivore.app>
Date: Thu, 1 Aug 2024 11:56:36 +0800
Subject: [PATCH] add script to update original_content to NULL in db

---
 packages/db/remove_original_content.py | 49 ++++++++++++++++++++++++++
 1 file changed, 49 insertions(+)
 create mode 100755 packages/db/remove_original_content.py

diff --git a/packages/db/remove_original_content.py b/packages/db/remove_original_content.py
new file mode 100755
index 000000000..29285e134
--- /dev/null
+++ b/packages/db/remove_original_content.py
@@ -0,0 +1,49 @@
+#!/usr/bin/python3
+import os
+
+import psycopg2
+
+PG_HOST = os.getenv('PG_HOST', 'localhost')
+PG_PORT = os.getenv('PG_PORT', 5432)
+PG_USER = os.getenv('PG_USER', 'app_user')
+PG_PASSWORD = os.getenv('PG_PASSWORD', 'app_pass')
+PG_DB = os.getenv('PG_DB', 'omnivore')
+PG_TIMEOUT = os.getenv('PG_TIMEOUT', 10)
+
+
+def batch_update_library_items(conn):
+    batch_size = 100
+    # update original_content to NULL in batches
+    with conn.cursor() as cursor:
+        while True:
+            cursor.execute(f"""
+                UPDATE omnivore.library_item
+                SET original_content = NULL
+                WHERE ctid IN (
+                    SELECT ctid
+                    FROM omnivore.library_item
+                    WHERE original_content IS NOT NULL
+                    LIMIT {batch_size}
+                )
+            """)
+            rows_updated = cursor.rowcount
+            conn.commit()
+            if rows_updated == 0:
+                break
+
+
+# postgres connection
+conn = psycopg2.connect(
+    f'host={PG_HOST} port={PG_PORT} dbname={PG_DB} user={PG_USER} \
+    password={PG_PASSWORD} connect_timeout={PG_TIMEOUT}')
+print('Postgres connection:', conn.info)
+
+try:
+    print('Starting migration')
+    batch_update_library_items(conn)
+    print('Migration complete')
+except Exception as err:
+    print('Migration error', err)
+finally:
+    print('Closing connections')
+    conn.close()