fix: Update AmazonBedrockDocumentEmbedder to not modify Documents in place

[GunaPalanivel]

Description

partially addresses #2174

Updates AmazonBedrockDocumentEmbedder to create new Document instances instead of modifying them in place, following the pattern established in Haystack core (#9693) and recent fixes in FastEmbed (#2678), Optimum (#2675), and Nvidia (#2680) integrations.

Changes

Code Changes

• Added from dataclasses import replace import
• Updated _embed_cohere() method to create new document instances using replace(doc, embedding=emb)
• Updated _embed_titan() method to create new document instances using replace(doc, embedding=emb)

Test Changes

• Added test_run_cohere_does_not_modify_original_documents() to verify Cohere path creates new instances
• Added test_run_titan_does_not_modify_original_documents() to verify Titan path creates new instances
• Both tests confirm original documents remain unchanged and new instances contain embeddings

Documentation

• Updated CHANGELOG.md with bug fix entry

Testing

All tests pass successfully:

• ✅ Format check: hatch run fmt-check
• ✅ Type checking: hatch run test:types
• ✅ Unit tests: hatch run test:unit (213 passed including 2 new immutability tests)

Notes

• AmazonBedrockDocumentImageEmbedder already uses the correct pattern (implemented in feat: Add AmazonBedrockDocumentImageEmbedder component #2185)
• No AmazonBedrockSparseDocumentEmbedder exists in this integration
• Follows the same immutability pattern as other integrations for consistency across the codebase

[sjrl]

@GunaPalanivel thanks for the changes! Could you also update the code here

haystack-core-integrations/integrations/amazon_bedrock/src/haystack_integrations/components/rankers/amazon_bedrock/ranker.py

Lines 249 to 255 in 1017c72

	sorted_docs = []
	for result in results:
	idx = result["index"]
	score = result["relevanceScore"]
	doc = documents[idx]
	doc.score = score
	sorted_docs.append(doc)

in the AmazonBedrockRanker to use replace as well

[GunaPalanivel]

@GunaPalanivel thanks for the changes! Could you also update the code here

haystack-core-integrations/integrations/amazon_bedrock/src/haystack_integrations/components/rankers/amazon_bedrock/ranker.py

Lines 249 to 255 in 1017c72

	sorted_docs = []
	for result in results:
	idx = result["index"]
	score = result["relevanceScore"]
	doc = documents[idx]
	doc.score = score
	sorted_docs.append(doc)

in the AmazonBedrockRanker to use replace as well

Added from dataclasses import replace import
Changed doc.score = score to replace(doc, score=score)
Now creates new document instances instead of modifying in place

[sjrl]

Thanks!