airbyte-cdk 6.7.0rc1__tar.gz → 6.7.0rc3__tar.gz

{airbyte_cdk-6.7.0rc1 → airbyte_cdk-6.7.0rc3}/PKG-INFO

@@ -1,13 +1,13 @@
 Metadata-Version: 2.1
 Name: airbyte-cdk
-Version: 6.7.0rc1
+Version: 6.7.0rc3
 Summary: A framework for writing Airbyte Connectors.
 Home-page: https://airbyte.com
 License: MIT
 Keywords: airbyte,connector-development-kit,cdk
 Author: Airbyte
 Author-email: contact@airbyte.io
-Requires-Python: >=3.10,<4.0
+Requires-Python: >=3.10,<3.13
 Classifier: Development Status :: 3 - Alpha
 Classifier: Intended Audience :: Developers
 Classifier: License :: OSI Approved :: MIT License
@@ -15,7 +15,6 @@ Classifier: Programming Language :: Python :: 3
 Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: Python :: 3.12
-Classifier: Programming Language :: Python :: 3.13
 Classifier: Topic :: Scientific/Engineering
 Classifier: Topic :: Software Development :: Libraries :: Python Modules
 Provides-Extra: file-based
@@ -38,11 +37,11 @@ Requires-Dist: fastavro (>=1.8.0,<1.9.0) ; extra == "file-based"
 Requires-Dist: genson (==1.3.0)
 Requires-Dist: isodate (>=0.6.1,<0.7.0)
 Requires-Dist: jsonref (>=0.2,<0.3)
-Requires-Dist: jsonschema (>=3.2.0,<3.3.0)
+Requires-Dist: jsonschema (>=4.17.3,<4.18.0)
 Requires-Dist: langchain (==0.1.16) ; extra == "vector-db-based"
 Requires-Dist: langchain_core (==0.1.42)
 Requires-Dist: markdown ; extra == "file-based"
-Requires-Dist: nltk (==3.8.1)
+Requires-Dist: nltk (==3.9.1)
 Requires-Dist: numpy (<2)
 Requires-Dist: openai[embeddings] (==0.27.9) ; extra == "vector-db-based"
 Requires-Dist: orjson (>=3.10.7,<4.0.0)
@@ -61,6 +60,7 @@ Requires-Dist: python-dateutil
 Requires-Dist: python-snappy (==0.7.3) ; extra == "file-based"
 Requires-Dist: python-ulid (>=3.0.0,<4.0.0)
 Requires-Dist: pytz (==2024.1)
+Requires-Dist: rapidfuzz (>=3.10.1,<4.0.0)
 Requires-Dist: requests
 Requires-Dist: requests_cache
 Requires-Dist: serpyco-rs (>=1.10.2,<2.0.0)

airbyte_cdk-6.7.0rc3/airbyte_cdk/cli/source_declarative_manifest/__init__.py

@@ -0,0 +1,6 @@
+from airbyte_cdk.cli.source_declarative_manifest._run import run
+
+
+__all__ = [
+    "run",
+]

airbyte_cdk-6.7.0rc3/airbyte_cdk/cli/source_declarative_manifest/_run.py

@@ -0,0 +1,223 @@
+# Copyright (c) 2024 Airbyte, Inc., all rights reserved.
+"""Defines the `source-declarative-manifest` connector, which installs alongside CDK.
+
+This file was originally imported from the dedicated connector directory, under the
+`airbyte` monorepo.
+
+Usage:
+
+```
+pipx install airbyte-cdk
+source-declarative-manifest --help
+source-declarative-manifest spec
+...
+```
+"""
+
+from __future__ import annotations
+
+import json
+import pkgutil
+import sys
+import traceback
+from collections.abc import Mapping
+from datetime import datetime
+from pathlib import Path
+from typing import Any, cast
+
+from airbyte_cdk.entrypoint import AirbyteEntrypoint, launch
+from airbyte_cdk.models import (
+    AirbyteErrorTraceMessage,
+    AirbyteMessage,
+    AirbyteMessageSerializer,
+    AirbyteStateMessage,
+    AirbyteTraceMessage,
+    ConfiguredAirbyteCatalog,
+    ConnectorSpecificationSerializer,
+    TraceType,
+    Type,
+)
+from airbyte_cdk.sources.declarative.concurrent_declarative_source import (
+    ConcurrentDeclarativeSource,
+)
+from airbyte_cdk.sources.declarative.yaml_declarative_source import YamlDeclarativeSource
+from airbyte_cdk.sources.source import TState
+from orjson import orjson
+
+
+class SourceLocalYaml(YamlDeclarativeSource):
+    """
+    Declarative source defined by a yaml file in the local filesystem
+    """
+
+    def __init__(
+        self,
+        catalog: ConfiguredAirbyteCatalog | None,
+        config: Mapping[str, Any] | None,
+        state: TState,
+        **kwargs: Any,
+    ) -> None:
+        """
+        HACK!
+            Problem: YamlDeclarativeSource relies on the calling module name/path to find the yaml file.
+            Implication: If you call YamlDeclarativeSource directly it will look for the yaml file in the wrong place. (e.g. the airbyte-cdk package)
+            Solution: Subclass YamlDeclarativeSource from the same location as the manifest to load.
+
+            When can we remove this?
+                When the airbyte-cdk is updated to not rely on the calling module name/path to find the yaml file.
+                When all manifest connectors are updated to use the new airbyte-cdk.
+                When all manifest connectors are updated to use the source-declarative-manifest as the base image.
+        """
+        super().__init__(
+            catalog=catalog,
+            config=config,
+            state=state,
+            path_to_yaml="manifest.yaml",
+        )
+
+
+def _is_local_manifest_command(args: list[str]) -> bool:
+    # Check for a local manifest.yaml file
+    return Path("/airbyte/integration_code/source_declarative_manifest/manifest.yaml").exists()
+
+
+def handle_command(args: list[str]) -> None:
+    if _is_local_manifest_command(args):
+        handle_local_manifest_command(args)
+    else:
+        handle_remote_manifest_command(args)
+
+
+def _get_local_yaml_source(args: list[str]) -> SourceLocalYaml:
+    try:
+        config, catalog, state = _parse_inputs_into_config_catalog_state(args)
+        return SourceLocalYaml(config=config, catalog=catalog, state=state)
+    except Exception as error:
+        print(
+            orjson.dumps(
+                AirbyteMessageSerializer.dump(
+                    AirbyteMessage(
+                        type=Type.TRACE,
+                        trace=AirbyteTraceMessage(
+                            type=TraceType.ERROR,
+                            emitted_at=int(datetime.now().timestamp() * 1000),
+                            error=AirbyteErrorTraceMessage(
+                                message=f"Error starting the sync. This could be due to an invalid configuration or catalog. Please contact Support for assistance. Error: {error}",
+                                stack_trace=traceback.format_exc(),
+                            ),
+                        ),
+                    )
+                )
+            ).decode()
+        )
+        raise error
+
+
+def handle_local_manifest_command(args: list[str]) -> None:
+    source = _get_local_yaml_source(args)
+    launch(
+        source=source,
+        args=args,
+    )
+
+
+def handle_remote_manifest_command(args: list[str]) -> None:
+    """Overrides the spec command to return the generalized spec for the declarative manifest source.
+
+    This is different from a typical low-code, but built and published separately source built as a ManifestDeclarativeSource,
+    because that will have a spec method that returns the spec for that specific source. Other than spec,
+    the generalized connector behaves the same as any other, since the manifest is provided in the config.
+    """
+    if args[0] == "spec":
+        json_spec = pkgutil.get_data(
+            "airbyte_cdk.cli.source_declarative_manifest",
+            "spec.json",
+        )
+        if json_spec is None:
+            raise FileNotFoundError(
+                "Could not find `spec.json` file for source-declarative-manifest"
+            )
+
+        spec_obj = json.loads(json_spec)
+        spec = ConnectorSpecificationSerializer.load(spec_obj)
+
+        message = AirbyteMessage(type=Type.SPEC, spec=spec)
+        print(AirbyteEntrypoint.airbyte_message_to_string(message))
+    else:
+        source = create_declarative_source(args)
+        launch(
+            source=source,
+            args=args,
+        )
+
+
+def create_declarative_source(args: list[str]) -> ConcurrentDeclarativeSource:
+    """Creates the source with the injected config.
+
+    This essentially does what other low-code sources do at build time, but at runtime,
+    with a user-provided manifest in the config. This better reflects what happens in the
+    connector builder.
+    """
+    try:
+        config, catalog, state = _parse_inputs_into_config_catalog_state(args)
+        if "__injected_declarative_manifest" not in config:
+            raise ValueError(
+                f"Invalid config: `__injected_declarative_manifest` should be provided at the root of the config but config only has keys {list(config.keys())}"
+            )
+        return ConcurrentDeclarativeSource(
+            config=config,
+            catalog=catalog,
+            state=state,
+            source_config=cast(dict[str, Any], config["__injected_declarative_manifest"]),
+        )
+    except Exception as error:
+        print(
+            orjson.dumps(
+                AirbyteMessageSerializer.dump(
+                    AirbyteMessage(
+                        type=Type.TRACE,
+                        trace=AirbyteTraceMessage(
+                            type=TraceType.ERROR,
+                            emitted_at=int(datetime.now().timestamp() * 1000),
+                            error=AirbyteErrorTraceMessage(
+                                message=f"Error starting the sync. This could be due to an invalid configuration or catalog. Please contact Support for assistance. Error: {error}",
+                                stack_trace=traceback.format_exc(),
+                            ),
+                        ),
+                    )
+                )
+            ).decode()
+        )
+        raise error
+
+
+def _parse_inputs_into_config_catalog_state(
+    args: list[str],
+) -> tuple[
+    Mapping[str, Any] | None,
+    ConfiguredAirbyteCatalog | None,
+    list[AirbyteStateMessage],
+]:
+    parsed_args = AirbyteEntrypoint.parse_args(args)
+    config = (
+        ConcurrentDeclarativeSource.read_config(parsed_args.config)
+        if hasattr(parsed_args, "config")
+        else None
+    )
+    catalog = (
+        ConcurrentDeclarativeSource.read_catalog(parsed_args.catalog)
+        if hasattr(parsed_args, "catalog")
+        else None
+    )
+    state = (
+        ConcurrentDeclarativeSource.read_state(parsed_args.state)
+        if hasattr(parsed_args, "state")
+        else []
+    )
+
+    return config, catalog, state
+
+
+def run() -> None:
+    args: list[str] = sys.argv[1:]
+    handle_command(args)

airbyte_cdk-6.7.0rc3/airbyte_cdk/cli/source_declarative_manifest/spec.json

@@ -0,0 +1,17 @@
+{
+  "documentationUrl": "https://docs.airbyte.com/integrations/sources/low-code",
+  "connectionSpecification": {
+    "$schema": "http://json-schema.org/draft-07/schema#",
+    "title": "Low-code source spec",
+    "type": "object",
+    "required": ["__injected_declarative_manifest"],
+    "additionalProperties": true,
+    "properties": {
+      "__injected_declarative_manifest": {
+        "title": "Low-code manifest",
+        "type": "object",
+        "description": "The low-code manifest that defines the components of the source."
+      }
+    }
+  }
+}

{airbyte_cdk-6.7.0rc1 → airbyte_cdk-6.7.0rc3}/airbyte_cdk/sources/declarative/declarative_component_schema.yaml

@@ -2790,21 +2790,21 @@ interpolation:
         - created_at: "2020-01-01 00:00:00.000+00:00"
         - updated_at: "2020-01-02 00:00:00.000+00:00"
   macros:
-    - title: Now (UTC)
+    - title: now_utc
       description: Returns the current date and time in the UTC timezone.
       arguments: {}
       return_type: Datetime
       examples:
         - "'{{ now_utc() }}' -> '2021-09-01 00:00:00+00:00'"
         - "'{{ now_utc().strftime('%Y-%m-%d') }}' -> '2021-09-01'"
-    - title: Today (UTC)
+    - title: today_utc
       description: Returns the current date in UTC timezone. The output is a date object.
       arguments: {}
       return_type: Date
       examples:
         - "'{{ today_utc() }}' -> '2021-09-01'"
         - "'{{ today_utc().strftime('%Y/%m/%d')}}' -> '2021/09/01'"
-    - title: Timestamp
+    - title: timestamp
       description: Converts a number or a string representing a datetime (formatted as ISO8601) to a timestamp. If the input is a number, it is converted to an int. If no timezone is specified, the string is interpreted as UTC.
       arguments:
         datetime: A string formatted as ISO8601 or an integer representing a unix timestamp
@@ -2815,7 +2815,7 @@ interpolation:
         - "'{{ timestamp('2022-02-28T00:00:00Z') }}' -> 1646006400"
         - "'{{ timestamp('2022-02-28 00:00:00Z') }}' -> 1646006400"
         - "'{{ timestamp('2022-02-28T00:00:00-08:00') }}' -> 1646035200"
-    - title: Max
+    - title: max
       description: Returns the largest object of a iterable, or or two or more arguments.
       arguments:
         args: iterable or a sequence of two or more arguments
@@ -2823,7 +2823,7 @@ interpolation:
       examples:
         - "'{{ max(2, 3) }}' -> 3"
         - "'{{ max([2, 3]) }}' -> 3"
-    - title: Day Delta
+    - title: day_delta
       description: Returns the datetime of now() + num_days.
       arguments:
         num_days: The number of days to add to now
@@ -2833,8 +2833,8 @@ interpolation:
         - "'{{ day_delta(1) }}' -> '2021-09-02T00:00:00.000000+0000'"
         - "'{{ day_delta(-1) }}' -> '2021-08-31:00:00.000000+0000'"
         - "'{{ day_delta(25, format='%Y-%m-%d') }}' -> '2021-09-02'"
-    - title: Duration
-      description: Converts an ISO8601 duratioin to datetime.timedelta.
+    - title: duration
+      description: Converts an ISO8601 duration to datetime timedelta.
       arguments:
         duration_string: "A string representing an ISO8601 duration. See https://www.digi.com/resources/documentation/digidocs//90001488-13/reference/r_iso_8601_duration_format.htm for more details."
       return_type: datetime.timedelta
@@ -2842,7 +2842,7 @@ interpolation:
         - "'{{ duration('P1D') }}' -> '1 day, 0:00:00'"
         - "'{{ duration('P6DT23H') }}' -> '6 days, 23:00:00'"
         - "'{{ (now_utc() - duration('P1D')).strftime('%Y-%m-%dT%H:%M:%SZ') }}' -> '2021-08-31T00:00:00Z'"
-    - title: Format Datetime
+    - title: format_datetime
       description: Converts a datetime or a datetime-string to the specified format.
       arguments:
         datetime: The datetime object or a string to convert. If datetime is a string, it must be formatted as ISO8601.
@@ -2854,7 +2854,7 @@ interpolation:
         - "{{ format_datetime(config['start_date'], '%Y-%m-%dT%H:%M:%S.%fZ') }}"
         - "{{ format_datetime(config['start_date'], '%Y-%m-%dT%H:%M:%S.%fZ', '%a, %d %b %Y %H:%M:%S %z') }}"
   filters:
-    - title: Hash
+    - title: hash
       description: Convert the specified value to a hashed string.
       arguments:
         hash_type: Valid hash type for converts ('md5' as default value).
@@ -2864,26 +2864,26 @@ interpolation:
         - "{{ 'Test client_secret' | hash() }} -> '3032d57a12f76b61a820e47b9a5a0cbb'"
         - "{{ 'Test client_secret' | hash('md5') }} -> '3032d57a12f76b61a820e47b9a5a0cbb'"
         - "{{ 'Test client_secret' | hash('md5', salt='salt') }} -> '5011a0168579c2d94cbbe1c6ad14327c'"
-    - title: Base64 encoder
+    - title: base64encode
       description: Convert the specified value to a string in the base64 format.
       arguments: {}
       return_type: str
       examples:
         - "{{ 'Test client_secret' | base64encode }} -> 'VGVzdCBjbGllbnRfc2VjcmV0'"
-    - title: Base64 decoder
+    - title: base64decode
       description: Decodes the specified base64 format value into a common string.
       arguments: {}
       return_type: str
       examples:
         - "{{ 'ZmFrZSByZWZyZXNoX3Rva2VuIHZhbHVl' | base64decode }} -> 'fake refresh_token value'"
-    - title: String
+    - title: string
       description: Converts the specified value to a string.
       arguments: {}
       return_type: str
       examples:
         - '{{ 1 | string }} -> "1"'
         - '{{ ["hello", "world" | string }} -> "["hello", "world"]"'
-    - title: Regex Search
+    - title: regex_search
       description: Match the input string against a regular expression and return the first match.
       arguments:
         regex: The regular expression to search for. It must include a capture group.

{airbyte_cdk-6.7.0rc1 → airbyte_cdk-6.7.0rc3}/airbyte_cdk/sources/declarative/manifest_declarative_source.py

@@ -5,10 +5,10 @@
 import json
 import logging
 import pkgutil
-import re
 from copy import deepcopy
 from importlib import metadata
-from typing import Any, Dict, Iterator, List, Mapping, Optional, Tuple
+from typing import Any, Dict, Iterator, List, Mapping, Optional
+from packaging.version import Version, InvalidVersion
 
 import yaml
 from airbyte_cdk.models import (
@@ -245,45 +245,54 @@ class ManifestDeclarativeSource(DeclarativeSource):
                 "Validation against json schema defined in declarative_component_schema.yaml schema failed"
             ) from e
 
-        cdk_version = metadata.version("airbyte_cdk")
-        cdk_major, cdk_minor, cdk_patch = self._get_version_parts(cdk_version, "airbyte-cdk")
-        manifest_version = self._source_config.get("version")
-        if manifest_version is None:
+        cdk_version_str = metadata.version("airbyte_cdk")
+        cdk_version = self._parse_version(cdk_version_str, "airbyte-cdk")
+        manifest_version_str = self._source_config.get("version")
+        if manifest_version_str is None:
             raise RuntimeError(
                 "Manifest version is not defined in the manifest. This is unexpected since it should be a required field. Please contact support."
             )
-        manifest_major, manifest_minor, manifest_patch = self._get_version_parts(
-            manifest_version, "manifest"
-        )
+        manifest_version = self._parse_version(manifest_version_str, "manifest")
 
-        if cdk_version.startswith("0.0.0"):
+        if (cdk_version.major, cdk_version.minor, cdk_version.micro) == (0, 0, 0):
             # Skipping version compatibility check on unreleased dev branch
             pass
-        elif cdk_major < manifest_major or (
-            cdk_major == manifest_major and cdk_minor < manifest_minor
+        elif (cdk_version.major, cdk_version.minor) < (
+            manifest_version.major,
+            manifest_version.minor,
         ):
             raise ValidationError(
-                f"The manifest version {manifest_version} is greater than the airbyte-cdk package version ({cdk_version}). Your "
+                f"The manifest version {manifest_version!s} is greater than the airbyte-cdk package version ({cdk_version!s}). Your "
                 f"manifest may contain features that are not in the current CDK version."
             )
-        elif manifest_major == 0 and manifest_minor < 29:
+        elif (manifest_version.major, manifest_version.minor) < (0, 29):
             raise ValidationError(
                 f"The low-code framework was promoted to Beta in airbyte-cdk version 0.29.0 and contains many breaking changes to the "
-                f"language. The manifest version {manifest_version} is incompatible with the airbyte-cdk package version "
-                f"{cdk_version} which contains these breaking changes."
+                f"language. The manifest version {manifest_version!s} is incompatible with the airbyte-cdk package version "
+                f"{cdk_version!s} which contains these breaking changes."
             )
 
     @staticmethod
-    def _get_version_parts(version: str, version_type: str) -> Tuple[int, int, int]:
-        """
-        Takes a semantic version represented as a string and splits it into a tuple of its major, minor, and patch versions.
+    def _parse_version(
+        version: str,
+        version_type: str,
+    ) -> Version:
+        """Takes a semantic version represented as a string and splits it into a tuple.
+
+        The fourth part (prerelease) is not returned in the tuple.
+
+        Returns:
+            Version: the parsed version object
         """
-        version_parts = re.split(r"\.", version)
-        if len(version_parts) != 3 or not all([part.isdigit() for part in version_parts]):
+        try:
+            parsed_version = Version(version)
+        except InvalidVersion as ex:
             raise ValidationError(
-                f"The {version_type} version {version} specified is not a valid version format (ex. 1.2.3)"
-            )
-        return tuple(int(part) for part in version_parts)  # type: ignore # We already verified there were 3 parts and they are all digits
+                f"The {version_type} version '{version}' is not a valid version format."
+            ) from ex
+        else:
+            # No exception
+            return parsed_version
 
     def _stream_configs(self, manifest: Mapping[str, Any]) -> List[Dict[str, Any]]:
         # This has a warning flag for static, but after we finish part 4 we'll replace manifest with self._source_config

{airbyte_cdk-6.7.0rc1 → airbyte_cdk-6.7.0rc3}/airbyte_cdk/sources/file_based/file_types/unstructured_parser.py

@@ -29,16 +29,25 @@ from airbyte_cdk.sources.file_based.schema_helpers import SchemaType
 from airbyte_cdk.utils import is_cloud_environment
 from airbyte_cdk.utils.traced_exception import AirbyteTracedException
 from unstructured.file_utils.filetype import (
+    EXT_TO_FILETYPE,
     FILETYPE_TO_MIMETYPE,
     STR_TO_FILETYPE,
     FileType,
     detect_filetype,
 )
+import nltk
 
 unstructured_partition_pdf = None
 unstructured_partition_docx = None
 unstructured_partition_pptx = None
 
+try:
+    nltk.data.find("tokenizers/punkt.zip")
+    nltk.data.find("tokenizers/punkt_tab.zip")
+except LookupError:
+    nltk.download("punkt")
+    nltk.download("punkt_tab")
+
 
 def optional_decode(contents: Union[str, bytes]) -> str:
     if isinstance(contents, bytes):
@@ -108,9 +117,11 @@ class UnstructuredParser(FileTypeParser):
         format = _extract_format(config)
         with stream_reader.open_file(file, self.file_read_mode, None, logger) as file_handle:
             filetype = self._get_filetype(file_handle, file)
-
             if filetype not in self._supported_file_types() and not format.skip_unprocessable_files:
-                raise self._create_parse_error(file, self._get_file_type_error_message(filetype))
+                raise self._create_parse_error(
+                    file,
+                    self._get_file_type_error_message(filetype),
+                )
 
             return {
                 "content": {
@@ -159,6 +170,10 @@ class UnstructuredParser(FileTypeParser):
                     logger.warn(f"File {file.uri} cannot be parsed. Skipping it.")
                 else:
                     raise e
+            except Exception as e:
+                exception_str = str(e)
+                logger.error(f"File {file.uri} caused an error during parsing: {exception_str}.")
+                raise e
 
     def _read_file(
         self,
@@ -176,20 +191,32 @@ class UnstructuredParser(FileTypeParser):
             # check whether unstructured library is actually available for better error message and to ensure proper typing (can't be None after this point)
             raise Exception("unstructured library is not available")
 
-        filetype = self._get_filetype(file_handle, remote_file)
+        filetype: FileType | None = self._get_filetype(file_handle, remote_file)
 
-        if filetype == FileType.MD or filetype == FileType.TXT:
+        if filetype is None or filetype not in self._supported_file_types():
+            raise self._create_parse_error(
+                remote_file,
+                self._get_file_type_error_message(filetype),
+            )
+        if filetype in {FileType.MD, FileType.TXT}:
             file_content: bytes = file_handle.read()
             decoded_content: str = optional_decode(file_content)
             return decoded_content
-        if filetype not in self._supported_file_types():
-            raise self._create_parse_error(remote_file, self._get_file_type_error_message(filetype))
         if format.processing.mode == "local":
-            return self._read_file_locally(file_handle, filetype, format.strategy, remote_file)
+            return self._read_file_locally(
+                file_handle,
+                filetype,
+                format.strategy,
+                remote_file,
+            )
         elif format.processing.mode == "api":
             try:
                 result: str = self._read_file_remotely_with_retries(
-                    file_handle, format.processing, filetype, format.strategy, remote_file
+                    file_handle,
+                    format.processing,
+                    filetype,
+                    format.strategy,
+                    remote_file,
                 )
             except Exception as e:
                 # If a parser error happens during remotely processing the file, this means the file is corrupted. This case is handled by the parse_records method, so just rethrow.
@@ -336,7 +363,11 @@ class UnstructuredParser(FileTypeParser):
 
         return self._render_markdown([element.to_dict() for element in elements])
 
-    def _create_parse_error(self, remote_file: RemoteFile, message: str) -> RecordParseError:
+    def _create_parse_error(
+        self,
+        remote_file: RemoteFile,
+        message: str,
+    ) -> RecordParseError:
         return RecordParseError(
             FileBasedSourceError.ERROR_PARSING_RECORD, filename=remote_file.uri, message=message
         )
@@ -360,32 +391,51 @@ class UnstructuredParser(FileTypeParser):
         # detect_filetype is either using the file name or file content
         # if possible, try to leverage the file name to detect the file type
         # if the file name is not available, use the file content
-        file_type = detect_filetype(
-            filename=remote_file.uri,
-        )
-        if file_type is not None and not file_type == FileType.UNK:
+        file_type: FileType | None = None
+        try:
+            file_type = detect_filetype(
+                filename=remote_file.uri,
+            )
+        except Exception:
+            # Path doesn't exist locally. Try something else...
+            pass
+
+        if file_type and file_type != FileType.UNK:
             return file_type
 
         type_based_on_content = detect_filetype(file=file)
+        file.seek(0)  # detect_filetype is reading to read the file content, so we need to reset
 
-        # detect_filetype is reading to read the file content
-        file.seek(0)
+        if type_based_on_content and type_based_on_content != FileType.UNK:
+            return type_based_on_content
 
-        return type_based_on_content
+        extension = "." + remote_file.uri.split(".")[-1].lower()
+        if extension in EXT_TO_FILETYPE:
+            return EXT_TO_FILETYPE[extension]
+
+        return None
 
     def _supported_file_types(self) -> List[Any]:
         return [FileType.MD, FileType.PDF, FileType.DOCX, FileType.PPTX, FileType.TXT]
 
-    def _get_file_type_error_message(self, file_type: FileType) -> str:
+    def _get_file_type_error_message(
+        self,
+        file_type: FileType | None,
+    ) -> str:
         supported_file_types = ", ".join([str(type) for type in self._supported_file_types()])
-        return f"File type {file_type} is not supported. Supported file types are {supported_file_types}"
+        return f"File type {file_type or 'None'!s} is not supported. Supported file types are {supported_file_types}"
 
     def _render_markdown(self, elements: List[Any]) -> str:
         return "\n\n".join((self._convert_to_markdown(el) for el in elements))
 
     def _convert_to_markdown(self, el: Dict[str, Any]) -> str:
         if dpath.get(el, "type") == "Title":
-            heading_str = "#" * (dpath.get(el, "metadata/category_depth", default=1) or 1)
+            category_depth = dpath.get(el, "metadata/category_depth", default=1) or 1
+            if not isinstance(category_depth, int):
+                category_depth = (
+                    int(category_depth) if isinstance(category_depth, (str, float)) else 1
+                )
+            heading_str = "#" * category_depth
             return f"{heading_str} {dpath.get(el, 'text')}"
         elif dpath.get(el, "type") == "ListItem":
             return f"- {dpath.get(el, 'text')}"

airbyte_cdk-6.7.0rc3/airbyte_cdk/test/utils/__init__.py

@@ -0,0 +1 @@
+# Copyright (c) 2024 Airbyte, Inc., all rights reserved.