OCI AI Language Service - Python SDK

Python is great!! With just few lines of code, we can implement a service that authenticates, calls an OCI service, and prints the result 🙂

import oci

signer = oci.auth.signers.InstancePrincipalsSecurityTokenSigner()

ai_language_client = oci.ai_language.AIServiceLanguageClient(config={'region': 'eu-frankfurt-1'}, signer=signer)

detect_language_sentiments_response = ai_language_client.detect_language_sentiments(
    detect_language_sentiments_details=oci.ai_language.models.DetectLanguageSentimentsDetails(
        text="I do not like PSG"), opc_request_id="ZSDZFYMMEQIJ2HXSFSD23>")

print(detect_language_sentiments_response.data)

Granted, this authentication method (Instance Principals) only works from within an OCI Compute instance. Locally we would use the .oci config file for authentication.

config = oci.config.from_file(file_location=’C:/Users/DANIMMAR/Documents/.oci/config’)

import oci

config = oci.config.from_file(file_location='C:/Users/DANIMMAR/Documents/.oci/config')

language_client = oci.ai_language.AIServiceLanguageClient(config)

detect_dominant_language_response = language_client.detect_dominant_language(
    detect_dominant_language_details=oci.ai_language.models.DetectDominantLanguageDetails(
        text="Muito bom dia. Estou a fazer um teste do serviço de inteligência artificial."), opc_request_id="ZSDZFYMMEQIJ2HXWVI7SSDFGSDFG")


print(detect_dominant_language_response.data)

The batch request supports also Sentence based analysis

Instead of analyzing aspect-based (words) sentiment, we can also have sentence-based.

batch_detect_language_sentiments_response = ai_language_client.batch_detect_language_sentiments(
    batch_detect_language_sentiments_details=oci.ai_language.models.BatchDetectLanguageSentimentsDetails(
        documents=[
            oci.ai_language.models.SentimentsDocument(
                key="01",
                text="<your text>",
                )]),
    opc_request_id="CIXQHCIX0WNU7J6Y564564",
    level=["SENTENCE"])