Functions — github.com/aws-samples/bedrock-access-gateway

hub / github.com/aws-samples/bedrock-access-gateway / functions

Functions48 in github.com/aws-samples/bedrock-access-gateway

⨍ Functions 48 ◇ Types & classes 40 ↳ Endpoints 5

↓ 7 callersMethod_resolve_to_foundation_model

Resolve any model identifier to foundation model ID for feature detection. Handles: - Cross-region profiles (us.*, eu.*, apa

src/api/models/bedrock.py:253

↓ 4 callersMethod_create_response

( self, model: str, message_id: str, content: list[dict] | None = None,

src/api/models/bedrock.py:912

↓ 4 callersMethodstream_response_to_bytes

(response: ChatStreamResponse | Error | None = None)

src/api/models/base.py:50

↓ 3 callersMethod_invoke_model

(self, args: dict, model_id: str)

src/api/models/bedrock.py:1288

↓ 2 callersMethod_convert_finish_reason

Below is a list of finish reason according to OpenAI doc: - stop: if the model hit a natural stop point or a provided stop sequence,

src/api/models/bedrock.py:1259

↓ 2 callersMethod_invoke_bedrock

Common logic for invoke bedrock models

src/api/models/bedrock.py:335

↓ 2 callersMethod_parse_content_parts

( self, message: UserMessage | AssistantMessage, model_id: str, )

src/api/models/bedrock.py:1187

↓ 2 callersMethod_supports_prompt_caching

Check if model supports prompt caching based on model ID pattern. Uses pattern matching instead of hardcoded whitelist for better ma

src/api/models/bedrock.py:275

↓ 2 callersMethodgenerate_message_id

()

src/api/models/base.py:46

↓ 2 callersFunctionlist_bedrock_models

Automatically getting a list of supported models. Returns a model list combines: - ON_DEMAND models. - Cross-Region Inference Pro

src/api/models/bedrock.py:111

↓ 2 callersMethodlist_models

Return a list of supported models

src/api/models/base.py:27

↓ 1 callersMethod_async_iterate

Helper method to convert sync iterator to async iterator

src/api/models/bedrock.py:411

↓ 1 callersMethod_calc_budget_tokens

( self, max_tokens: int, reasoning_effort: Literal["low", "medium", "high"] )

src/api/models/bedrock.py:1245

↓ 1 callersMethod_convert_tool_spec

(self, func: Function)

src/api/models/bedrock.py:1234

↓ 1 callersMethod_create_response_stream

Parsing the Bedrock stream response chunk. Ref: https://docs.aws.amazon.com/bedrock/latest/userguide/conversation-inference.html#message-infe

src/api/models/bedrock.py:1005

↓ 1 callersMethod_estimate_reasoning_tokens

Estimate reasoning tokens from reasoningContent blocks. Bedrock doesn't separately report reasoning tokens, so we estimate t

src/api/models/bedrock.py:895

↓ 1 callersMethod_extract_tool_content

Extract text content from various OpenAI SDK tool message formats. Handles: - String content (legacy format) - List of conten

src/api/models/bedrock.py:618

↓ 1 callersMethod_get_max_cache_tokens

Get maximum cacheable tokens limit for the model. Different models have different caching limits: - Claude: No explicit limi

src/api/models/bedrock.py:310

↓ 1 callersMethod_parse_args

(self, embeddings_request: EmbeddingsRequest)

src/api/models/bedrock.py:1341

↓ 1 callersMethod_parse_args

(self, embeddings_request: EmbeddingsRequest)

src/api/models/bedrock.py:1386

↓ 1 callersMethod_parse_args

(self, text: str, dimensions: int | None = None)

src/api/models/bedrock.py:1430

↓ 1 callersMethod_parse_image

Try to get the raw data from an image url. Ref: https://docs.aws.amazon.com/bedrock/latest/APIReference/API_runtime_ImageSource.html

src/api/models/bedrock.py:1158

↓ 1 callersMethod_parse_messages

Converse API only support user and assistant messages. example output: [{ "role": "user", "content": [{"text

src/api/models/bedrock.py:530

↓ 1 callersMethod_parse_request

Create default converse request body. Also perform validations to tool call etc. Ref: https://docs.aws.amazon.com/bedrock/latest/API

src/api/models/bedrock.py:767

↓ 1 callersMethod_parse_system_prompts

Create system prompts with optional prompt caching support. Prompt caching can be enabled via: 1. ENABLE_PROMPT_CACHING environment v

src/api/models/bedrock.py:465

↓ 1 callersMethod_reframe_multi_payloard

Receive messages and reformat them to comply with the Claude format With OpenAI format requests, it's not a problem to repeatedly receive mes

src/api/models/bedrock.py:669

↓ 1 callersMethodchat

Default implementation for Chat API.

src/api/models/bedrock.py:373

↓ 1 callersMethodchat_stream

Default implementation for Chat Stream API

src/api/models/bedrock.py:417

↓ 1 callersMethodembed

Handle a basic embeddings request.

src/api/models/base.py:73

↓ 1 callersFunctionget_embeddings_model

(model_id: str)

src/api/models/bedrock.py:1509

↓ 1 callersMethodis_supported_modality

(model_id: str, modality: str = "IMAGE")

src/api/models/bedrock.py:1227

↓ 1 callersMethodvalidate

Perform basic validation on requests

src/api/models/bedrock.py:221

↓ 1 callersFunctionvalidate_model_id

(model_id: str)

src/api/routers/model.py:18

Method_create_response

( self, embeddings: list[float], model: str, input_tokens: int = 0, ou

src/api/models/bedrock.py:1310

Functionapi_key_auth

( credentials: Annotated[HTTPAuthorizationCredentials, Depends(security)], )

src/api/auth.py:39

Methodchat

Handle a basic chat completion requests.

src/api/models/base.py:36

Functionchat_completions

( chat_request: Annotated[ ChatRequest, Body( examples=[ {

src/api/routers/chat.py:21

Methodchat_stream

Handle a basic chat completion requests with stream response.

src/api/models/base.py:41

Methodembed

(self, embeddings_request: EmbeddingsRequest)

src/api/models/bedrock.py:1370

Methodembed

(self, embeddings_request: EmbeddingsRequest)

src/api/models/bedrock.py:1410

Methodembed

(self, embeddings_request: EmbeddingsRequest)

src/api/models/bedrock.py:1446

Functionembeddings

( embeddings_request: Annotated[ EmbeddingsRequest, Body( examples=[

src/api/routers/embeddings.py:17

Functionget_model

( model_id: Annotated[ str, Path(description="Model ID", example="anthropic.claude-3-sonne

src/api/routers/model.py:33

Functionhealth

For health check if needed

src/api/app.py:49

Functionlist_models

()

src/api/routers/model.py:24

Methodlist_models

Always refresh the latest model list

src/api/models/bedrock.py:215

Methodvalidate

Validate chat completion requests.

src/api/models/base.py:31

Functionvalidation_exception_handler

(request, exc)

src/api/app.py:55