✨ add support for remote sources (#344)

Author: sebastianMindee

.github/workflows/_test-integrations.yml

@@ -49,6 +49,7 @@ jobs:
           WORKFLOW_ID: ${{ secrets.WORKFLOW_ID_SE_TESTS }}
           MINDEE_V2_API_KEY: ${{ secrets.MINDEE_V2_SE_TESTS_API_KEY }}
           MINDEE_V2_FINDOC_MODEL_ID: ${{ secrets.MINDEE_V2_SE_TESTS_FINDOC_MODEL_ID }}
+          MINDEE_V2_SE_TESTS_BLANK_PDF_URL: ${{ secrets.MINDEE_V2_SE_TESTS_BLANK_PDF_URL }}
       run: |
         pytest -m integration
 

mindee/client.py

@@ -562,16 +562,3 @@ def create_endpoint(
             )
             version = "1"
         return self._build_endpoint(endpoint_name, account_name, version)
-
-    @staticmethod
-    def source_from_url(
-        url: str,
-    ) -> UrlInputSource:
-        """
-        Load a document from a URL.
-
-        :param url: Raw byte input
-        """
-        return UrlInputSource(
-            url,
-        )

mindee/client_mixin.py

@@ -2,7 +2,7 @@
 from typing import BinaryIO, Union
 
 from mindee.error import MindeeClientError
-from mindee.input import Base64Input, BytesInput, FileInput, PathInput
+from mindee.input import Base64Input, BytesInput, FileInput, PathInput, UrlInputSource
 
 
 class ClientMixin:
@@ -89,3 +89,16 @@ def _validate_async_params(
             )
         if max_retries < min_retries:
             raise MindeeClientError(f"Cannot set retries to less than {min_retries}.")
+
+    @staticmethod
+    def source_from_url(
+        url: str,
+    ) -> UrlInputSource:
+        """
+        Load a document from a URL.
+
+        :param url: Raw byte input
+        """
+        return UrlInputSource(
+            url,
+        )

mindee/client_v2.py

@@ -1,9 +1,10 @@
 from time import sleep
-from typing import Optional
+from typing import Optional, Union
 
 from mindee.client_mixin import ClientMixin
 from mindee.error.mindee_error import MindeeError
 from mindee.error.mindee_http_error_v2 import handle_error_v2
+from mindee.input import UrlInputSource
 from mindee.input.inference_parameters import InferenceParameters
 from mindee.input.polling_options import PollingOptions
 from mindee.input.sources.local_input_source import LocalInputSource
@@ -13,7 +14,7 @@
     is_valid_get_response,
     is_valid_post_response,
 )
-from mindee.parsing.v2.common_response import CommonStatus
+from mindee.parsing.v2.field.common_response import CommonStatus
 from mindee.parsing.v2.inference_response import InferenceResponse
 from mindee.parsing.v2.job_response import JobResponse
 
@@ -38,13 +39,14 @@ def __init__(self, api_key: Optional[str] = None) -> None:
         self.mindee_api = MindeeApiV2(api_key)
 
     def enqueue_inference(
-        self, input_source: LocalInputSource, params: InferenceParameters
+        self,
+        input_source: Union[LocalInputSource, UrlInputSource],
+        params: InferenceParameters,
     ) -> JobResponse:
         """
         Enqueues a document to a given model.
 
-        :param input_source: The document/source file to use.
-            Has to be created beforehand.
+        :param input_source: The document/source file to use. Can be local or remote.
 
         :param params: Parameters to set when sending a file.
         :return: A valid inference response.
@@ -95,13 +97,14 @@ def get_inference(self, inference_id: str) -> InferenceResponse:
         return InferenceResponse(dict_response)
 
     def enqueue_and_get_inference(
-        self, input_source: LocalInputSource, params: InferenceParameters
+        self,
+        input_source: Union[LocalInputSource, UrlInputSource],
+        params: InferenceParameters,
     ) -> InferenceResponse:
         """
         Enqueues to an asynchronous endpoint and automatically polls for a response.
 
-        :param input_source: The document/source file to use.
-            Has to be created beforehand.
+        :param input_source: The document/source file to use. Can be local or remote.
 
         :param params: Parameters to set when sending a file.
 

mindee/input/local_response.py

@@ -7,7 +7,7 @@
 from typing import Any, BinaryIO, Dict, Type, TypeVar, Union
 
 from mindee.error.mindee_error import MindeeError
-from mindee.parsing.v2.common_response import CommonResponse
+from mindee.parsing.v2.field.common_response import CommonResponse
 
 
 class LocalResponse:

mindee/mindee_http/mindee_api_v2.py

@@ -1,10 +1,10 @@
 import os
-from typing import Dict, Optional
+from typing import Dict, Optional, Union
 
 import requests
 
 from mindee.error.mindee_error import MindeeApiV2Error
-from mindee.input import LocalInputSource
+from mindee.input import LocalInputSource, UrlInputSource
 from mindee.input.inference_parameters import InferenceParameters
 from mindee.logger import logger
 from mindee.mindee_http.base_settings import USER_AGENT
@@ -68,7 +68,9 @@ def set_from_env(self) -> None:
                 logger.debug("Value was set from env: %s", name)
 
     def req_post_inference_enqueue(
-        self, input_source: LocalInputSource, params: InferenceParameters
+        self,
+        input_source: Union[LocalInputSource, UrlInputSource],
+        params: InferenceParameters,
     ) -> requests.Response:
         """
         Make an asynchronous request to POST a document for prediction on the V2 API.
@@ -87,14 +89,25 @@ def req_post_inference_enqueue(
         if params.alias and len(params.alias):
             data["alias"] = params.alias
 
-        files = {"file": input_source.read_contents(params.close_file)}
-        response = requests.post(
-            url=url,
-            files=files,
-            headers=self.base_headers,
-            data=data,
-            timeout=self.request_timeout,
-        )
+        if isinstance(input_source, LocalInputSource):
+            files = {"file": input_source.read_contents(params.close_file)}
+            response = requests.post(
+                url=url,
+                files=files,
+                headers=self.base_headers,
+                data=data,
+                timeout=self.request_timeout,
+            )
+        elif isinstance(input_source, UrlInputSource):
+            data["url"] = input_source.url
+            response = requests.post(
+                url=url,
+                headers=self.base_headers,
+                data=data,
+                timeout=self.request_timeout,
+            )
+        else:
+            raise MindeeApiV2Error("Invalid input source.")
         return response
 
     def req_get_job(self, job_id: str) -> requests.Response:

mindee/parsing/v2/field/__init__.py

@@ -1,6 +1,6 @@
 from typing import List, Optional
 
-from mindee.parsing.v2.dynamic_field import DynamicField
+from mindee.parsing.v2.field.dynamic_field import DynamicField
 
 
 class BaseField(DynamicField):

mindee/parsing/v2/base_field.py mindee/parsing/v2/field/base_field.pymindee/parsing/v2/base_field.py renamed to mindee/parsing/v2/field/base_field.py

@@ -34,13 +34,13 @@ def get_field_type(raw_response: StringDict, indent_level: int = 0) -> DynamicFi
     """Get appropriate field types."""
     if isinstance(raw_response, dict):
         if "value" in raw_response:
-            field_file = import_module("mindee.parsing.v2.simple_field")
+            field_file = import_module("mindee.parsing.v2.field.simple_field")
             field_class = getattr(field_file, FieldType.SIMPLE.value)
         elif "items" in raw_response:
-            field_file = import_module("mindee.parsing.v2.list_field")
+            field_file = import_module("mindee.parsing.v2.field.list_field")
             field_class = getattr(field_file, FieldType.LIST.value)
         elif "fields" in raw_response:
-            field_file = import_module("mindee.parsing.v2.object_field")
+            field_file = import_module("mindee.parsing.v2.field.object_field")
             field_class = getattr(field_file, FieldType.OBJECT.value)
         else:
             raise MindeeApiV2Error(f"Unrecognized field format in {raw_response}.")