feat(anthropic): add ChatAnthropicBedrock wrapper

[NITIN Madan (Nitin75408)]

1. PR title:feat(anthropic): add ChatAnthropicBedrock wrapper for AWS Bedrock with full feature parity

2. PR description:

Adds ChatAnthropicBedrock to langchain-anthropic, wrapping Anthropic's AnthropicBedrock client to provide feature parity with ChatAnthropic using AWS credentials. This addresses gaps in ChatBedrockConverse (in langchain-aws), which uses AWS's Converse API and exposes only a subset of Anthropic features.
Fixes #34757

Changes
Core Implementation:
Added ChatAnthropicBedrock inheriting from ChatAnthropic to reuse generation, streaming, and tool binding
Uses AnthropicBedrock/AsyncAnthropicBedrock clients instead of the standard Anthropic client
Supports AWS credentials (access key, secret key, session token, region) with environment variable fallback

Shared Utilities:
Created langchain_anthropic/utils.py with resolve_aws_credentials() and create_bedrock_client_params() for reusable AWS credential handling

Extensibility Hook:
Added _get_ls_provider() to BaseChatModel to allow subclasses to customize LangSmith provider identification
Provider-agnostic and backward compatible

Features Supported
parallel_tool_calls control (enable/disable parallel tool execution)
thinking parameter (extended thinking mode)
strict tool schema validation
Configurable prompt caching (5m or 1h TTL)
Reliable streaming with tool use
Full LCEL compatibility

Files Modified
libs/partners/anthropic/langchain_anthropic/chat_models.py - Added ChatAnthropicBedrock class
libs/partners/anthropic/langchain_anthropic/init.py - Exported ChatAnthropicBedrock
libs/partners/anthropic/langchain_anthropic/utils.py - Added shared AWS credential utilities
libs/core/langchain_core/language_models/chat_models.py - Added _get_ls_provider() extensibility hook

Testing
Added 12 unit tests for utils functions covering credential resolution edge cases
Added 2 integration tests verifying ChatAnthropicBedrock uses utils correctly
Added 3 tests for the core extensibility hook
All existing tests pass; no breaking changes

Verification
Ran make format, make lint, and make test in libs/partners/anthropic and libs/core
Verified ChatAnthropicBedrock initializes with AWS credentials
Confirmed inheritance of ChatAnthropic methods (bind_tools, with_structured_output, streaming)
Tested _get_ls_provider() override for custom provider identification
Validated backward compatibility with existing ChatAnthropic usage

AI Disclosure
This PR was developed with assistance from AI coding tools. The implementation follows LangChain patterns, maintains backward compatibility, and includes tests.

Breaking Changes
None. This is an additive change that does not modify existing ChatAnthropic or ChatBedrockConverse behavior.

[codspeed-hq]

Merging this PR will improve performance by 28.57%

⚠️ Unknown Walltime execution environment detected

Using the Walltime instrument on standard Hosted Runners will lead to inconsistent data.

For the most accurate results, we recommend using CodSpeed Macro Runners: bare-metal machines fine-tuned for performance measurement consistency.

⚡ 7 improved benchmarks
✅ 9 untouched benchmarks
🆕 1 new benchmark
⏩ 19 skipped benchmarks¹

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
⚡	WallTime	test_import_time[tool]	542.3 ms	490.8 ms	+10.49%
⚡	WallTime	test_import_time[CallbackManager]	488.6 ms	443.4 ms	+10.19%
⚡	WallTime	test_import_time[PydanticOutputParser]	563.5 ms	505 ms	+11.58%
⚡	WallTime	test_import_time[LangChainTracer]	472.1 ms	421 ms	+12.11%
⚡	WallTime	test_async_callbacks_in_sync	23.5 ms	18.3 ms	+28.57%
⚡	WallTime	test_import_time[Runnable]	520.1 ms	471.7 ms	+10.27%
⚡	WallTime	test_import_time[Document]	191.6 ms	172.9 ms	+10.86%
🆕	Simulation	test_init_time	N/A	1.1 ms	N/A

---

_{Comparing Nitin75408:Fixes/34757 (296f1ae) with master (0382606)}

Footnotes

1. 19 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩