initial commit

odd_collector_profiler/datasource/database/repository.py

@@ -1,10 +1,14 @@
+import sys
+
+import oracledb
+
 from contextlib import contextmanager
 
 from sqlalchemy import create_engine, inspect
 from sqlalchemy.engine import Connection, Engine, Inspector
 
 from odd_collector_profiler.datasource.database.table import Table
-from odd_collector_profiler.domain.config import DatabaseConfig
+from odd_collector_profiler.domain.config import DatabaseConfig, OracleConfig
 from odd_collector_profiler.logger import logger
 
 
@@ -83,3 +87,12 @@ def get_tables(self) -> list[Table]:
         for table_name in self._get_tables_by(self.config.database):
             if not self.config.filters or table_name in self.config.filters:
                 yield Table(database=self.config.database, schema=None, name=table_name)
+
+
+class OracleRepository(RDBRepository):
+    def __init__(self, config: OracleConfig, skip_schemas: set[str] = None) -> None:
+        super().__init__(config, skip_schemas)
+        oracledb.version = "8.3.0"
+        sys.modules["cx_Oracle"] = oracledb
+        if config.thick_mode:
+            oracledb.init_oracle_client()

odd_collector_profiler/domain/config.py

@@ -104,3 +104,18 @@ class ClickHouseConfig(DatabaseConfig):
     def connection_str(self) -> str:
         conn_str = f"clickhouse://{self.username}:{self.password}@{self.host}:{self.port}/{self.database}"
         return conn_str
+
+
+class OracleConfig(DatabaseConfig):
+    type = "oracle"
+    host: str
+    port: str
+    user: str
+    database: str
+    password: SecretStr
+    thick_mode: Optional[bool] = False
+
+    def connection_str(self) -> str:
+        conn_str = f"oracle+cx_oracle://{self.user}:{self.password.get_secret_value()}@{self.host}:{self.port}" \
+                   f"/?service_name={self.database}"
+        return conn_str

odd_collector_profiler/profilers/oracle/profiler.py

@@ -0,0 +1,32 @@
+from typing import Any
+
+from oddrn_generator import OracleGenerator
+
+from odd_collector_profiler.datasource.database.profiler import RDBProfiler
+from odd_collector_profiler.datasource.database.repository import OracleRepository
+from odd_collector_profiler.domain.config import OracleConfig
+from odd_collector_profiler.domain.profiler import Profiler
+from odd_collector_profiler.profilers import DATA_PROFILER
+
+
+SYS_SCHEMAS = {"xdb", "xs$null", "anonymous", "apex_040000", "apex_public_user", "ctxsys", "flows_files", "hr", "mdsys",
+               "outln", "sys", "system"}
+
+
+class OracleProfiler(RDBProfiler):
+    config_model = OracleConfig
+    """Profiler for Oracle datasource"""
+
+
+def register_profiler(config: dict[str, Any]) -> Profiler:
+    config = OracleConfig.parse_obj(config)
+    repository = OracleRepository(config=config, skip_schemas=SYS_SCHEMAS)
+    generator = OracleGenerator(
+        host_settings=config.host, databases=config.database, schemas=config.user
+    )
+    return OracleProfiler(
+        config=config,
+        generator=generator,
+        repository=repository,
+        profiler=DATA_PROFILER,
+    )