Fix s3 client instantiation in _get_document_images_from_path

Fix s3 client instantiation in _get_document_images_from_path

Author: Belval

textractor/entities/lazy_document.py

@@ -27,6 +27,7 @@ def __init__(
         job_id: str,
         api: TextractAPI,
         textract_client=None,
+        s3_client=None,
         images=None,
         output_config: OutputConfig = None,
     ):
@@ -38,6 +39,7 @@ def __init__(
         self.job_id = job_id
         self._api = api
         self._textract_client = textract_client
+        self._s3_client = s3_client
         self._document = None
         self._images = images
         self._output_config = output_config
@@ -104,6 +106,7 @@ def __getattr__(self, __name: str) -> Any:
             "job_id",
             "_api",
             "_textract_client",
+            "_s3_client",
             "_document",
             "_images",
             "s3_polling_interval",
@@ -116,14 +119,13 @@ def __getattr__(self, __name: str) -> Any:
 
         if self._document is None:
             if self._output_config:
-                s3_client = boto3.client("s3")
                 start = time.time()
                 response = None
                 while not results_exist(
                     self.job_id,
                     self._output_config.s3_bucket,
                     self._output_config.s3_prefix,
-                    s3_client,
+                    self._s3_client,
                 ):
                     time.sleep(self._s3_polling_interval)
                     if time.time() - start > self._textract_polling_interval:
@@ -158,7 +160,7 @@ def __getattr__(self, __name: str) -> Any:
                     response = get_full_json_from_output_config(
                         self._output_config,
                         self.job_id,
-                        s3_client,
+                        self._s3_client,
                     )
             else:
                 if not self._textract_client:

textractor/textractor.py

@@ -127,12 +127,7 @@ def _get_document_images_from_path(self, filepath: str) -> List[Image.Image]:
             bucket = edit_filepath.split("/")[0]
             key = edit_filepath[edit_filepath.index("/") + 1 :]
 
-            s3_client = (
-                boto3.session.Session(profile_name=self.profile_name).client("s3")
-                if self.profile_name is not None
-                else boto3.session.Session(region_name=self.region_name).client("s3")
-            )
-            file_obj = s3_client.get_object(Bucket=bucket, Key=key).get("Body").read()
+            file_obj = self.s3_client.get_object(Bucket=bucket, Key=key).get("Body").read()
             if filepath.lower().endswith(".pdf"):
                 if IS_PDF_RENDERING_ENABLED:
                     images = rasterize_pdf(file_obj)
@@ -339,6 +334,7 @@ def start_document_text_detection(
             response["JobId"],
             TextractAPI.DETECT_TEXT,
             textract_client=self.textract_client,
+            s3_client=self.s3_client,
             images=images,
         )
 
@@ -582,6 +578,7 @@ def start_document_analysis(
             response["JobId"],
             TextractAPI.ANALYZE,
             textract_client=self.textract_client,
+            s3_client=self.s3_client,
             images=images,
             output_config=output_config,
         )
@@ -812,6 +809,7 @@ def start_expense_analysis(
             response["JobId"],
             TextractAPI.EXPENSE,
             textract_client=self.textract_client,
+            s3_client=self.s3_client,
             images=images,
         )