Generative AI Application Integration Patterns by Juan Pablo Bustos Luis Lopez Soria

Author:Juan Pablo Bustos, Luis Lopez Soria
Language: eng
Format: epub, pdf
Publisher: Packt
Published: 2024-08-29T06:42:15+00:00

The interactions between these components are represented in the following diagram:

Figure 6.2: Batch summarization use case architecture diagram

Here’s a breakdown of the flow of the architecture:

Client applications are uploaded to Google Cloud Storage, triggering a cloud function as soon as the documents are in the bucket.

The cloud function reads the list of applications from Cloud Storage and publishes messages to a Pub/Sub topic, effectively creating a queue of applications to be processed.

Another cloud function, subscribed to the Pub/Sub topic, is triggered for each message tied to a given application in the queue.

This second cloud function invokes the LLM (for example, Google Gemini) hosted on Vertex AI, passing the application content as input along with a tailored prompt.

The LLM processes the application, leveraging its natural language understanding capabilities to generate a concise summary while adhering to any specified rules or guidelines.

The generated summary is then stored in a structured format (for example, BigQuery or Cloud Firestore) for further analysis, integration, and consumption.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.