weather

Description

The workflow takes latitude and longitude as input, retrieves weather forecast data using an open API and asks Gemini to prepare a brief summary.

Usage

Create custom images for functions:

docker build -t geminifunc gemini/
docker build -t weatherfunc weather/

Create functions and workflow (assuming CLI env. variable contains the path of Serverledge CLI executable):

$CLI create -u -f weather --memory 500 --runtime custom --custom_image weatherfunc \
    --input "gemini_api_key:Text" --input "latitude:Float" --input "longitude:Float"\
    --output "gemini_api_key:Text" \
    --output "current_temperature:Float" --output "daily_rain_sum:ArrayFloat" \
    --output "daily_max_temp:ArrayFloat" --output "daily_min_temp:ArrayFloat" 

$CLI create -u -f adapter --memory 200 --runtime python310 --handler function.handler --src adapter/function.py  \
    --input "gemini_api_key:Text" \
    --input "current_temperature:Float" --input "daily_rain_sum:ArrayFloat" \
    --input "daily_max_temp:ArrayFloat" --input "daily_min_temp:ArrayFloat" \
    --output "prompt:Text" --output "gemini_api_key:Text"

$CLI create -u -f gemini --memory 500 --runtime custom --custom_image geminifunc \
    --input "gemini_api_key:Text" --input "prompt:Text" \
    --output "response:Text"

$CLI create-workflow -s workflow.json -f weatherForecast

Invoke the workflow with an example input:

$CLI invoke-workflow -f weatherForecast -j weather/input.json > output.txt