feat: refactor

internal/dataset/dataset.go

@@ -0,0 +1,158 @@
+package dataset
+
+import "fmt"
+
+// Dataset shape constants.
+// Shape: (65, 47, 3, 361, 720) = (hour, pressure_level, variable, latitude, longitude)
+// This matches the reference predictor exactly.
+const (
+	NumHours      = 65  // 0, 3, 6, ..., 192
+	NumLevels     = 47  // pressure levels
+	NumVariables  = 3   // height, wind_u, wind_v
+	NumLatitudes  = 361 // -90.0 to +90.0 in 0.5 degree steps
+	NumLongitudes = 720 // 0.0 to 359.5 in 0.5 degree steps
+
+	HourStep   = 3     // hours between forecast time steps
+	MaxHour    = 192   // maximum forecast hour
+	Resolution = 0.5   // grid resolution in degrees
+	LatStart   = -90.0 // first latitude in the dataset
+	LonStart   = 0.0   // first longitude in the dataset
+
+	// Variable indices within the dataset.
+	VarHeight = 0
+	VarWindU  = 1
+	VarWindV  = 2
+
+	ElementSize = 4 // float32 = 4 bytes
+)
+
+// DatasetSize is the total size of the dataset file in bytes.
+// 65 * 47 * 3 * 361 * 720 * 4 = 9,528,667,200
+const DatasetSize int64 = int64(NumHours) * int64(NumLevels) * int64(NumVariables) *
+	int64(NumLatitudes) * int64(NumLongitudes) * int64(ElementSize)
+
+// LevelSet identifies which GRIB file set a pressure level belongs to.
+type LevelSet int
+
+const (
+	LevelSetA LevelSet = iota // pgrb2 (primary)
+	LevelSetB                 // pgrb2b (secondary)
+)
+
+// Pressures contains the 47 pressure levels in hPa, sorted descending.
+// Index 0 = 1000 hPa (near surface), Index 46 = 1 hPa (high atmosphere).
+var Pressures = [NumLevels]int{
+	1000, 975, 950, 925, 900, 875, 850, 825, 800, 775,
+	750, 725, 700, 675, 650, 625, 600, 575, 550, 525,
+	500, 475, 450, 425, 400, 375, 350, 325, 300, 275,
+	250, 225, 200, 175, 150, 125, 100, 70, 50, 30,
+	20, 10, 7, 5, 3, 2, 1,
+}
+
+// pressureIndex maps pressure in hPa to its index in the Pressures array.
+var pressureIndex map[int]int
+
+// pressureLevelSet maps pressure in hPa to its GRIB file set.
+var pressureLevelSet map[int]LevelSet
+
+func init() {
+	pressureIndex = make(map[int]int, NumLevels)
+	for i, p := range Pressures {
+		pressureIndex[p] = i
+	}
+
+	pressureLevelSet = make(map[int]LevelSet, NumLevels)
+	for _, p := range PressuresPgrb2 {
+		pressureLevelSet[p] = LevelSetA
+	}
+	for _, p := range PressuresPgrb2b {
+		pressureLevelSet[p] = LevelSetB
+	}
+}
+
+// PressuresPgrb2 contains levels found in the primary pgrb2 file (26 levels).
+var PressuresPgrb2 = []int{
+	10, 20, 30, 50, 70, 100, 150, 200, 250, 300, 350, 400,
+	450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 925,
+	950, 975, 1000,
+}
+
+// PressuresPgrb2b contains levels found in the secondary pgrb2b file (21 levels).
+var PressuresPgrb2b = []int{
+	1, 2, 3, 5, 7, 125, 175, 225, 275, 325, 375, 425,
+	475, 525, 575, 625, 675, 725, 775, 825, 875,
+}
+
+// PressureIndex returns the dataset index for a given pressure level in hPa.
+// Returns -1 if the level is not found.
+func PressureIndex(hPa int) int {
+	idx, ok := pressureIndex[hPa]
+	if !ok {
+		return -1
+	}
+	return idx
+}
+
+// PressureLevelSet returns which GRIB file set a pressure level belongs to.
+func PressureLevelSet(hPa int) (LevelSet, bool) {
+	ls, ok := pressureLevelSet[hPa]
+	return ls, ok
+}
+
+// HourIndex returns the dataset time index for a forecast hour.
+// Returns -1 if the hour is invalid (not a multiple of HourStep or out of range).
+func HourIndex(hour int) int {
+	if hour < 0 || hour > MaxHour || hour%HourStep != 0 {
+		return -1
+	}
+	return hour / HourStep
+}
+
+// Hours returns all forecast hours as a slice: [0, 3, 6, ..., 192].
+func Hours() []int {
+	out := make([]int, 0, NumHours)
+	for h := 0; h <= MaxHour; h += HourStep {
+		out = append(out, h)
+	}
+	return out
+}
+
+// S3 URL configuration for NOAA GFS data.
+const S3BaseURL = "https://noaa-gfs-bdp-pds.s3.amazonaws.com"
+
+// GribURL returns the S3 URL for a primary (pgrb2) GRIB file.
+func GribURL(date string, runHour, forecastStep int) string {
+	return fmt.Sprintf("%s/gfs.%s/%02d/atmos/gfs.t%02dz.pgrb2.0p50.f%03d",
+		S3BaseURL, date, runHour, runHour, forecastStep)
+}
+
+// GribURLB returns the S3 URL for a secondary (pgrb2b) GRIB file.
+func GribURLB(date string, runHour, forecastStep int) string {
+	return fmt.Sprintf("%s/gfs.%s/%02d/atmos/gfs.t%02dz.pgrb2b.0p50.f%03d",
+		S3BaseURL, date, runHour, runHour, forecastStep)
+}
+
+// GribFileName returns the local filename for a primary GRIB file.
+func GribFileName(runHour, forecastStep int) string {
+	return fmt.Sprintf("gfs.t%02dz.pgrb2.0p50.f%03d", runHour, forecastStep)
+}
+
+// GribFileNameB returns the local filename for a secondary GRIB file.
+func GribFileNameB(runHour, forecastStep int) string {
+	return fmt.Sprintf("gfs.t%02dz.pgrb2b.0p50.f%03d", runHour, forecastStep)
+}
+
+// VariableIndex returns the dataset variable index for a GRIB parameter.
+// Returns -1 if the parameter is not recognized.
+func VariableIndex(parameterCategory, parameterNumber int) int {
+	switch {
+	case parameterCategory == 3 && parameterNumber == 5:
+		return VarHeight // Geopotential Height
+	case parameterCategory == 2 && parameterNumber == 2:
+		return VarWindU // U-component of wind
+	case parameterCategory == 2 && parameterNumber == 3:
+		return VarWindV // V-component of wind
+	default:
+		return -1
+	}
+}