Merge branch 'develop' of https://github.com/aws-solutions-library-samples/accelerated-intelligent-document-processing-on-aws into develop

Author: Bob Strahan

lib/idp_common_pkg/idp_common/extraction/agentic_idp.py

@@ -47,6 +47,33 @@
     view_todo_list,
 )
 
+# Supported image formats for Bedrock API
+SUPPORTED_IMAGE_FORMATS = {"jpeg", "png", "gif", "webp"}
+
+
+def detect_image_format(image_bytes: bytes) -> str:
+    """
+    Detect the image format from raw bytes.
+
+    Args:
+        image_bytes: Raw image bytes
+
+    Returns:
+        Image format string suitable for Bedrock API ('jpeg', 'png', 'gif', 'webp')
+
+    Raises:
+        ValueError: If the image format is unsupported or cannot be detected
+    """
+    img = Image.open(io.BytesIO(image_bytes))
+
+    if not img.format or img.format.lower() not in SUPPORTED_IMAGE_FORMATS:
+        raise ValueError(
+            f"Unsupported image format: {img.format}. "
+            f"Supported formats: {', '.join(SUPPORTED_IMAGE_FORMATS)}"
+        )
+    return img.format.lower()
+
+
 # Use AWS Lambda Powertools Logger for structured logging
 # Automatically logs as JSON with Lambda context, request_id, timestamp, etc.
 # In Lambda: Full JSON structured logs
@@ -207,11 +234,15 @@ def view_image(image_index: int, agent: Agent) -> dict:
         # Get the base image (already has grid overlay)
         img_bytes = page_images[image_index]
 
+        # Detect actual image format from bytes
+        img_format = detect_image_format(img_bytes)
+
         logger.info(
             "Returning image to agent",
             extra={
                 "image_index": image_index,
                 "image_size_bytes": len(img_bytes),
+                "image_format": img_format,
             },
         )
 
@@ -220,7 +251,7 @@ def view_image(image_index: int, agent: Agent) -> dict:
             "content": [
                 {
                     "image": {
-                        "format": "png",
+                        "format": img_format,
                         "source": {
                             "bytes": img_bytes,
                         },
@@ -775,12 +806,17 @@ def _prepare_prompt_content(
             extra={"image_count": len(page_images)},
         )
 
-        prompt_content += [
-            ContentBlock(
-                image=ImageContent(format="png", source=ImageSource(bytes=img_bytes))
+        for img_bytes in page_images:
+            # Detect actual image format from bytes
+            img_format = detect_image_format(img_bytes)
+            prompt_content.append(
+                ContentBlock(
+                    image=ImageContent(
+                        format=img_format,  # pyright: ignore[reportArgumentType]
+                        source=ImageSource(bytes=img_bytes),
+                    )
+                )
             )
-            for img_bytes in page_images
-        ]
 
     # Add existing data context if provided
     if existing_data:

lib/idp_common_pkg/tests/unit/test_delete_tests.py

@@ -14,13 +14,19 @@
 sys.modules["idp_common_pkg"] = Mock()
 sys.modules["idp_common_pkg.logger"] = Mock()
 
-# Add the lambda directory to the path for importing
-lambda_path = os.path.join(
-    os.path.dirname(__file__), "../../../../src/lambda/delete_tests"
-)
-sys.path.insert(0, lambda_path)
+# Mock boto3 before importing the Lambda module to prevent NoRegionError
+# The Lambda creates boto3 clients at module level which requires AWS region
+with patch("boto3.resource") as mock_resource, patch("boto3.client") as mock_client:
+    mock_resource.return_value = Mock()
+    mock_client.return_value = Mock()
+
+    # Add the lambda directory to the path for importing
+    lambda_path = os.path.join(
+        os.path.dirname(__file__), "../../../../src/lambda/delete_tests"
+    )
+    sys.path.insert(0, lambda_path)
 
-import index  # type: ignore[import-untyped]  # noqa: E402
+    import index  # type: ignore[import-untyped]  # noqa: E402
 
 
 @pytest.mark.unit

lib/idp_common_pkg/tests/unit/test_results_resolver.py

@@ -8,18 +8,24 @@
 
 import pytest
 
-# Import the specific lambda module using importlib to avoid conflicts
-spec = importlib.util.spec_from_file_location(
-    "results_index",
-    os.path.join(
-        os.path.dirname(__file__),
-        "../../../../src/lambda/test_results_resolver/index.py",
-    ),
-)
-if spec is None or spec.loader is None:
-    raise ImportError("Could not load test_results_resolver module")
-index = importlib.util.module_from_spec(spec)
-spec.loader.exec_module(index)
+# Mock boto3 before importing the Lambda module to prevent NoRegionError
+# The Lambda creates boto3 clients at module level which requires AWS region
+with patch("boto3.resource") as mock_resource, patch("boto3.client") as mock_client:
+    mock_resource.return_value = Mock()
+    mock_client.return_value = Mock()
+
+    # Import the specific lambda module using importlib to avoid conflicts
+    spec = importlib.util.spec_from_file_location(
+        "results_index",
+        os.path.join(
+            os.path.dirname(__file__),
+            "../../../../src/lambda/test_results_resolver/index.py",
+        ),
+    )
+    if spec is None or spec.loader is None:
+        raise ImportError("Could not load test_results_resolver module")
+    index = importlib.util.module_from_spec(spec)
+    spec.loader.exec_module(index)
 
 
 @pytest.mark.unit