egutierrez
28e50b921c
- Added Router_de_agentes.py to manage agent interactions for sales data analysis. - Created Analizador_de_datos_de_ventas.yaml for generating structured text reports from sales data. - Developed Generador_sql_ventas.yaml for generating SQL queries to analyze sales data. - Established Router_de_agente.yaml as a routing mechanism for agent requests. - Compiled centros_disponibles.md listing available sales centers. - Introduced primera_ejecucion_de_un_agente.py as an example for executing agents. - Added ver_los_prompts_de_un_agente.py to inspect prompts sent to OpenAI. - Included service account key for BigQuery access in rag-datasets-reader-sa-key.json. - Defined schema for sales data in Objeto_ventas.json. - Implemented utility functions for querying BigQuery in conseguir_datos_bq.py. - Created data transformation utilities in transformar_datos.py for handling decimal and date formats.
68 lines
2.1 KiB
Python
68 lines
2.1 KiB
Python
from google.cloud import bigquery
|
|
from typing import Generator, List, Dict
|
|
import os
|
|
|
|
os.environ["GOOGLE_APPLICATION_CREDENTIALS"] = "rag-datasets-reader-sa-key.json"
|
|
|
|
|
|
def consultar_bigquery_paginado(
|
|
sql_query: str,
|
|
page_size: int = 1000,
|
|
project_id: str | None = None
|
|
) -> Generator[List[Dict], None, None]:
|
|
"""
|
|
Ejecuta una consulta SQL de BigQuery y devuelve los resultados paginados.
|
|
Limpia el texto SQL automáticamente antes de ejecutarlo.
|
|
|
|
Args:
|
|
sql_query (str): Texto SQL (puede incluir espacios o punto y coma al final).
|
|
page_size (int): Número de filas por página.
|
|
project_id (str | None): ID del proyecto GCP (opcional si está configurado el entorno).
|
|
|
|
Yields:
|
|
list[dict]: Página de resultados (cada fila convertida a dict).
|
|
"""
|
|
# 🧹 Limpieza del texto SQL
|
|
cleaned_query = (
|
|
sql_query.strip() # elimina espacios y saltos al inicio y fin
|
|
.removeprefix("```sql").removeprefix("```").removesuffix("```") # limpia delimitadores markdown si los hay
|
|
.strip() # vuelve a limpiar por si quedaron espacios
|
|
)
|
|
|
|
# Quita el punto y coma final si lo tiene
|
|
if cleaned_query.endswith(";"):
|
|
cleaned_query = cleaned_query[:-1].strip()
|
|
|
|
# Inicializa el cliente
|
|
client = bigquery.Client(project=project_id)
|
|
|
|
# Ejecuta la consulta
|
|
query_job = client.query(cleaned_query)
|
|
|
|
# Devuelve resultados paginados
|
|
iterator = query_job.result(page_size=page_size)
|
|
for page in iterator.pages:
|
|
yield [dict(row) for row in page]
|
|
|
|
|
|
if __name__ == "__main__":
|
|
resultado = consultar_bigquery_paginado("""```sql
|
|
SELECT
|
|
Productos___Producto_NavId__description AS producto,
|
|
Centros___Centro_NavId__name AS region,
|
|
SUM(cantidad) AS total_vendido,
|
|
SUM(cantidad * precioUnitario) AS ingreso_total
|
|
FROM
|
|
`autingo-159109.rag_datasets.Objeto_Ventas`
|
|
WHERE
|
|
Fecha >= DATE_SUB(CURRENT_DATE(), INTERVAL 3 MONTH)
|
|
GROUP BY
|
|
producto, region
|
|
ORDER BY
|
|
total_vendido DESC;
|
|
```""")
|
|
|
|
for pagina in resultado:
|
|
for fila in pagina:
|
|
print(fila)
|
|
print("----- Fin de página -----") |