transform.c File Reference

#include "transform.h"

Include dependency graph for transform.c:

Functions
static float	T_Normalise (const float value, const float min, const float max)
	Helper function for nomalising data. More...
void	T_BuildWeatherDB (T_LocationsLookup_TypeDef *locations, PGconn *psql_conn)
	Build weather table in PostgreSQL database. More...
void	T_FloodPrediction (PGconn *psql_conn)
	Transforms data from all harvest programs into flood prediction. More...
void	T_WindowDataset (PGconn *psql_conn, const int program_id)
	Window dataset into sum of next 8-days and previous 5-days. More...
void	T_NormaliseWindowedPrecipitation (PGconn *psql_conn, const int program_id)
	Normalise summed moving window precipitation between 0 and 1. More...

Function Documentation

T_BuildWeatherDB()

void T_BuildWeatherDB	(	T_LocationsLookup_TypeDef *	locations,
		PGconn *	psql_conn
	)

Build weather table in PostgreSQL database.

Main weather table for each location.

This table contains both forecasted and historical weather information for each harvest location. One column contains precipitation data where forecasted weather is added first, then overwritten using observed values. Both forecasted and observed data are stored in seperate columns.

Note

This function requires weather_ibm_eis and weather_bom tables to be populated.

Parameters

locations	Unique locations information.
psql_conn	PostgreSQL connection handler.

T_FloodPrediction()

void T_FloodPrediction

(

PGconn *

psql_conn

)

Transforms data from all harvest programs into flood prediction.

@breif Main entry point for calculating the risk of a harvest area closure.

Gets a list of the harvest areas. Loops through each of these locations to calcualte a windowed moving average of precipitation and then normalises this window to provide a value between 0 and 1.

Parameters

psql_conn

PostgreSQL connection.

T_Normalise()

static float T_Normalise ( const float  value, const float  min, const float  max  )

inlinestatic

Helper function for nomalising data.

Parameters

value	Raw input value.
min	Raw maxiumum value.
max	Raw miniumum value.

Returns

Normalised input value (between 0 and 1).

T_NormaliseWindowedPrecipitation()

void T_NormaliseWindowedPrecipitation	(	PGconn *	psql_conn,
		const int	program_id
	)

Normalise summed moving window precipitation between 0 and 1.

@breif Normalise precipiation data between 0 and 1.

After a summed window average has been completed by T_WindowDataset the data can then be normalised. This normalisation is program specific i.e. each normalisations max and min are based on that program only and not the max and min of all programs in NSW.

Parameters

psql_conn

PostgreSQL connection.

T_WindowDataset()

void T_WindowDataset	(	PGconn *	psql_conn,
		const int	program_id
	)

Window dataset into sum of next 8-days and previous 5-days.

Window transform Z-Score values into a probability of flooding event.

For each program_id calcualte the sum of the next 8-days precipitation and the previous 5-days preciptiation. This acts as a moving window across all available data.

Note

The outlook and hindlook can be adjusted by altering the w_start and w_end variables in this function.

Parameters

psql_conn	PostgreSQL connection.
program_id	Food authority program ID.