PredictorProjGPT

Introducción

El marco de aviso PredictorProjGPT tiene la capacidad de aceptar resultados de pruebas de rendimiento como entrada y proporcionar una salida que incluye indicadores clave de rendimiento (KPI) asignados a sus objetos de marco de datos correspondientes.

Documentación de OpenAI : https://platform.openai.com/docs/libraries/python-library

Requisitos previos

1. Clave API de OpenAI
2. Paquetes OpenAI
3. Paquetes Pandas
4. Paquetes Streamlit

Instalación de paquetes

1. Instalar paquetes OpenAI

 pip install openai

2. Instalar Pandas

 pip install pandas

3. Instale Streamlit

 pip install streamlit

Producción

Solicitud de informe de prueba de rendimiento de entrada

 A performance test was conducted on 24th April 2020 on a Salesforce application. The following observation was 
recorded by a performance tester after conclusion of the test.
Configuration:
1. There are total 40 transactions involved. The wait time between each transaction is 5 seconds.
2. Total Pacing Time is configured as 180 seconds.
    
Summary:
1. Average Response Time for the entire duration of the test was 6.5 seconds.
2. CPU Utilization peaked 75% in the first hour of the steady state.
3. Memory Consumption of 88% was observed at the last hour of the steady state.
4. Entire duration of the test was 5 hours. Endurance Test was performed.
5. Observed 10% error on the entire duration of the test
6. Predict the outcome

Resultado final


Instalación

1. Clona el repositorio en el directorio de tu proyecto.

 git clone 'enter repository url here'

2. Construir imagen

 docker build -t predictor-gpt-app .

2. Ejecute el contenedor

 docker run -it predictor-gpt-app

Actualización - 18/03/2024

 1. Recent Changes in OpenAI commands - https://github.com/openai/openai-python

 2. OpenAI secret key is now stored as environment variable