Skip to content

Bedrock titan embedding driver

BedrockTitanEmbeddingDriver

Bases: BaseEmbeddingDriver

Attributes:

Name Type Description
model str

Embedding model name. Defaults to DEFAULT_MODEL.
dimensions int

Vector dimensions. Defaults to DEFAULT_MAX_TOKENS.
tokenizer BedrockTitanTokenizer

Optionally provide custom BedrockTitanTokenizer.
session Session

Optionally provide custom boto3.Session.
bedrock_client Any

Optionally provide custom bedrock-runtime client.
Source code in griptape/griptape/drivers/embedding/bedrock_titan_embedding_driver.py 
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
@define
class BedrockTitanEmbeddingDriver(BaseEmbeddingDriver):
    """
    Attributes:
        model: Embedding model name. Defaults to DEFAULT_MODEL.
        dimensions: Vector dimensions. Defaults to DEFAULT_MAX_TOKENS.
        tokenizer: Optionally provide custom `BedrockTitanTokenizer`.
        session: Optionally provide custom `boto3.Session`.
        bedrock_client: Optionally provide custom `bedrock-runtime` client.
    """

    DEFAULT_MODEL = "amazon.titan-embed-text-v1"
    DEFAULT_MAX_TOKENS = 1536

    model: str = field(default=DEFAULT_MODEL, kw_only=True)
    dimensions: int = field(default=DEFAULT_MAX_TOKENS, kw_only=True)
    session: boto3.Session = field(default=Factory(lambda: import_optional_dependency("boto3").Session()), kw_only=True)
    tokenizer: BedrockTitanTokenizer = field(
        default=Factory(lambda self: BedrockTitanTokenizer(model=self.model, session=self.session), takes_self=True),
        kw_only=True,
    )
    bedrock_client: Any = field(
        default=Factory(lambda self: self.session.client("bedrock-runtime"), takes_self=True), kw_only=True
    )

    def try_embed_chunk(self, chunk: str) -> list[float]:
        payload = {"inputText": chunk}

        response = self.bedrock_client.invoke_model(
            body=json.dumps(payload), modelId=self.model, accept="application/json", contentType="application/json"
        )
        response_body = json.loads(response.get("body").read())

        return response_body.get("embedding")

DEFAULT_MAX_TOKENS = 1536 class-attribute instance-attribute

DEFAULT_MODEL = 'amazon.titan-embed-text-v1' class-attribute instance-attribute

bedrock_client: Any = field(default=Factory(lambda : self.session.client('bedrock-runtime'), takes_self=True), kw_only=True) class-attribute instance-attribute

dimensions: int = field(default=DEFAULT_MAX_TOKENS, kw_only=True) class-attribute instance-attribute

model: str = field(default=DEFAULT_MODEL, kw_only=True) class-attribute instance-attribute

session: boto3.Session = field(default=Factory(lambda : import_optional_dependency('boto3').Session()), kw_only=True) class-attribute instance-attribute

tokenizer: BedrockTitanTokenizer = field(default=Factory(lambda : BedrockTitanTokenizer(model=self.model, session=self.session), takes_self=True), kw_only=True) class-attribute instance-attribute

try_embed_chunk(chunk)

Source code in griptape/griptape/drivers/embedding/bedrock_titan_embedding_driver.py 
38
39
40
41
42
43
44
45
46
def try_embed_chunk(self, chunk: str) -> list[float]:
    payload = {"inputText": chunk}

    response = self.bedrock_client.invoke_model(
        body=json.dumps(payload), modelId=self.model, accept="application/json", contentType="application/json"
    )
    response_body = json.loads(response.get("body").read())

    return response_body.get("embedding")