🐛 Source File: fix check method

airbyte-config/init/src/main/resources/seed/source_definitions.yaml

@@ -317,7 +317,7 @@
 - name: File
   sourceDefinitionId: 778daa7c-feaf-4db6-96f3-70fd645acc77
   dockerRepository: airbyte/source-file
-  dockerImageTag: 0.2.26
+  dockerImageTag: 0.2.27
   documentationUrl: https://docs.airbyte.com/integrations/sources/file
   icon: file.svg
   sourceType: file

airbyte-config/init/src/main/resources/seed/source_specs.yaml

@@ -3123,7 +3123,7 @@
     supportsNormalization: false
     supportsDBT: false
     supported_destination_sync_modes: []
-- dockerImage: "airbyte/source-file:0.2.26"
+- dockerImage: "airbyte/source-file:0.2.27"
   spec:
     documentationUrl: "https://docs.airbyte.com/integrations/sources/file"
     connectionSpecification:

airbyte-integrations/connectors/source-file/Dockerfile

@@ -17,5 +17,5 @@ COPY source_file ./source_file
 ENV AIRBYTE_ENTRYPOINT "python /airbyte/integration_code/main.py"
 ENTRYPOINT ["python", "/airbyte/integration_code/main.py"]
 
-LABEL io.airbyte.version=0.2.26
+LABEL io.airbyte.version=0.2.27
 LABEL io.airbyte.name=airbyte/source-file

airbyte-integrations/connectors/source-file/integration_tests/file_formats_test.py

@@ -9,7 +9,7 @@
 import pytest
 from airbyte_cdk import AirbyteLogger
 from source_file import SourceFile
-from source_file.client import Client
+from source_file.client import Client, ConfigurationError
 
 SAMPLE_DIRECTORY = Path(__file__).resolve().parent.joinpath("sample_files/formats")
 
@@ -42,6 +42,28 @@ def test_local_file_read(file_format, extension, expected_columns, expected_rows
     check_read(configs, expected_columns, expected_rows)
 
 
+@pytest.mark.parametrize(
+    "file_format, extension, wrong_format, filename",
+    [
+        ("excel", "xls", "csv", "demo"),
+        ("excel", "xlsx", "csv", "demo"),
+        ("csv", "csv", "excel", "demo"),
+        ("csv", "csv", "excel", "demo"),
+        ("jsonl", "jsonl", "excel", "jsonl_nested"),
+        ("feather", "feather", "csv", "demo"),
+        ("parquet", "parquet", "feather", "demo"),
+        ("yaml", "yaml", "json", "demo"),
+    ],
+)
+def test_raises_file_wrong_format(file_format, extension, wrong_format, filename):
+    file_directory = SAMPLE_DIRECTORY.joinpath(file_format)
+    file_path = str(file_directory.joinpath(f"{filename}.{extension}"))
+    configs = {"dataset_name": "test", "format": wrong_format, "url": file_path, "provider": {"storage": "local"}}
+    client = Client(**configs)
+    with pytest.raises((TypeError, ValueError, ConfigurationError)):
+        list(client.read())
+
+
 def run_load_dataframes(config, expected_columns=10, expected_rows=42):
     df_list = SourceFile.load_dataframes(config=config, logger=AirbyteLogger(), skip_data=False)
     assert len(df_list) == 1  # Properly load 1 DataFrame

airbyte-integrations/connectors/source-file/source_file/source.py

@@ -21,7 +21,7 @@
 )
 from airbyte_cdk.sources import Source
 
-from .client import Client
+from .client import Client, ConfigurationError
 from .utils import dropbox_force_download
 
 
@@ -96,12 +96,25 @@ def check(self, logger, config: Mapping) -> AirbyteConnectionStatus:
         """
         config = self._validate_and_transform(config)
         client = self._get_client(config)
-        logger.info(f"Checking access to {client.reader.full_url}...")
+        source_url = client.reader.full_url
+        logger.info(f"Checking access to {source_url}...")
+        if "docs.google.com/spreadsheets" in source_url:
+            reason = f"Failed to load {source_url}: please use the Official Google Sheets Source connector"
+            logger.error(reason)
+            return AirbyteConnectionStatus(status=Status.FAILED, message=reason)
         try:
             with client.reader.open():
+                list(client.streams)
                 return AirbyteConnectionStatus(status=Status.SUCCEEDED)
+        except (TypeError, ValueError, ConfigurationError) as err:
+            reason = (
+                f"Failed to load {source_url}\n Please check File Format and Reader Options are set correctly"
+                f"\n{repr(err)}\n{traceback.format_exc()}"
+            )
+            logger.error(reason)
+            return AirbyteConnectionStatus(status=Status.FAILED, message=reason)
         except Exception as err:
-            reason = f"Failed to load {client.reader.full_url}: {repr(err)}\n{traceback.format_exc()}"
+            reason = f"Failed to load {source_url}: {repr(err)}\n{traceback.format_exc()}"
             logger.error(reason)
             return AirbyteConnectionStatus(status=Status.FAILED, message=reason)
 

docs/integrations/sources/file.md

@@ -128,7 +128,8 @@ In order to read large files from a remote location, this connector uses the [sm
 ## Changelog
 
 | Version | Date       | Pull Request                                             | Subject                                                  |
-| ------- | ---------- | -------------------------------------------------------- | -------------------------------------------------------- |
+|---------|------------| -------------------------------------------------------- |----------------------------------------------------------|
+| 0.2.27  | 2022-10-26 | [18481](https://github.com/airbytehq/airbyte/pull/18481) | Fix check for wrong format                               |
 | 0.2.26  | 2022-10-18 | [18116](https://github.com/airbytehq/airbyte/pull/18116) | Transform Dropbox shared link                            |
 | 0.2.25  | 2022-10-14 | [17994](https://github.com/airbytehq/airbyte/pull/17994) | Handle `UnicodeDecodeError` during discover step.        |
 | 0.2.24  | 2022-10-03 | [17504](https://github.com/airbytehq/airbyte/pull/17504) | Validate data for `HTTPS` while `check_connection`       |