starrocks-br 0.5.2__py3-none-any.whl → 0.7.0a1__py3-none-any.whl

starrocks_br/schema.py

@@ -15,71 +15,117 @@
 from . import logger
 
 
-def initialize_ops_schema(db) -> None:
+def initialize_ops_schema(
+    db, ops_database: str = "ops", table_inventory_entries: list[tuple[str, str, str]] | None = None
+) -> None:
     """Initialize the ops database and all required control tables.
 
-    Creates empty ops tables. Does NOT populate with sample data.
-    Users must manually insert their table inventory records.
+    Creates empty ops tables. Optionally populates table_inventory from entries.
+
+    Args:
+        db: Database connection
+        ops_database: Name of the ops database (defaults to "ops")
+        table_inventory_entries: Optional list of (group, database, table) tuples to bootstrap
     """
 
-    logger.info("Creating ops database...")
-    db.execute("CREATE DATABASE IF NOT EXISTS ops")
-    logger.success("ops database created")
+    logger.info(f"Creating {ops_database} database...")
+    db.execute(f"CREATE DATABASE IF NOT EXISTS {ops_database}")
+    logger.success(f"{ops_database} database created")
+
+    logger.info(f"Creating {ops_database}.table_inventory...")
+    db.execute(get_table_inventory_schema(ops_database=ops_database))
+    logger.success(f"{ops_database}.table_inventory created")
+
+    logger.info(f"Creating {ops_database}.backup_history...")
+    db.execute(get_backup_history_schema(ops_database=ops_database))
+    logger.success(f"{ops_database}.backup_history created")
 
-    logger.info("Creating ops.table_inventory...")
-    db.execute(get_table_inventory_schema())
-    logger.success("ops.table_inventory created")
+    logger.info(f"Creating {ops_database}.restore_history...")
+    db.execute(get_restore_history_schema(ops_database=ops_database))
+    logger.success(f"{ops_database}.restore_history created")
 
-    logger.info("Creating ops.backup_history...")
-    db.execute(get_backup_history_schema())
-    logger.success("ops.backup_history created")
+    logger.info(f"Creating {ops_database}.run_status...")
+    db.execute(get_run_status_schema(ops_database=ops_database))
+    logger.success(f"{ops_database}.run_status created")
 
-    logger.info("Creating ops.restore_history...")
-    db.execute(get_restore_history_schema())
-    logger.success("ops.restore_history created")
+    logger.info(f"Creating {ops_database}.backup_partitions...")
+    db.execute(get_backup_partitions_schema(ops_database=ops_database))
+    logger.success(f"{ops_database}.backup_partitions created")
 
-    logger.info("Creating ops.run_status...")
-    db.execute(get_run_status_schema())
-    logger.success("ops.run_status created")
+    if table_inventory_entries:
+        logger.info(f"Bootstrapping {ops_database}.table_inventory from configuration...")
+        bootstrap_table_inventory(db, table_inventory_entries, ops_database=ops_database)
+        logger.success(
+            f"{ops_database}.table_inventory bootstrapped with {len(table_inventory_entries)} entries"
+        )
 
-    logger.info("Creating ops.backup_partitions...")
-    db.execute(get_backup_partitions_schema())
-    logger.success("ops.backup_partitions created")
     logger.info("")
     logger.success("Schema initialized successfully!")
 
 
-def ensure_ops_schema(db) -> bool:
+def ensure_ops_schema(db, ops_database: str = "ops") -> bool:
     """Ensure ops schema exists, creating it if necessary.
 
     Returns True if schema was created, False if it already existed.
     This is called automatically before backup/restore operations.
     """
     try:
-        result = db.query("SHOW DATABASES LIKE 'ops'")
+        result = db.query(f"SHOW DATABASES LIKE '{ops_database}'")
 
         if not result:
-            initialize_ops_schema(db)
+            initialize_ops_schema(db, ops_database=ops_database)
             return True
 
-        db.execute("USE ops")
+        db.execute(f"USE {ops_database}")
         tables_result = db.query("SHOW TABLES")
 
         if not tables_result or len(tables_result) < 5:
-            initialize_ops_schema(db)
+            initialize_ops_schema(db, ops_database=ops_database)
             return True
 
         return False
 
     except Exception:
-        initialize_ops_schema(db)
+        initialize_ops_schema(db, ops_database=ops_database)
         return True
 
 
-def get_table_inventory_schema() -> str:
+def bootstrap_table_inventory(
+    db, entries: list[tuple[str, str, str]], ops_database: str = "ops"
+) -> None:
+    """Bootstrap table_inventory table with entries from configuration.
+
+    Args:
+        db: Database connection
+        entries: List of (group, database, table) tuples
+        ops_database: Name of the ops database (defaults to "ops")
+    """
+    if not entries:
+        return
+
+    unique_databases = {database for _, database, _ in entries}
+
+    for database_name in unique_databases:
+        result = db.query(f"SHOW DATABASES LIKE '{database_name}'")
+        if not result:
+            logger.warning(
+                f"Database '{database_name}' does not exist. "
+                f"Table inventory entries will be created, but backups will fail until the database is created."
+            )
+
+    for group, database, table in entries:
+        sql = f"""
+            INSERT INTO {ops_database}.table_inventory
+            (inventory_group, database_name, table_name)
+            VALUES ('{group}', '{database}', '{table}')
+        """
+        db.execute(sql)
+
+
+def get_table_inventory_schema(ops_database: str = "ops") -> str:
     """Get CREATE TABLE statement for table_inventory."""
-    return """
-    CREATE TABLE IF NOT EXISTS ops.table_inventory (
+    return f"""
+    CREATE TABLE IF NOT EXISTS {ops_database}.table_inventory (
         inventory_group STRING NOT NULL COMMENT "Group name for a set of tables",
         database_name STRING NOT NULL COMMENT "Database name",
         table_name STRING NOT NULL COMMENT "Table name, or '*' for all tables in database",
@@ -92,10 +138,10 @@ def get_table_inventory_schema() -> str:
     """
 
 
-def get_backup_history_schema() -> str:
+def get_backup_history_schema(ops_database: str = "ops") -> str:
     """Get CREATE TABLE statement for backup_history."""
-    return """
-    CREATE TABLE IF NOT EXISTS ops.backup_history (
+    return f"""
+    CREATE TABLE IF NOT EXISTS {ops_database}.backup_history (
         label STRING NOT NULL COMMENT "Unique backup snapshot label",
         backup_type STRING NOT NULL COMMENT "Type of backup: full or incremental",
         status STRING NOT NULL COMMENT "Final backup status: FINISHED, FAILED, CANCELLED, TIMEOUT",
@@ -110,10 +156,10 @@ def get_backup_history_schema() -> str:
     """
 
 
-def get_restore_history_schema() -> str:
+def get_restore_history_schema(ops_database: str = "ops") -> str:
     """Get CREATE TABLE statement for restore_history."""
-    return """
-    CREATE TABLE IF NOT EXISTS ops.restore_history (
+    return f"""
+    CREATE TABLE IF NOT EXISTS {ops_database}.restore_history (
         job_id STRING NOT NULL COMMENT "Unique restore job identifier",
         backup_label STRING NOT NULL COMMENT "Source backup snapshot label",
         restore_type STRING NOT NULL COMMENT "Type of restore: partition, table, or database",
@@ -130,10 +176,10 @@ def get_restore_history_schema() -> str:
     """
 
 
-def get_run_status_schema() -> str:
+def get_run_status_schema(ops_database: str = "ops") -> str:
     """Get CREATE TABLE statement for run_status."""
-    return """
-    CREATE TABLE IF NOT EXISTS ops.run_status (
+    return f"""
+    CREATE TABLE IF NOT EXISTS {ops_database}.run_status (
         scope STRING NOT NULL COMMENT "Job scope: backup or restore",
         label STRING NOT NULL COMMENT "Job label or identifier",
         state STRING NOT NULL DEFAULT "ACTIVE" COMMENT "Job state: ACTIVE, FINISHED, FAILED, or CANCELLED",
@@ -145,12 +191,12 @@ def get_run_status_schema() -> str:
     """
 
 
-def get_backup_partitions_schema() -> str:
+def get_backup_partitions_schema(ops_database: str = "ops") -> str:
     """Get CREATE TABLE statement for backup_partitions."""
-    return """
-    CREATE TABLE IF NOT EXISTS ops.backup_partitions (
+    return f"""
+    CREATE TABLE IF NOT EXISTS {ops_database}.backup_partitions (
         key_hash STRING NOT NULL COMMENT "MD5 hash of composite key (label, database_name, table_name, partition_name)",
-        label STRING NOT NULL COMMENT "The backup label this partition belongs to. FK to ops.backup_history.label.",
+        label STRING NOT NULL COMMENT "The backup label this partition belongs to. FK to {ops_database}.backup_history.label.",
         database_name STRING NOT NULL COMMENT "The name of the database the partition belongs to.",
         table_name STRING NOT NULL COMMENT "The name of the table the partition belongs to.",
         partition_name STRING NOT NULL COMMENT "The name of the specific partition.",

{starrocks_br-0.5.2.dist-info → starrocks_br-0.7.0a1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: starrocks-br
-Version: 0.5.2
+Version: 0.7.0a1
 Summary: StarRocks Backup and Restore automation tool
 Requires-Python: >=3.10
 Description-Content-Type: text/markdown
@@ -56,6 +56,7 @@ This tool adds **incremental backup capabilities** to StarRocks by leveraging na
 - ✅ **Complete operation tracking** - Every backup and restore is logged with status, timestamps, and error details
 - ✅ **Intelligent restore** - Automatically resolves backup chains (full + incremental) for you
 - ✅ **Inventory groups** - Organize tables into groups with different backup strategies
+- ✅ **Backup lifecycle management** - Prune old backups with flexible retention policies (keep-last, older-than, specific snapshots)
 - ✅ **Job concurrency control** - Prevents conflicting operations
 - ✅ **Safe restores** - Atomic rename mechanism prevents data loss during restore
 - ✅ **Metadata management** - Dedicated `ops` database tracks all backup metadata and partition manifests
@@ -95,6 +96,15 @@ port: 9030              # MySQL protocol port
 user: "root"            # Database user with backup/restore privileges
 database: "your_database"   # Database containing tables to backup
 repository: "your_repo_name"  # Repository created via CREATE REPOSITORY in StarRocks
+
+# Optional: Define table inventory groups directly in config
+table_inventory:
+  - group: "production"
+    tables:
+      - database: "mydb"
+        table: "users"
+      - database: "mydb"
+        table: "orders"
 ```
 
 Set password:
@@ -111,7 +121,11 @@ See [Configuration Reference](docs/configuration.md) for TLS and advanced option
 starrocks-br init --config config.yaml
 ```
 
-**Define inventory groups** (in StarRocks):
+This creates the `ops` database and automatically populates table inventory from your config (if defined).
+
+**Note:** If you modify the `table_inventory` in your config file, rerun `starrocks-br init --config config.yaml` to update the database.
+
+**Alternative: Define inventory groups manually** (in StarRocks):
 ```sql
 INSERT INTO ops.table_inventory (inventory_group, database_name, table_name)
 VALUES
@@ -134,6 +148,15 @@ starrocks-br backup incremental --config config.yaml --group production
 starrocks-br restore --config config.yaml --target-label mydb_20251118_full
 ```
 
+**Prune old backups:**
+```bash
+# Keep only last 5 backups
+starrocks-br prune --config config.yaml --keep-last 5
+
+# Delete backups older than a date
+starrocks-br prune --config config.yaml --older-than "2024-01-01 00:00:00"
+```
+
 See [Commands Reference](docs/commands.md) for all options.
 
 ## How It Works

starrocks_br-0.7.0a1.dist-info/RECORD

@@ -0,0 +1,25 @@
+starrocks_br/__init__.py,sha256=SeQqVuym4h6mvf4ar0UpBs_qDWbP0XOYH65Qj7WT4HI,600
+starrocks_br/cli.py,sha256=X0qUWDsocBsZnH7HvtgEqH_ZicyUKmyBEHm3IqhTM8w,34243
+starrocks_br/concurrency.py,sha256=k06-9hPdAayXGnN_9gfJNkdZTVGnFsHGsws5RK7GauE,6586
+starrocks_br/config.py,sha256=JooRqqNm39UqbXSFqgFn3x84heOb-i0ZezNKpRTxqj0,5463
+starrocks_br/db.py,sha256=abGG12e4JN4ydrz1ZdahKvOkk9CKbkGbkLnRxjUMFA8,5518
+starrocks_br/error_handler.py,sha256=PnmjUWDzP5P_5SymT-cYTiyrSe8M0DDW4pXUNYv_HOc,15890
+starrocks_br/exceptions.py,sha256=i1XUEmV1q60oTSN8u3xb7SQb3ZCaCcKvvAua4ZXd0lM,5061
+starrocks_br/executor.py,sha256=w3jMKuchdqlNf5xDA6L-AagQ1yHgktzhJyARBu6xCp0,12238
+starrocks_br/health.py,sha256=4CEX6nFKN1Vdc2cP_b89IVFZq6eaHKGMrs-FFDC4T9Q,1594
+starrocks_br/history.py,sha256=UOQqw8-bKfImQkvq6_gIIb8i7R5BaIjKiDiku-koaas,3621
+starrocks_br/labels.py,sha256=IqoNgPd-LCbZYGrVZxa0Lj9TNa8XcNCxp1oU2yiLHHU,2397
+starrocks_br/logger.py,sha256=a44M_VDhVqnaU5sxqJ0tVoXBdY4tAmWG5P3DRNOdbWE,2004
+starrocks_br/planner.py,sha256=B0K9Qi0ILoEJq_oc6q8sPBofALLSBpo94QgbjzCl7MU,12758
+starrocks_br/prune.py,sha256=QbPYQDginF1wQDcxksvvC-en_qalg6uH4lN6ELdk5Nk,7101
+starrocks_br/repository.py,sha256=hcl5ehoHdi0wgWKYl1wQYVjSpkdgrpSUPbDS2vV6BwM,1807
+starrocks_br/restore.py,sha256=mLU42XbeSr6zcLOCrbiZA4JXpwBBHITJCD4DpZcFS9w,26654
+starrocks_br/schema.py,sha256=KCQEm6h6umEWA1lb80_Sfa8ZDBARMcWNZYcF2UAlMLs,9072
+starrocks_br/timezone.py,sha256=DLoOJ2-Z65a__PEcZS1eYauVIbRJNWOOX5lpfdHfqGc,4150
+starrocks_br/utils.py,sha256=bmEZpD_CQ1_7hy51ZkRG7KKYOglhqW3EZYsWwFPUPow,2749
+starrocks_br-0.7.0a1.dist-info/licenses/LICENSE,sha256=HV8aogcPFdRayUZanMcDj73f-B3qg2YstudemA_Qi5U,11369
+starrocks_br-0.7.0a1.dist-info/METADATA,sha256=UVEsR7lD3jWosNhEa6PN19WsweGjvE7UuO99sI7Cuow,6777
+starrocks_br-0.7.0a1.dist-info/WHEEL,sha256=wUyA8OaulRlbfwMtmQsvNngGrxQHAvkKcvRmdizlJi0,92
+starrocks_br-0.7.0a1.dist-info/entry_points.txt,sha256=AKUt01G2MAlh85s1Q9kNQDOUio14kaTnT3dmg9gjdNg,54
+starrocks_br-0.7.0a1.dist-info/top_level.txt,sha256=CU1tGVo0kjulhDr761Sndg-oTeRKsisDnWm8UG95aBE,13
+starrocks_br-0.7.0a1.dist-info/RECORD,,

{starrocks_br-0.5.2.dist-info → starrocks_br-0.7.0a1.dist-info}/WHEEL

@@ -1,5 +1,5 @@
 Wheel-Version: 1.0
-Generator: setuptools (80.9.0)
+Generator: setuptools (80.10.2)
 Root-Is-Purelib: true
 Tag: py3-none-any
 

starrocks_br-0.5.2.dist-info/RECORD

@@ -1,24 +0,0 @@
-starrocks_br/__init__.py,sha256=SeQqVuym4h6mvf4ar0UpBs_qDWbP0XOYH65Qj7WT4HI,600
-starrocks_br/cli.py,sha256=9DQzFDYwqDVfoQSNamHBtQrwBKH99agsIFWrCs096wM,23496
-starrocks_br/concurrency.py,sha256=aLHLjXy8ccmtfOXUuL-HJx_xQYIR0lpQE4TI4LI9y1k,6130
-starrocks_br/config.py,sha256=6A4BdHC7OT6RQSpK_cZ_8e8M6oHIMJwu6jKDa4ZtHRg,2822
-starrocks_br/db.py,sha256=abGG12e4JN4ydrz1ZdahKvOkk9CKbkGbkLnRxjUMFA8,5518
-starrocks_br/error_handler.py,sha256=vHNaEiLkTVBkhcrNJyjboaqikpxeTMA0ZsNsE01ysB8,13657
-starrocks_br/exceptions.py,sha256=Z9zoiCdSjKAxYKIKf0rwYjTlkoCichIz7WwoFIfnmi8,4482
-starrocks_br/executor.py,sha256=aQA6lQuUc0TGy9CqqpAi9f7JDJsdwatmI-TvETuFTPs,12016
-starrocks_br/health.py,sha256=4CEX6nFKN1Vdc2cP_b89IVFZq6eaHKGMrs-FFDC4T9Q,1594
-starrocks_br/history.py,sha256=FM8uwqjIuDbKYJSf1XKr1Sila98a6uz1CESD2MsVjtI,3536
-starrocks_br/labels.py,sha256=93AROzHxTa5eCEVZlqJi9Jjv7w-YpYTANqTu-n3ujP8,2234
-starrocks_br/logger.py,sha256=a44M_VDhVqnaU5sxqJ0tVoXBdY4tAmWG5P3DRNOdbWE,2004
-starrocks_br/planner.py,sha256=zJwOx16y3Chn_WEmPgKdJ_Zfy-8Hfnp_w212qdrfxao,11326
-starrocks_br/repository.py,sha256=jYzu2EMR7RPp86c6PkuHTA-cld6qX-Jj0yTD-lryZR0,1823
-starrocks_br/restore.py,sha256=B9m6bFONyUiUGTYnb6ac3CUpW-Ev-S820-q-LD7R5Po,20677
-starrocks_br/schema.py,sha256=KayDhOL3NQq_ShwCvGrTxbFkvX7BaBybfEAe2Dxuuu0,6771
-starrocks_br/timezone.py,sha256=DLoOJ2-Z65a__PEcZS1eYauVIbRJNWOOX5lpfdHfqGc,4150
-starrocks_br/utils.py,sha256=bmEZpD_CQ1_7hy51ZkRG7KKYOglhqW3EZYsWwFPUPow,2749
-starrocks_br-0.5.2.dist-info/licenses/LICENSE,sha256=HV8aogcPFdRayUZanMcDj73f-B3qg2YstudemA_Qi5U,11369
-starrocks_br-0.5.2.dist-info/METADATA,sha256=Np4XYVJQmW3hkIkpbfCeOvTZmFHLLYLzUqUV9ShHU6A,5931
-starrocks_br-0.5.2.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-starrocks_br-0.5.2.dist-info/entry_points.txt,sha256=AKUt01G2MAlh85s1Q9kNQDOUio14kaTnT3dmg9gjdNg,54
-starrocks_br-0.5.2.dist-info/top_level.txt,sha256=CU1tGVo0kjulhDr761Sndg-oTeRKsisDnWm8UG95aBE,13
-starrocks_br-0.5.2.dist-info/RECORD,,