starrocks-br 0.3.0__py3-none-any.whl → 0.5.0__py3-none-any.whl

starrocks_br/planner.py

@@ -1,17 +1,16 @@
-from typing import List, Dict, Optional
 import datetime
 import hashlib
 
-from starrocks_br import logger, timezone
+from starrocks_br import logger, timezone, utils
 
 
-def find_latest_full_backup(db, database: str) -> Optional[Dict[str, str]]:
+def find_latest_full_backup(db, database: str) -> dict[str, str] | None:
     """Find the latest successful full backup for a database.
-    
+
     Args:
         db: Database connection
         database: Database name to search for
-        
+
     Returns:
         Dictionary with keys: label, backup_type, finished_at, or None if no full backup found.
         The finished_at value is returned as a string in the cluster timezone format.
@@ -21,125 +20,126 @@ def find_latest_full_backup(db, database: str) -> Optional[Dict[str, str]]:
     FROM ops.backup_history
     WHERE backup_type = 'full'
     AND status = 'FINISHED'
-    AND label LIKE '{database}_%'
+    AND label LIKE {utils.quote_value(f"{database}_%")}
     ORDER BY finished_at DESC
     LIMIT 1
     """
-    
+
     rows = db.query(query)
-    
+
     if not rows:
         return None
-    
+
     row = rows[0]
     finished_at = row[2]
-    
+
     if isinstance(finished_at, datetime.datetime):
-        cluster_tz = db.timezone
-        finished_at = finished_at.strftime("%Y-%m-%d %H:%M:%S")
+        finished_at_normalized = timezone.normalize_datetime_to_tz(finished_at, db.timezone)
+        finished_at = finished_at_normalized.strftime("%Y-%m-%d %H:%M:%S")
     elif not isinstance(finished_at, str):
         finished_at = str(finished_at)
-    
-    return {
-        "label": row[0],
-        "backup_type": row[1],
-        "finished_at": finished_at
-    }
 
+    return {"label": row[0], "backup_type": row[1], "finished_at": finished_at}
 
-def find_tables_by_group(db, group_name: str) -> List[Dict[str, str]]:
+
+def find_tables_by_group(db, group_name: str) -> list[dict[str, str]]:
     """Find tables belonging to a specific inventory group.
-    
+
     Returns list of dictionaries with keys: database, table.
     Supports '*' table wildcard which signifies all tables in a database.
     """
     query = f"""
     SELECT database_name, table_name
     FROM ops.table_inventory
-    WHERE inventory_group = '{group_name}'
+    WHERE inventory_group = {utils.quote_value(group_name)}
     ORDER BY database_name, table_name
     """
     rows = db.query(query)
-    return [
-        {"database": row[0], "table": row[1]} for row in rows
-    ]
+    return [{"database": row[0], "table": row[1]} for row in rows]
 
 
-def find_recent_partitions(db, database: str, baseline_backup_label: Optional[str] = None, *, group_name: str) -> List[Dict[str, str]]:
+def find_recent_partitions(
+    db, database: str, baseline_backup_label: str | None = None, *, group_name: str
+) -> list[dict[str, str]]:
     """Find partitions updated since baseline for tables in the given inventory group.
-    
+
     Args:
         db: Database connection
         database: Database name (StarRocks database scope for backup)
         baseline_backup_label: Optional specific backup label to use as baseline.
         group_name: Inventory group whose tables will be considered
-    
+
     Returns list of dictionaries with keys: database, table, partition_name.
     Only partitions of tables within the specified database are returned.
     """
     cluster_tz = db.timezone
-    
+
     if baseline_backup_label:
         baseline_query = f"""
         SELECT finished_at
         FROM ops.backup_history
-        WHERE label = '{baseline_backup_label}'
+        WHERE label = {utils.quote_value(baseline_backup_label)}
         AND status = 'FINISHED'
         """
         baseline_rows = db.query(baseline_query)
         if not baseline_rows:
-            raise ValueError(f"Baseline backup '{baseline_backup_label}' not found or not successful")
+            raise ValueError(
+                f"Baseline backup '{baseline_backup_label}' not found or not successful"
+            )
         baseline_time_raw = baseline_rows[0][0]
     else:
         latest_backup = find_latest_full_backup(db, database)
         if not latest_backup:
-            raise ValueError(f"No successful full backup found for database '{database}'. Run a full database backup first.")
-        baseline_time_raw = latest_backup['finished_at']
-    
+            raise ValueError(
+                f"No successful full backup found for database '{database}'. Run a full database backup first."
+            )
+        baseline_time_raw = latest_backup["finished_at"]
+
     if isinstance(baseline_time_raw, datetime.datetime):
         baseline_time_str = baseline_time_raw.strftime("%Y-%m-%d %H:%M:%S")
     elif isinstance(baseline_time_raw, str):
         baseline_time_str = baseline_time_raw
     else:
         baseline_time_str = str(baseline_time_raw)
-    
+
     baseline_dt = timezone.parse_datetime_with_tz(baseline_time_str, cluster_tz)
-    
+
     group_tables = find_tables_by_group(db, group_name)
 
     if not group_tables:
         return []
 
-    db_group_tables = [t for t in group_tables if t['database'] == database]
+    db_group_tables = [t for t in group_tables if t["database"] == database]
 
     if not db_group_tables:
         return []
-    
+
     concrete_tables = []
     for table_entry in db_group_tables:
-        if table_entry['table'] == '*':
-            show_tables_query = f"SHOW TABLES FROM {table_entry['database']}"
+        if table_entry["table"] == "*":
+            show_tables_query = (
+                f"SHOW TABLES FROM {utils.quote_identifier(table_entry['database'])}"
+            )
             tables_rows = db.query(show_tables_query)
             for row in tables_rows:
-                concrete_tables.append({
-                    'database': table_entry['database'],
-                    'table': row[0]
-                })
+                concrete_tables.append({"database": table_entry["database"], "table": row[0]})
         else:
             concrete_tables.append(table_entry)
-    
+
     recent_partitions = []
     for table_entry in concrete_tables:
-        db_name = table_entry['database']
-        table_name = table_entry['table']
-        
-        show_partitions_query = f"SHOW PARTITIONS FROM {db_name}.{table_name}"
+        db_name = table_entry["database"]
+        table_name = table_entry["table"]
+
+        show_partitions_query = (
+            f"SHOW PARTITIONS FROM {utils.build_qualified_table_name(db_name, table_name)}"
+        )
         try:
             partition_rows = db.query(show_partitions_query)
         except Exception as e:
             logger.error(f"Error showing partitions for table {db_name}.{table_name}: {e}")
             continue
-        
+
         for row in partition_rows:
             # FOR SHARED NOTHING CLUSTER:
             # PartitionId, PartitionName, VisibleVersion, VisibleVersionTime, VisibleVersionHash, State, PartitionKey, Range, DistributionKey, Buckets, ReplicationNum, StorageMedium, CooldownTime, LastConsistencyCheckTime, DataSize, StorageSize, IsInMemory, RowCount, DataVersion, VersionEpoch, VersionTxnType
@@ -152,86 +152,90 @@ def find_recent_partitions(db, database: str, baseline_backup_label: Optional[st
                 visible_version_time_str = visible_version_time
             else:
                 visible_version_time_str = str(visible_version_time)
-            
-            visible_version_dt = timezone.parse_datetime_with_tz(visible_version_time_str, cluster_tz)
-            
+
+            visible_version_dt = timezone.parse_datetime_with_tz(
+                visible_version_time_str, cluster_tz
+            )
+
             if visible_version_dt > baseline_dt:
-                recent_partitions.append({
-                    'database': db_name,
-                    'table': table_name,
-                    'partition_name': partition_name
-                })
-    
+                recent_partitions.append(
+                    {"database": db_name, "table": table_name, "partition_name": partition_name}
+                )
+
     return recent_partitions
 
 
-def build_incremental_backup_command(partitions: List[Dict[str, str]], repository: str, label: str, database: str) -> str:
+def build_incremental_backup_command(
+    partitions: list[dict[str, str]], repository: str, label: str, database: str
+) -> str:
     """Build BACKUP command for incremental backup of specific partitions.
-    
+
     Args:
         partitions: List of partitions to backup
         repository: Repository name
         label: Backup label
         database: Database name (StarRocks requires BACKUP to be database-specific)
-    
+
     Note: Filters partitions to only include those from the specified database.
     """
     if not partitions:
         return ""
-    
-    db_partitions = [p for p in partitions if p['database'] == database]
-    
+
+    db_partitions = [p for p in partitions if p["database"] == database]
+
     if not db_partitions:
         return ""
-    
+
     table_partitions = {}
     for partition in db_partitions:
-        table_name = partition['table']
+        table_name = partition["table"]
         if table_name not in table_partitions:
             table_partitions[table_name] = []
-        table_partitions[table_name].append(partition['partition_name'])
-    
+        table_partitions[table_name].append(partition["partition_name"])
+
     on_clauses = []
     for table, parts in table_partitions.items():
-        partitions_str = ", ".join(parts)
-        on_clauses.append(f"TABLE {table} PARTITION ({partitions_str})")
-    
+        partitions_str = ", ".join(utils.quote_identifier(p) for p in parts)
+        on_clauses.append(f"TABLE {utils.quote_identifier(table)} PARTITION ({partitions_str})")
+
     on_clause = ",\n    ".join(on_clauses)
-    
-    command = f"""BACKUP DATABASE {database} SNAPSHOT {label}
-    TO {repository}
+
+    command = f"""BACKUP DATABASE {utils.quote_identifier(database)} SNAPSHOT {utils.quote_identifier(label)}
+    TO {utils.quote_identifier(repository)}
     ON ({on_clause})"""
-    
+
     return command
 
 
-def build_full_backup_command(db, group_name: str, repository: str, label: str, database: str) -> str:
+def build_full_backup_command(
+    db, group_name: str, repository: str, label: str, database: str
+) -> str:
     """Build BACKUP command for an inventory group.
-    
+
     If the group contains '*' for any entry in the target database, generate a
     simple BACKUP DATABASE command. Otherwise, generate ON (TABLE ...) list for
     the specific tables within the database.
     """
     tables = find_tables_by_group(db, group_name)
 
-    db_entries = [t for t in tables if t['database'] == database]
+    db_entries = [t for t in tables if t["database"] == database]
     if not db_entries:
         return ""
 
-    if any(t['table'] == '*' for t in db_entries):
-        return f"""BACKUP DATABASE {database} SNAPSHOT {label}
-    TO {repository}"""
+    if any(t["table"] == "*" for t in db_entries):
+        return f"""BACKUP DATABASE {utils.quote_identifier(database)} SNAPSHOT {utils.quote_identifier(label)}
+    TO {utils.quote_identifier(repository)}"""
 
     on_clauses = []
     for t in db_entries:
-        on_clauses.append(f"TABLE {t['table']}")
+        on_clauses.append(f"TABLE {utils.quote_identifier(t['table'])}")
     on_clause = ",\n        ".join(on_clauses)
-    return f"""BACKUP DATABASE {database} SNAPSHOT {label}
-    TO {repository}
+    return f"""BACKUP DATABASE {utils.quote_identifier(database)} SNAPSHOT {utils.quote_identifier(label)}
+    TO {utils.quote_identifier(repository)}
     ON ({on_clause})"""
 
 
-def record_backup_partitions(db, label: str, partitions: List[Dict[str, str]]) -> None:
+def record_backup_partitions(db, label: str, partitions: list[dict[str, str]]) -> None:
     """Record partition metadata for a backup in ops.backup_partitions table.
 
     Args:
@@ -243,62 +247,59 @@ def record_backup_partitions(db, label: str, partitions: List[Dict[str, str]]) -
         return
 
     for partition in partitions:
-        composite_key = f"{label}|{partition['database']}|{partition['table']}|{partition['partition_name']}"
-        key_hash = hashlib.md5(composite_key.encode('utf-8')).hexdigest()
+        composite_key = (
+            f"{label}|{partition['database']}|{partition['table']}|{partition['partition_name']}"
+        )
+        key_hash = hashlib.md5(composite_key.encode("utf-8")).hexdigest()
 
         db.execute(f"""
             INSERT INTO ops.backup_partitions
             (key_hash, label, database_name, table_name, partition_name)
-            VALUES ('{key_hash}', '{label}', '{partition['database']}', '{partition['table']}', '{partition['partition_name']}')
+            VALUES ({utils.quote_value(key_hash)}, {utils.quote_value(label)}, {utils.quote_value(partition["database"])}, {utils.quote_value(partition["table"])}, {utils.quote_value(partition["partition_name"])})
         """)
 
 
-def get_all_partitions_for_tables(db, database: str, tables: List[Dict[str, str]]) -> List[Dict[str, str]]:
+def get_all_partitions_for_tables(
+    db, database: str, tables: list[dict[str, str]]
+) -> list[dict[str, str]]:
     """Get all existing partitions for the specified tables.
-    
+
     Args:
         db: Database connection
         database: Database name
         tables: List of tables with keys: database, table
-        
+
     Returns:
         List of partitions with keys: database, table, partition_name
     """
     if not tables:
         return []
-    
-    db_tables = [t for t in tables if t['database'] == database]
+
+    db_tables = [t for t in tables if t["database"] == database]
     if not db_tables:
         return []
-    
-    where_conditions = [f"DB_NAME = '{database}'", "PARTITION_NAME IS NOT NULL"]
-    
+
+    where_conditions = [f"DB_NAME = {utils.quote_value(database)}", "PARTITION_NAME IS NOT NULL"]
+
     table_conditions = []
     for table in db_tables:
-        if table['table'] == '*':
+        if table["table"] == "*":
             pass
         else:
-            table_conditions.append(f"TABLE_NAME = '{table['table']}'")
-    
+            table_conditions.append(f"TABLE_NAME = {utils.quote_value(table['table'])}")
+
     if table_conditions:
         where_conditions.append("(" + " OR ".join(table_conditions) + ")")
-    
+
     where_clause = " AND ".join(where_conditions)
-    
+
     query = f"""
     SELECT DB_NAME, TABLE_NAME, PARTITION_NAME
     FROM information_schema.partitions_meta 
     WHERE {where_clause}
     ORDER BY TABLE_NAME, PARTITION_NAME
     """
-    
+
     rows = db.query(query)
-    
-    return [
-        {
-            "database": row[0],
-            "table": row[1], 
-            "partition_name": row[2]
-        }
-        for row in rows
-    ]
+
+    return [{"database": row[0], "table": row[1], "partition_name": row[2]} for row in rows]

starrocks_br/repository.py

@@ -3,11 +3,11 @@ from __future__ import annotations
 
 def ensure_repository(db, name: str) -> None:
     """Verify that the specified repository exists and is accessible.
-    
+
     Args:
         db: Database connection
         name: Repository name to verify
-        
+
     Raises:
         RuntimeError: If repository doesn't exist or has errors
     """
@@ -18,7 +18,7 @@ def ensure_repository(db, name: str) -> None:
             f"  CREATE REPOSITORY {name} WITH BROKER ON LOCATION '...' PROPERTIES(...)\n"
             f"For examples, see: https://docs.starrocks.io/docs/sql-reference/sql-statements/data-definition/backup_restore/CREATE_REPOSITORY/"
         )
-    
+
     # SHOW REPOSITORIES returns: RepoId, RepoName, CreateTime, IsReadOnly, Location, Broker, ErrMsg
     err_msg = existing[6]
     if err_msg and str(err_msg).strip().upper() not in {"", "NULL", "NONE"}:
@@ -32,5 +32,3 @@ def _find_repository(db, name: str):
         if row and row[1] == name:
             return row
     return None
-
-