google.ai.generativelanguage.GenerativeServiceClient

API for using Large Models that generate multimodal content and have additional capabilities beyond text generation.

google.ai.generativelanguage.GenerativeServiceClient(
    *,
    credentials: Optional[ga_credentials.Credentials] = None,
    transport: Optional[Union[str, GenerativeServiceTransport]] = None,
    client_options: Optional[Union[client_options_lib.ClientOptions, dict]] = None,
    client_info: gapic_v1.client_info.ClientInfo = DEFAULT_CLIENT_INFO
) -> None

credentials Optional[google.auth.credentials.Credentials]

The authorization credentials to attach to requests. These credentials identify the application to the service; if none are specified, the client will attempt to ascertain the credentials from the environment.

transport Union[str, GenerativeServiceTransport]

The transport to use. If set to None, a transport is chosen automatically.

client_options Optional[Union[google.api_core.client_options.ClientOptions, dict]]

Custom options for the client. It won't take effect if a transport instance is provided. (1) The api_endpoint property can be used to override the default endpoint provided by the client. GOOGLE_API_USE_MTLS_ENDPOINT environment variable can also be used to override the endpoint: "always" (always use the default mTLS endpoint), "never" (always use the default regular endpoint) and "auto" (auto switch to the default mTLS endpoint if client certificate is present, this is the default value). However, the api_endpoint property takes precedence if provided. (2) If GOOGLE_API_USE_CLIENT_CERTIFICATE environment variable is "true", then the client_cert_source property can be used to provide client certificate for mutual TLS transport. If not provided, the default SSL client certificate will be used if present. If GOOGLE_API_USE_CLIENT_CERTIFICATE is "false" or not set, no client certificate will be used.

client_info google.api_core.gapic_v1.client_info.ClientInfo

The client info used to send a user-agent string along with API requests. If None, then default info will be used. Generally, you only need to set this if you're developing your own client library.

google.auth.exceptions.MutualTLSChannelError If mutual TLS transport creation failed for any reason.

transport Returns the transport used by the client instance.

Methods

batch_embed_contents

View source

batch_embed_contents(
    request: Optional[Union[google.ai.generativelanguage.BatchEmbedContentsRequest, dict]] = None,
    *,
    model: Optional[str] = None,
    requests: Optional[MutableSequence[generative_service.EmbedContentRequest]] = None,
    retry: OptionalRetry = gapic_v1.method.DEFAULT,
    timeout: Union[float, object] = gapic_v1.method.DEFAULT,
    metadata: Sequence[Tuple[str, str]] = ()
) -> google.ai.generativelanguage.BatchEmbedContentsResponse

Generates multiple embeddings from the model given input text in a synchronous call.

# This snippet has been automatically generated and should be regarded as a
# code template only.
# It will require modifications to work:
# - It may require correct/in-range values for request initialization.
# - It may require specifying regional endpoints when creating the service
#   client as shown in:
#   <a href="https://googleapis.dev/python/google-api-core/latest/client_options.html">https://googleapis.dev/python/google-api-core/latest/client_options.html</a>
from google.ai import generativelanguage_v1beta

def sample_batch_embed_contents():
    # Create a client
    client = generativelanguage_v1beta.GenerativeServiceClient()

    # Initialize request argument(s)
    requests = generativelanguage_v1beta.EmbedContentRequest()
    requests.model = "model_value"

    request = generativelanguage_v1beta.BatchEmbedContentsRequest(
        model="model_value",
        requests=requests,
    )

    # Make the request
    response = client.batch_embed_contents(request=request)

    # Handle the response
    print(response)

Args
request Union[google.ai.generativelanguage.BatchEmbedContentsRequest, dict]

The request object. Batch request to get embeddings from the model for a list of prompts.

model str

Required. The model's resource name. This serves as an ID for the Model to use.

This name should match a model name returned by the ListModels method.

Format: models/{model}

This corresponds to the model field on the request instance; if request is provided, this should not be set.

requests MutableSequence[google.ai.generativelanguage.EmbedContentRequest]

Required. Embed requests for the batch. The model in each of these requests must match the model specified BatchEmbedContentsRequest.model.

This corresponds to the requests field on the request instance; if request is provided, this should not be set.

retry google.api_core.retry.Retry

Designation of what errors, if any, should be retried.

timeout float

The timeout for this request.

metadata Sequence[Tuple[str, str]]

Strings which should be sent along with the request as metadata.

Returns
google.ai.generativelanguage.BatchEmbedContentsResponse The response to a BatchEmbedContentsRequest.

count_tokens

View source

count_tokens(
    request: Optional[Union[google.ai.generativelanguage.CountTokensRequest, dict]] = None,
    *,
    model: Optional[str] = None,
    contents: Optional[MutableSequence[content.Content]] = None,
    retry: OptionalRetry = gapic_v1.method.DEFAULT,
    timeout: Union[float, object] = gapic_v1.method.DEFAULT,
    metadata: Sequence[Tuple[str, str]] = ()
) -> google.ai.generativelanguage.CountTokensResponse

Runs a model's tokenizer on input content and returns the token count.

# This snippet has been automatically generated and should be regarded as a
# code template only.
# It will require modifications to work:
# - It may require correct/in-range values for request initialization.
# - It may require specifying regional endpoints when creating the service
#   client as shown in:
#   <a href="https://googleapis.dev/python/google-api-core/latest/client_options.html">https://googleapis.dev/python/google-api-core/latest/client_options.html</a>
from google.ai import generativelanguage_v1beta

def sample_count_tokens():
    # Create a client
    client = generativelanguage_v1beta.GenerativeServiceClient()

    # Initialize request argument(s)
    request = generativelanguage_v1beta.CountTokensRequest(
        model="model_value",
    )

    # Make the request
    response = client.count_tokens(request=request)

    # Handle the response
    print(response)

Args
request Union[google.ai.generativelanguage.CountTokensRequest, dict]

The request object. Counts the number of tokens in the prompt sent to a model.

Models may tokenize text differently, so each model may return a different token_count.

model str

Required. The model's resource name. This serves as an ID for the Model to use.

This name should match a model name returned by the ListModels method.

Format: models/{model}

This corresponds to the model field on the request instance; if request is provided, this should not be set.

contents MutableSequence[google.ai.generativelanguage.Content]

Required. The input given to the model as a prompt.

This corresponds to the contents field on the request instance; if request is provided, this should not be set.

retry google.api_core.retry.Retry

Designation of what errors, if any, should be retried.

timeout float

The timeout for this request.

metadata Sequence[Tuple[str, str]]

Strings which should be sent along with the request as metadata.

Returns
google.ai.generativelanguage.CountTokensResponse A response from CountTokens.

It returns the model's token_count for the prompt.

embed_content

View source

embed_content(
    request: Optional[Union[google.ai.generativelanguage.EmbedContentRequest, dict]] = None,
    *,
    model: Optional[str] = None,
    content: Optional[google.ai.generativelanguage.Content] = None,
    retry: OptionalRetry = gapic_v1.method.DEFAULT,
    timeout: Union[float, object] = gapic_v1.method.DEFAULT,
    metadata: Sequence[Tuple[str, str]] = ()
) -> google.ai.generativelanguage.EmbedContentResponse

Generates an embedding from the model given an input Content.

# This snippet has been automatically generated and should be regarded as a
# code template only.
# It will require modifications to work:
# - It may require correct/in-range values for request initialization.
# - It may require specifying regional endpoints when creating the service
#   client as shown in:
#   <a href="https://googleapis.dev/python/google-api-core/latest/client_options.html">https://googleapis.dev/python/google-api-core/latest/client_options.html</a>
from google.ai import generativelanguage_v1beta

def sample_embed_content():
    # Create a client
    client = generativelanguage_v1beta.GenerativeServiceClient()

    # Initialize request argument(s)
    request = generativelanguage_v1beta.EmbedContentRequest(
        model="model_value",
    )

    # Make the request
    response = client.embed_content(request=request)

    # Handle the response
    print(response)

Args
request Union[google.ai.generativelanguage.EmbedContentRequest, dict]

The request object. Request containing the Content for the model to embed.

model str

Required. The model's resource name. This serves as an ID for the Model to use.

This name should match a model name returned by the ListModels method.

Format: models/{model}

This corresponds to the model field on the request instance; if request is provided, this should not be set.

content google.ai.generativelanguage.Content

Required. The content to embed. Only the parts.text fields will be counted.

This corresponds to the content field on the request instance; if request is provided, this should not be set.

retry google.api_core.retry.Retry

Designation of what errors, if any, should be retried.

timeout float

The timeout for this request.

metadata Sequence[Tuple[str, str]]

Strings which should be sent along with the request as metadata.

Returns
google.ai.generativelanguage.EmbedContentResponse The response to an EmbedContentRequest.

from_service_account_file

View source

@classmethod
from_service_account_file(
    filename: str, *args, **kwargs
)

Creates an instance of this client using the provided credentials file.

Args
filename str

The path to the service account private key json file.

args Additional arguments to pass to the constructor.
kwargs Additional arguments to pass to the constructor.

Returns
GenerativeServiceClient The constructed client.

from_service_account_info

View source

@classmethod
from_service_account_info(
    info: dict, *args, **kwargs
)

Creates an instance of this client using the provided credentials info.

Args
info dict

The service account private key info.

args Additional arguments to pass to the constructor.
kwargs Additional arguments to pass to the constructor.

Returns
GenerativeServiceClient The constructed client.

from_service_account_json

View source

@classmethod
from_service_account_json(
    filename: str, *args, **kwargs
)

Creates an instance of this client using the provided credentials file.

Args
filename str

The path to the service account private key json file.

args Additional arguments to pass to the constructor.
kwargs Additional arguments to pass to the constructor.

Returns
GenerativeServiceClient The constructed client.

generate_answer

View source

generate_answer(
    request: Optional[Union[google.ai.generativelanguage.GenerateAnswerRequest, dict]] = None,
    *,
    model: Optional[str] = None,
    contents: Optional[MutableSequence[content.Content]] = None,
    safety_settings: Optional[MutableSequence[safety.SafetySetting]] = None,
    answer_style: Optional[google.ai.generativelanguage.GenerateAnswerRequest.AnswerStyle] = None,
    retry: OptionalRetry = gapic_v1.method.DEFAULT,
    timeout: Union[float, object] = gapic_v1.method.DEFAULT,
    metadata: Sequence[Tuple[str, str]] = ()
) -> google.ai.generativelanguage.GenerateAnswerResponse

Generates a grounded answer from the model given an input GenerateAnswerRequest.

# This snippet has been automatically generated and should be regarded as a
# code template only.
# It will require modifications to work:
# - It may require correct/in-range values for request initialization.
# - It may require specifying regional endpoints when creating the service
#   client as shown in:
#   <a href="https://googleapis.dev/python/google-api-core/latest/client_options.html">https://googleapis.dev/python/google-api-core/latest/client_options.html</a>
from google.ai import generativelanguage_v1beta

def sample_generate_answer():
    # Create a client
    client = generativelanguage_v1beta.GenerativeServiceClient()

    # Initialize request argument(s)
    request = generativelanguage_v1beta.GenerateAnswerRequest(
        model="model_value",
        answer_style="VERBOSE",
    )

    # Make the request
    response = client.generate_answer(request=request)

    # Handle the response
    print(response)

Args
request Union[google.ai.generativelanguage.GenerateAnswerRequest, dict]

The request object. Request to generate a grounded answer from the model.

model str

Required. The name of the Model to use for generating the grounded response.

Format: model=models/{model}.

This corresponds to the model field on the request instance; if request is provided, this should not be set.

contents MutableSequence[google.ai.generativelanguage.Content]

Required. The content of the current conversation with the model. For single-turn queries, this is a single question to answer. For multi-turn queries, this is a repeated field that contains conversation history and the last Content in the list containing the question.

This corresponds to the contents field on the request instance; if request is provided, this should not be set.

safety_settings MutableSequence[google.ai.generativelanguage.SafetySetting]

Optional. A list of unique SafetySetting instances for blocking unsafe content.

This will be enforced on the GenerateAnswerRequest.contents and GenerateAnswerResponse.candidate. There should not be more than one setting for each SafetyCategory type. The API will block any contents and responses that fail to meet the thresholds set by these settings. This list overrides the default settings for each SafetyCategory specified in the safety_settings. If there is no SafetySetting for a given SafetyCategory provided in the list, the API will use the default safety setting for that category.

This corresponds to the safety_settings field on the request instance; if request is provided, this should not be set.

answer_style google.ai.generativelanguage.GenerateAnswerRequest.AnswerStyle

Required. Style in which answers should be returned.

This corresponds to the answer_style field on the request instance; if request is provided, this should not be set.

retry google.api_core.retry.Retry

Designation of what errors, if any, should be retried.

timeout float

The timeout for this request.

metadata Sequence[Tuple[str, str]]

Strings which should be sent along with the request as metadata.

Returns
google.ai.generativelanguage.GenerateAnswerResponse Response from the model for a grounded answer.

generate_content

View source

generate_content(
    request: Optional[Union[google.ai.generativelanguage.GenerateContentRequest, dict]] = None,
    *,
    model: Optional[str] = None,
    contents: Optional[MutableSequence[content.Content]] = None,
    retry: OptionalRetry = gapic_v1.method.DEFAULT,
    timeout: Union[float, object] = gapic_v1.method.DEFAULT,
    metadata: Sequence[Tuple[str, str]] = ()
) -> google.ai.generativelanguage.GenerateContentResponse

Generates a response from the model given an input GenerateContentRequest.

# This snippet has been automatically generated and should be regarded as a
# code template only.
# It will require modifications to work:
# - It may require correct/in-range values for request initialization.
# - It may require specifying regional endpoints when creating the service
#   client as shown in:
#   <a href="https://googleapis.dev/python/google-api-core/latest/client_options.html">https://googleapis.dev/python/google-api-core/latest/client_options.html</a>
from google.ai import generativelanguage_v1beta

def sample_generate_content():
    # Create a client
    client = generativelanguage_v1beta.GenerativeServiceClient()

    # Initialize request argument(s)
    request = generativelanguage_v1beta.GenerateContentRequest(
        model="model_value",
    )

    # Make the request
    response = client.generate_content(request=request)

    # Handle the response
    print(response)

Args
request Union[google.ai.generativelanguage.GenerateContentRequest, dict]

The request object. Request to generate a completion from the model.

model str

Required. The name of the Model to use for generating the completion.

Format: name=models/{model}.

This corresponds to the model field on the request instance; if request is provided, this should not be set.

contents MutableSequence[google.ai.generativelanguage.Content]

Required. The content of the current conversation with the model. For single-turn queries, this is a single instance. For multi-turn queries, this is a repeated field that contains conversation history + latest request.

This corresponds to the contents field on the request instance; if request is provided, this should not be set.

retry google.api_core.retry.Retry

Designation of what errors, if any, should be retried.

timeout float

The timeout for this request.

metadata Sequence[Tuple[str, str]]

Strings which should be sent along with the request as metadata.

Returns
google.ai.generativelanguage.GenerateContentResponse Response from the model supporting multiple candidates.

Note on safety ratings and content filtering. They are reported for both prompt in GenerateContentResponse.prompt_feedback and for each candidate in finish_reason and in safety_ratings. The API contract is that: - either all requested candidates are returned or no candidates at all - no candidates are returned only if there was something wrong with the prompt (see prompt_feedback) - feedback on each candidate is reported on finish_reason and safety_ratings.

get_mtls_endpoint_and_cert_source

View source

@classmethod
get_mtls_endpoint_and_cert_source(
    client_options: Optional[client_options_lib.ClientOptions] = None
)

Return the API endpoint and client cert source for mutual TLS.

The client cert source is determined in the following order: (1) if GOOGLE_API_USE_CLIENT_CERTIFICATE environment variable is not "true", the client cert source is None. (2) if client_options.client_cert_source is provided, use the provided one; if the default client cert source exists, use the default one; otherwise the client cert source is None.

The API endpoint is determined in the following order: (1) if client_options.api_endpoint if provided, use the provided one. (2) if GOOGLE_API_USE_CLIENT_CERTIFICATE environment variable is "always", use the default mTLS endpoint; if the environment variable is "never", use the default API endpoint; otherwise if client cert source exists, use the default mTLS endpoint, otherwise use the default API endpoint.

More details can be found at https://google.aip.dev/auth/4114

Args
client_options google.api_core.client_options.ClientOptions

Custom options for the client. Only the api_endpoint and client_cert_source properties may be used in this method.

Returns
Tuple[str, Callable[[], Tuple[bytes, bytes]]]: returns the API endpoint and the client cert source to use.

Raises
google.auth.exceptions.MutualTLSChannelError If any errors happen.

stream_generate_content

View source

stream_generate_content(
    request: Optional[Union[google.ai.generativelanguage.GenerateContentRequest, dict]] = None,
    *,
    model: Optional[str] = None,
    contents: Optional[MutableSequence[content.Content]] = None,
    retry: OptionalRetry = gapic_v1.method.DEFAULT,
    timeout: Union[float, object] = gapic_v1.method.DEFAULT,
    metadata: Sequence[Tuple[str, str]] = ()
) -> Iterable[google.ai.generativelanguage.GenerateContentResponse]

Generates a streamed response from the model given an input GenerateContentRequest.

# This snippet has been automatically generated and should be regarded as a
# code template only.
# It will require modifications to work:
# - It may require correct/in-range values for request initialization.
# - It may require specifying regional endpoints when creating the service
#   client as shown in:
#   <a href="https://googleapis.dev/python/google-api-core/latest/client_options.html">https://googleapis.dev/python/google-api-core/latest/client_options.html</a>
from google.ai import generativelanguage_v1beta

def sample_stream_generate_content():
    # Create a client
    client = generativelanguage_v1beta.GenerativeServiceClient()

    # Initialize request argument(s)
    request = generativelanguage_v1beta.GenerateContentRequest(
        model="model_value",
    )

    # Make the request
    stream = client.stream_generate_content(request=request)

    # Handle the response
    for response in stream:
        print(response)

Args
request Union[google.ai.generativelanguage.GenerateContentRequest, dict]

The request object. Request to generate a completion from the model.

model str

Required. The name of the Model to use for generating the completion.

Format: name=models/{model}.

This corresponds to the model field on the request instance; if request is provided, this should not be set.

contents MutableSequence[google.ai.generativelanguage.Content]

Required. The content of the current conversation with the model. For single-turn queries, this is a single instance. For multi-turn queries, this is a repeated field that contains conversation history + latest request.

This corresponds to the contents field on the request instance; if request is provided, this should not be set.

retry google.api_core.retry.Retry

Designation of what errors, if any, should be retried.

timeout float

The timeout for this request.

metadata Sequence[Tuple[str, str]]

Strings which should be sent along with the request as metadata.

Returns
Iterable[google.ai.generativelanguage.GenerateContentResponse]: Response from the model supporting multiple candidates.

Note on safety ratings and content filtering. They are reported for both prompt in GenerateContentResponse.prompt_feedback and for each candidate in finish_reason and in safety_ratings. The API contract is that: - either all requested candidates are returned or no candidates at all - no candidates are returned only if there was something wrong with the prompt (see prompt_feedback) - feedback on each candidate is reported on finish_reason and safety_ratings.

__enter__

View source

__enter__() -> 'GenerativeServiceClient'

__exit__

View source

__exit__(
    type, value, traceback
)

Releases underlying transport's resources.

.. warning:: ONLY use as a context manager if the transport is NOT shared with other clients! Exiting the with block will CLOSE the transport and may cause errors in other clients!

DEFAULT_ENDPOINT 'generativelanguage.googleapis.com'
DEFAULT_MTLS_ENDPOINT 'generativelanguage.mtls.googleapis.com'