Do not add the schema argument to the PostgreSQL connection string when not needed. (#248)

Author: droserasprout

CHANGELOG.md

@@ -2,6 +2,12 @@
 
 Please use [this](https://docs.gitlab.com/ee/development/changelog.html) document as guidelines to keep a changelog.
 
+## [unreleased]
+
+### Fixed
+
+* database: Do not add the `schema` argument to the PostgreSQL connection string when not needed.
+
 ## 4.2.4 - 2022-02-14
 
 ### Fixed

src/dipdup/config.py

@@ -54,11 +54,15 @@
 from dipdup.utils import pascal_to_snake
 from dipdup.utils import snake_to_pascal
 
-ENV_VARIABLE_REGEX = r'\${([\w]*):-(.*)}'
+ENV_VARIABLE_REGEX = r'\${([\w]*):-(.*)}'  # ${VARIABLE:-default}
 DEFAULT_RETRY_COUNT = 3
 DEFAULT_RETRY_SLEEP = 1
 DEFAULT_METADATA_URL = 'https://metadata.dipdup.net'
 DEFAULT_IPFS_URL = 'https://ipfs.io/ipfs'
+DEFAULT_POSTGRES_SCHEMA = 'public'
+DEFAULT_POSTGRES_USER = DEFAULT_POSTGRES_DATABASE = 'postgres'
+DEFAULT_POSTGRES_PORT = 5432
+DEFAULT_SQLITE_PATH = ':memory'
 
 _logger = logging.getLogger('dipdup.config')
 
@@ -73,7 +77,7 @@ class SqliteDatabaseConfig:
     """
 
     kind: Literal['sqlite']
-    path: str = ':memory:'
+    path: str = DEFAULT_SQLITE_PATH
 
     @cached_property
     def connection_string(self) -> str:
@@ -97,10 +101,10 @@ class PostgresDatabaseConfig:
 
     kind: Literal['postgres']
     host: str
-    user: str = 'postgres'
-    database: str = 'postgres'
-    port: int = 5432
-    schema_name: str = 'public'
+    user: str = DEFAULT_POSTGRES_USER
+    database: str = DEFAULT_POSTGRES_DATABASE
+    port: int = DEFAULT_POSTGRES_PORT
+    schema_name: str = DEFAULT_POSTGRES_SCHEMA
     password: str = ''
     immune_tables: Tuple[str, ...] = field(default_factory=tuple)
     connection_timeout: int = 60
@@ -109,7 +113,10 @@ class PostgresDatabaseConfig:
     def connection_string(self) -> str:
         # NOTE: `maxsize=1` is important! Concurrency will be broken otherwise.
         # NOTE: https://github.com/tortoise/tortoise-orm/issues/792
-        return f'{self.kind}://{self.user}:{self.password}@{self.host}:{self.port}/{self.database}?schema={self.schema_name}&maxsize=1'
+        connection_string = f'{self.kind}://{self.user}:{self.password}@{self.host}:{self.port}/{self.database}?maxsize=1'
+        if self.schema_name != DEFAULT_POSTGRES_SCHEMA:
+            connection_string += f'&schema={self.schema_name}'
+        return connection_string
 
     @validator('immune_tables')
     def valid_immune_tables(cls, v):
@@ -1009,7 +1016,10 @@ def __post_init_post_parse__(self):
 
     @cached_property
     def schema_name(self) -> str:
-        return self.database.schema_name if isinstance(self.database, PostgresDatabaseConfig) else 'public'
+        if isinstance(self.database, PostgresDatabaseConfig):
+            return self.database.schema_name
+        # NOTE: Not exactly correct; historical reason
+        return DEFAULT_POSTGRES_SCHEMA
 
     @cached_property
     def package_path(self) -> str:

src/dipdup/dipdup.py

@@ -71,7 +71,6 @@
 from dipdup.utils.database import generate_schema
 from dipdup.utils.database import get_schema_hash
 from dipdup.utils.database import prepare_models
-from dipdup.utils.database import set_schema
 from dipdup.utils.database import tortoise_wrapper
 from dipdup.utils.database import validate_models
 
@@ -417,9 +416,6 @@ async def _initialize_schema(self) -> None:
         schema_name = self._config.schema_name
         conn = get_connection(None)
 
-        if isinstance(self._config.database, PostgresDatabaseConfig):
-            await set_schema(conn, schema_name)
-
         # NOTE: Try to fetch existing schema
         try:
             self._schema = await Schema.get_or_none(name=schema_name)

src/dipdup/hasura.py

@@ -25,6 +25,7 @@
 from tortoise import fields
 from tortoise.transactions import get_connection
 
+from dipdup.config import DEFAULT_POSTGRES_SCHEMA
 from dipdup.config import HasuraConfig
 from dipdup.config import HTTPConfig
 from dipdup.config import PostgresDatabaseConfig
@@ -102,7 +103,8 @@ def __init__(
     async def configure(self, force: bool = False) -> None:
         """Generate Hasura metadata and apply to instance with credentials from `hasura` config section."""
 
-        if self._database_config.schema_name != 'public':
+        # TODO: Validate during config parsing
+        if self._database_config.schema_name != DEFAULT_POSTGRES_SCHEMA:
             raise ConfigurationError('Hasura integration requires `schema_name` to be `public`')
 
         self._logger.info('Configuring Hasura')

src/dipdup/utils/database.py

@@ -132,14 +132,6 @@ def get_schema_hash(conn: BaseDBAsyncClient) -> str:
     return hashlib.sha256(processed_schema_sql).hexdigest()
 
 
-async def set_schema(conn: BaseDBAsyncClient, name: str) -> None:
-    """Set schema for the connection"""
-    if isinstance(conn, SqliteClient):
-        raise NotImplementedError
-
-    await conn.execute_script(f'SET search_path TO {name}')
-
-
 async def create_schema(conn: BaseDBAsyncClient, name: str) -> None:
     if isinstance(conn, SqliteClient):
         raise NotImplementedError
@@ -164,7 +156,6 @@ async def generate_schema(conn: BaseDBAsyncClient, name: str) -> None:
         await Tortoise.generate_schemas()
     elif isinstance(conn, AsyncpgDBClient):
         await create_schema(conn, name)
-        await set_schema(conn, name)
         await Tortoise.generate_schemas()
 
         # NOTE: Apply built-in scripts before project ones