Update Shiny App to use the API once Michael deploys to Ferrlyland

[SamEdwardes]

ds-workflows-python/materials/05-shiny-app/src/module_model_explorer.py

Lines 289 to 358 in de07245

	@reactive.calc
	@time_function
	def predict_delay() -> float:
	"""
	The delay model is hosted on Posit Connect at this URL:
	https://connect.posit.it/content/823c479e-3d5e-4898-8801-a5c2cec97bb5
	"""
	# Based on the selected vessel name, get all of the data related to that
	# vessel.
	selected_vessel_data = (
	vessel_verbose.filter(pl.col("VesselName") == input.selected_vessel_name())
	.to_dicts()[0]
	)
	# Some of the vessels have not been rebuilt. When this applies, impute
	# the current year as the year rebuilt.
	if selected_vessel_data["YearRebuilt"]:
	year_rebuilt = selected_vessel_data["YearRebuilt"].year
	else:
	year_rebuilt = datetime.datetime.now().year
	# TODO: after Michael published the API to Ferryland bring this code
	# back into the fold
	prediction_input_data = {
	"Departing": get_starting_and_ending_terminal()[0],
	"Arriving": get_starting_and_ending_terminal()[1],
	"Month": input.selected_date().month,
	"Weekday": input.selected_date().weekday(),
	"Hour": input.selected_hour(),
	"ClassName": selected_vessel_data["ClassName"],
	"SpeedInKnots": selected_vessel_data["SpeedInKnots"],
	"EngineCount": selected_vessel_data["EngineCount"],
	"Horsepower": selected_vessel_data["Horsepower"],
	"MaxPassengerCount": selected_vessel_data["MaxPassengerCount"],
	"PassengerOnly": None, # selected_vessel_data["PassengerOnly"],
	"FastFerry": None, # selected_vessel_data["FastFerry"],
	"PropulsionInfo": selected_vessel_data["PropulsionInfo"],
	"YearBuilt": selected_vessel_data["YearBuilt"].year,
	"YearRebuilt": year_rebuilt,
	"departing_weather_code": int(input.selected_weather_code()),
	"departing_temperature_2m": input.selected_temperature(),
	"departing_precipitation": None, # input.selected_precipitation(),
	"departing_cloud_cover": input.selected_cloud_cover(),
	"departing_wind_speed_10m": input.selected_wind_speed(),
	"departing_wind_direction_10m": int(input.selected_wind_direction()),
	"departing_wind_gusts_10m": input.selected_wind_gust(),
	"arriving_weather_code": int(input.selected_weather_code()),
	"arriving_temperature_2m": input.selected_temperature(),
	"arriving_precipitation": None, # input.selected_precipitation(),
	"arriving_cloud_cover": input.selected_cloud_cover(),
	"arriving_wind_speed_10m": input.selected_wind_speed(),
	"arriving_wind_direction_10m": int(input.selected_wind_direction()),
	"arriving_wind_gusts_10m": input.selected_wind_gust(),
	}
	# Make the prediction
	# prediction_results_df = predict(
	# vetiver_endpoint(
	# "https://connect.posit.it/content/823c479e-3d5e-4898-8801-a5c2cec97bb5/predict"
	# ),
	# pd.DataFrame.from_records([prediction_input_data]),
	# headers={"Authorization": f'Key {os.environ["CONNECT_API_KEY"]}'},
	# )
	# prediction_results_value = prediction_results_df.iloc[0, 0]
	# return round(float(prediction_results_value), 2) # type: ignore
	# TEMPORARY - return a random number as the prediction
	return random.randint(-3, 23)