chore: swap mypy with pyrefly for type checking (#817)

Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Co-authored-by: Copilot Autofix powered by AI <62310815+github-advanced-security[bot]@users.noreply.github.com>

Author: 3 people

.github/workflows/python_lint.yml

@@ -9,6 +9,8 @@ on:
 env:
   AIRBYTE_ANALYTICS_ID: ${{ vars.AIRBYTE_ANALYTICS_ID }}
 
+permissions:
+  contents: read
 jobs:
   ruff-lint-check:
     name: Ruff Lint Check
@@ -58,8 +60,8 @@ jobs:
     - name: Check code format
       run: poetry run ruff format --diff .
 
-  mypy-check:
-    name: MyPy Check
+  python-type-checks:
+    name: Python Type Checks
     runs-on: ubuntu-latest
     steps:
     # Common steps:
@@ -81,5 +83,5 @@ jobs:
       run: poetry install
 
     # Job-specifc step(s):
-    - name: Check MyPy typing
-      run: poetry run mypy .
+    - name: Run Pyrefly Check
+      run: poetry run pyrefly check

.gitignore

@@ -118,7 +118,7 @@ venv.bak/
 # mkdocs documentation
 /site
 
-# mypy
+# Type checkers (mypy, pyrefly)
 .mypy_cache/
 .dmypy.json
 dmypy.json

.ruff.toml

@@ -80,6 +80,7 @@ ignore = [
     "RUF022",  # Allow unsorted __all__ (sometimes useful for grouping by type with pdoc)
     "S",       # flake8-bandit (noisy, security related)
     "SIM910",  # Allow "None" as second argument to Dict.get(). "Explicit is better than implicit."
+    "TC006",   # Require quoted type names in cast() calls
     "TD002",   # Require author for TODOs
     "ASYNC1",  # flake8-trio (opinionated, noisy)
     "INP001",  # Dir 'examples' is part of an implicit namespace package. Add an __init__.py.

airbyte/_connector_base.py

@@ -283,10 +283,10 @@ def print_config_spec(
             content = json.dumps(self.config_spec, indent=2)
 
         if output_file:
-            output_file.write_text(content)
+            output_file.write_text(content)  # pyrefly: ignore[unbound-name]
             return
 
-        syntax_highlighted = Syntax(content, format)
+        syntax_highlighted = Syntax(content, format)  # pyrefly: ignore[unbound-name]
         rich.print(syntax_highlighted, file=sys.stderr if stderr else None)
 
     @property
@@ -300,7 +300,7 @@ def _yaml_spec(self) -> str:
         for each connector.
         """
         spec_obj: ConnectorSpecification = self._get_spec()
-        spec_dict = spec_obj.dict(exclude_unset=True)
+        spec_dict = spec_obj.model_dump(exclude_unset=True)
         # convert to a yaml string
         return yaml.dump(spec_dict)
 
@@ -405,25 +405,33 @@ def _peek_airbyte_message(
             AirbyteConnectorFailedError: If a TRACE message of type ERROR is emitted.
         """
         if message.type == Type.LOG:
-            self._print_info_message(message.log.message)
+            self._print_info_message(message.log.message)  # pyrefly: ignore[missing-attribute]
             return
 
-        if message.type == Type.TRACE and message.trace.type == TraceType.ERROR:
-            self._print_error_message(message.trace.error.message)
+        if (
+            message.type == Type.TRACE
+            and message.trace.type == TraceType.ERROR  # pyrefly: ignore[missing-attribute]
+        ):
+            self._print_error_message(
+                message.trace.error.message  # pyrefly: ignore[missing-attribute]
+            )
             if raise_on_error:
                 raise exc.AirbyteConnectorFailedError(
                     connector_name=self.name,
-                    message=message.trace.error.message,
+                    message=message.trace.error.message,  # pyrefly: ignore[missing-attribute]
                     log_text=self._last_log_messages,
                 )
             return
 
         if (
             message.type == Type.CONTROL
-            and message.control.type == OrchestratorType.CONNECTOR_CONFIG
+            and message.control.type  # pyrefly: ignore[missing-attribute]
+            == OrchestratorType.CONNECTOR_CONFIG
             and self.config_change_callback is not None
         ):
-            self.config_change_callback(message.control.connectorConfig.config)
+            self.config_change_callback(
+                message.control.connectorConfig.config  # pyrefly: ignore[missing-attribute]
+            )
             return
 
     def _execute(

airbyte/_message_iterators.py

@@ -178,7 +178,9 @@ def generator() -> Generator[AirbyteMessage, None, None]:
                 if current_file_buffer is None:
                     try:
                         current_file = next(file_iterator)
-                        current_file_buffer = file_opener(current_file)
+                        current_file_buffer = file_opener(
+                            current_file  # pyrefly: ignore[bad-argument-type]
+                        )
                     except StopIteration:
                         # No more files to read; Exit the loop
                         break
@@ -192,7 +194,7 @@ def generator() -> Generator[AirbyteMessage, None, None]:
 
                 try:
                     # Let Pydantic handle the JSON decoding from the raw string
-                    yield (
+                    yield (  # pyrefly: ignore[invalid-yield]
                         AirbyteMessage.model_validate_json(next_line),
                         current_file,
                     )

airbyte/_processors/sql/bigquery.py

@@ -7,8 +7,10 @@
 from pathlib import Path
 from typing import TYPE_CHECKING, cast, final
 
+import google.auth
 import google.oauth2
 import sqlalchemy
+import sqlalchemy.exc
 from google.api_core.exceptions import NotFound
 from google.cloud import bigquery
 from google.oauth2 import service_account

airbyte/_processors/sql/postgres.py

@@ -71,5 +71,5 @@ class PostgresSqlProcessor(SqlProcessorBase):
     file_writer_class = JsonlWriter
     sql_config: PostgresConfig
 
-    normalizer = PostgresNormalizer
+    normalizer = PostgresNormalizer  # pyrefly: ignore[bad-override]
     """A Postgres-specific name normalizer for table and column name normalization."""

airbyte/_processors/sql/snowflake.py

@@ -204,7 +204,9 @@ class SnowflakeSqlProcessor(SqlProcessorBase):
     """A Snowflake implementation of the cache."""
 
     file_writer_class = JsonlWriter
-    type_converter_class: type[SnowflakeTypeConverter] = SnowflakeTypeConverter
+    type_converter_class: type[SnowflakeTypeConverter] = (  # pyrefly: ignore[bad-override]
+        SnowflakeTypeConverter
+    )
     supports_merge_insert = True
     sql_config: SnowflakeConfig
 

airbyte/_util/api_util.py

@@ -606,7 +606,7 @@ def create_destination(
     )
     definition_id_override: str | None = None
     if _get_destination_type_str(config) == "dev-null":
-        # TODO: We have to hard-code the definition ID for dev-null destination. (important-comment)
+        # TODO: We have to hard-code the definition ID for dev-null destination.
         #  https://github.com/airbytehq/PyAirbyte/issues/743
         definition_id_override = "a7bcc9d8-13b3-4e49-b80d-d020b90045e3"
     response: api.CreateDestinationResponse = airbyte_instance.destinations.create_destination(
@@ -660,9 +660,9 @@ def get_destination(
         }
 
         if destination_type in destination_mapping:
-            response.destination_response.configuration = destination_mapping[destination_type](
-                **raw_configuration
-            )
+            response.destination_response.configuration = destination_mapping[
+                destination_type  # pyrefly: ignore[index-error]
+            ](**raw_configuration)
         return response.destination_response
 
     raise AirbyteMissingResourceError(

airbyte/_util/meta.py

@@ -134,7 +134,7 @@ def get_notebook_name() -> str | None:
 @lru_cache
 def get_vscode_notebook_name() -> str | None:
     with suppress(Exception):
-        import IPython  # noqa: PLC0415
+        import IPython  # noqa: PLC0415  # pyrefly: ignore[missing-import]
 
         return Path(
             IPython.extract_module_locals()[1]["__vsc_ipynb_file__"],