import get from 'lodash-es/get.js'
import each from 'lodash-es/each.js'
import size from 'lodash-es/size.js'
import isNumber from 'lodash-es/isNumber.js'
import isestr from 'wsemi/src/isestr.mjs'
import iseobj from 'wsemi/src/iseobj.mjs'
import isbol from 'wsemi/src/isbol.mjs'
import isearr from 'wsemi/src/isearr.mjs'
import isnum from 'wsemi/src/isnum.mjs'
import cdbl from 'wsemi/src/cdbl.mjs'
import haskey from 'wsemi/src/haskey.mjs'
import ptsXYtoArr from './ptsXYtoArr.mjs'
import ptsXYZtoArr from './ptsXYZtoArr.mjs'
import interp2Normalize from './interp2Normalize.mjs'
import kriging from './kriging.mjs'
//https://github.com/lvisei/web-developer-resources/blob/master/webassembly/kriging.md
/**
* 克利金法(Kriging)內外插點數值
*
* Unit Test: {@link https://github.com/yuda-lyu/w-gis/blob/master/test/interp2Kriging.test.mjs Github}
* @memberOf w-gis
* @param {Array} psSrc 輸入二維座標加觀測數據點陣列，為[{x:x1,y:y1,z:z1},{x:x2,y:y2,z:z2},...]點物件之陣列
* @param {Array|Object} psTar 輸入二維座標點陣列或點物件，為[{x:x1,y:y1},{x:x2,y:y2},...]點物件之陣列，或{x:x1,y:y1}點物件
* @param {Object} [opt={}] 輸入設定物件，預設{}
* @param {String} [opt.keyX='x'] 輸入點物件之x欄位字串，為座標，預設'x'
* @param {String} [opt.keyY='y'] 輸入點物件之y欄位字串，為座標，預設'y'
* @param {String} [opt.keyZ='z'] 輸入點物件之z欄位字串，為觀測值，預設'z'
* @param {Number} [opt.scale=1] 輸入正規化範圍數值，因polybooljs處理多邊形時有數值容許誤差，故須通過縮放值域來減少問題，預設1是正規化0至1之間，使用scaleXY則是正規化為0至scaleXY之間，預設1
* @param {String} [opt.model='exponential'] 輸入擬合模式字串，可選'exponential'、'gaussian'、'spherical'，預設'exponential'
* @param {Number} [opt.sigma2=0] 輸入自動擬合參數sigma2數值，預設0
* @param {Number} [opt.alpha=100] 輸入自動擬合參數alpha數值，預設100
* @param {Boolean} [opt.returnWithVariogram=false] 輸入是否回傳擬合半變異數結果布林值，預設false
* @returns {Array|Object} 回傳點物件陣列或點物件，若使用returnWithVariogram=true則回傳物件資訊，若發生錯誤則回傳錯誤訊息物件
* @example
*
* let ps
* let p
* let r
*
* ps = [{ x: 243, y: 206, z: 95 }, { x: 233, y: 225, z: 146 }, { x: 21, y: 325, z: 22 }, { x: 953, y: 28, z: 223 }, { x: 1092, y: 290, z: 39 }, { x: 744, y: 200, z: 191 }, { x: 174, y: 3, z: 22 }, { x: 537, y: 368, z: 249 }, { x: 1151, y: 371, z: 86 }, { x: 814, y: 252, z: 125 }]
* p = {
* x: 243,
* y: 205,
* }
* r = interp2Kriging(ps, p)
* console.log(r)
* // => { x: 243, y: 205, z: 94.88479948418721 }
*
* ps = [{ x: 243, y: 206, z: 95 }, { x: 233, y: 225, z: 146 }, { x: 21, y: 325, z: 22 }, { x: 953, y: 28, z: 223 }, { x: 1092, y: 290, z: 39 }, { x: 744, y: 200, z: 191 }, { x: 174, y: 3, z: 22 }, { x: 537, y: 368, z: 249 }, { x: 1151, y: 371, z: 86 }, { x: 814, y: 252, z: 125 }]
* p = {
* x: 283,
* y: 205,
* }
* r = interp2Kriging(ps, p)
* console.log(r)
* // => { x: 283, y: 205, z: 116.32333499687805 }
*
* ps = [{ x: 243, y: 206, z: 95 }, { x: 233, y: 225, z: 146 }, { x: 21, y: 325, z: 22 }, { x: 953, y: 28, z: 223 }, { x: 1092, y: 290, z: 39 }, { x: 744, y: 200, z: 191 }, { x: 174, y: 3, z: 22 }, { x: 537, y: 368, z: 249 }, { x: 1151, y: 371, z: 86 }, { x: 814, y: 252, z: 125 }]
* p = {
* x: 1160,
* y: 380,
* }
* r = interp2Kriging(ps, p)
* console.log(r)
* // => { x: 1160, y: 380, z: 87.27045807621836 }
*
* ps = [{ a: 243, b: 206, c: 95 }, { a: 233, b: 225, c: 146 }, { a: 21, b: 325, c: 22 }, { a: 953, b: 28, c: 223 }, { a: 1092, b: 290, c: 39 }, { a: 744, b: 200, c: 191 }, { a: 174, b: 3, c: 22 }, { a: 537, b: 368, c: 249 }, { a: 1151, b: 371, c: 86 }, { a: 814, b: 252, c: 125 }]
* p = {
* a: 243,
* b: 205,
* }
* r = interp2Kriging(ps, p, { keyX: 'a', keyY: 'b', keyZ: 'c' })
* console.log(r)
* // => { a: 243, b: 205, c: 94.88479948418721 }
*
* ps = [{ x: 243, y: 206, z: 95 }, { x: 233, y: 225, z: 146 }, { x: 21, y: 325, z: 22 }, { x: 953, y: 28, z: 223 }, { x: 1092, y: 290, z: 39 }, { x: 744, y: 200, z: 191 }, { x: 174, y: 3, z: 22 }, { x: 537, y: 368, z: 249 }, { x: 1151, y: 371, z: 86 }, { x: 814, y: 252, z: 125 }]
* p = [
* {
* x: 243,
* y: 205,
* },
* {
* x: 283,
* y: 205,
* },
* ]
* r = interp2Kriging(ps, p)
* console.log(r)
* // => [
* // { x: 243, y: 205, z: 94.88479948418721 },
* // { x: 283, y: 205, z: 116.32333499687805 }
* // ]
*
* ps = [{ x: 243, y: 206, z: 95 }, { x: 233, y: 225, z: 146 }, { x: 21, y: 325, z: 22 }, { x: 953, y: 28, z: 223 }, { x: 1092, y: 290, z: 39 }, { x: 744, y: 200, z: 191 }, { x: 174, y: 3, z: 22 }, { x: 537, y: 368, z: 249 }, { x: 1151, y: 371, z: 86 }, { x: 814, y: 252, z: 125 }]
* p = {
* x: 243,
* y: 205,
* }
* r = interp2Kriging(ps, p, { scale: 1000 })
* console.log(r)
* // => { x: 243, y: 205, z: 94.88479948418878 }
*
* ps = [{ x: 243, y: 206, z: 95 }, { x: 233, y: 225, z: 146 }, { x: 21, y: 325, z: 22 }, { x: 953, y: 28, z: 223 }, { x: 1092, y: 290, z: 39 }, { x: 744, y: 200, z: 191 }, { x: 174, y: 3, z: 22 }, { x: 537, y: 368, z: 249 }, { x: 1151, y: 371, z: 86 }, { x: 814, y: 252, z: 125 }]
* p = {
* x: 243,
* y: 205,
* }
* r = interp2Kriging(ps, p, { model: 'gaussian' })
* console.log(r)
* // => { x: 243, y: 205, z: 92.39124139470005 }
*
* ps = [{ x: 243, y: 206, z: 95 }, { x: 233, y: 225, z: 146 }, { x: 21, y: 325, z: 22 }, { x: 953, y: 28, z: 223 }, { x: 1092, y: 290, z: 39 }, { x: 744, y: 200, z: 191 }, { x: 174, y: 3, z: 22 }, { x: 537, y: 368, z: 249 }, { x: 1151, y: 371, z: 86 }, { x: 814, y: 252, z: 125 }]
* p = {
* x: 243,
* y: 205,
* }
* r = interp2Kriging(ps, p, { sigma2: 0.001, alpha: 70 })
* console.log(r)
* // => { x: 243, y: 205, z: 90.88702949276343 }
*
* ps = [{ x: 243, y: 206, z: 95 }, { x: 233, y: 225, z: 146 }, { x: 21, y: 325, z: 22 }, { x: 953, y: 28, z: 223 }, { x: 1092, y: 290, z: 39 }, { x: 744, y: 200, z: 191 }, { x: 174, y: 3, z: 22 }, { x: 537, y: 368, z: 249 }, { x: 1151, y: 371, z: 86 }, { x: 814, y: 252, z: 125 }]
* p = {
* x: 243,
* y: 205,
* }
* r = interp2Kriging(ps, p, { returnWithVariogram: true })
* console.log(r)
* // => {
* // result: { x: 243, y: 205, z: 94.88479948418721 },
* // variogram: {
* // t: [
* // 0.32158590308370044,
* // 0.5462555066079295,
* // 0,
* // 0.8854625550660793,
* // 0.07488986784140969,
* // 0.7444933920704846,
* // 0,
* // 1,
* // 0.28193832599118945,
* // 0.45374449339207046
* // ],
* // x: [
* // 0.19646017699115045,
* // 0.18761061946902655,
* // 0,
* // 0.8247787610619469,
* // 0.947787610619469,
* // 0.6398230088495576,
* // 0.13539823008849558,
* // 0.45663716814159294,
* // 1,
* // 0.7017699115044248
* // ],
* // y: [
* // 0.5168141592920354,
* // 0.5336283185840708,
* // 0.6221238938053097,
* // 0.35929203539823007,
* // 0.5911504424778761,
* // 0.511504424778761,
* // 0.33716814159292036,
* // 0.6601769911504425,
* // 0.6628318584070797,
* // 0.5575221238938053
* // ],
* // nugget: 0.33528534923428144,
* // range: 0.9818274204120488,
* // sill: 0.4584580333443075,
* // A: 0.3333333333333333,
* // n: 10,
* // model: [Function: kriging_variogram_exponential],
* // svpd: { data: [Array], bars: [Array] },
* // K: [
* // -75.8517004175245, 67.925722923484, -0.16335841036964677,
* // 0.24131518978270877, -0.04234246865048732, 0.8770518396328912,
* // 5.616960362170279, 1.621659637440227, 0.027888001651894635,
* // -0.21415853416652803, 67.92572292348402, -75.61786478824868,
* // 5.413232079727064, 0.04906943438385958, 0.03679113229348465,
* // -0.11847139352698677, 0.3142012767297917, 1.9531078854430024,
* // 0.14216377854055343, -0.026503104672525174, -0.163358410369651,
* // 5.413232079727067, -10.341088741602475, 0.707690256415205,
* // 0.2717259477481375, -0.032751338327818, 2.4021510467734406,
* // 1.1039529303204958, 0.9343241199626711, 0.17593910677199553,
* // 0.24131518978268823, 0.04906943438388453, 0.7076902564152006,
* // -11.195895809332818, 3.6318326767168267, 3.0935386508887035,
* // 1.0451553499568136, 0.07681115159025198, 0.5500520981253416,
* // 2.176928207710048, -0.04234246865049055, 0.03679113229348578,
* // 0.27172594774813813, 3.6318326767168188, -21.833991833397526,
* // -0.3703544567031784, 0.1631940797086814, 0.1478485285302121,
* // 14.592661153646349, 3.529640759660949, 0.877051839632884,
* // -0.11847139352697111, -0.032751338327819686, 3.0935386508886986,
* // -0.37035445670318073, -23.428385057803695, 0.8346909075984952,
* // 3.683800415485778, -0.07094528659968152, 15.630834782735722,
* // 5.616960362170238, 0.31420127672983256, 2.4021510467734397,
* // 1.045155349956815, 0.16319407970868527, 0.8346909075984923,
* // -11.15967643574728, 0.47980505727623907, 0.7525119639869473,
* // -0.0399020216261384, 1.6216596374402739, 1.9531078854429533,
* // 1.1039529303204962, 0.07681115159024962, 0.1478485285302013,
* // 3.6838004154857935, 0.47980505727624, -11.594555930753835,
* // 0.9237284214140226, 1.8947859772223308, 0.027888001651917377,
* // 0.14216377854053472, 0.9343241199626732, 0.5500520981253476,
* // 14.592661153646347, -0.0709452865996735, 0.7525119639869486,
* // 0.9237284214140111, -18.251230245653318, 0.837959693391809,
* // -0.21415853416654693, -0.026503104672517874, 0.1759391067719989,
* // 2.176928207710053, 3.529640759660952, 15.630834782735704,
* // -0.03990202162613744, 1.894785977222343, 0.8379596933918148,
* // -23.89204017145281
* // ],
* // M: [
* // 15.107775528173802,
* // -17.52355316402437,
* // 4.974279767668829,
* // -6.014370838628359,
* // 7.175095294789402,
* // -3.7572657339807183,
* // 4.210920521157788,
* // -6.064323143863013,
* // -2.2280724452952487,
* // 5.035785385549117
* // ]
* // }
* // }
*
*/
function interp2Kriging(psSrc, psTar, opt = {}) {
//check psSrc
if (!isearr(psSrc)) {
return {
err: 'psSrc is not an array'
}
}
//check psTar
if (!iseobj(psTar) && !isearr(psTar)) {
return {
err: 'psTar is not an object or array'
}
}
//isOne
let isOne = iseobj(psTar)
if (isOne) {
psTar = [psTar]
}
//keyX
let keyX = get(opt, 'keyX')
if (!isestr(keyX)) {
keyX = 'x'
}
//keyY
let keyY = get(opt, 'keyY')
if (!isestr(keyY)) {
keyY = 'y'
}
//keyZ
let keyZ = get(opt, 'keyZ')
if (!isestr(keyZ)) {
keyZ = 'z'
}
//scale
let scale = get(opt, 'scale')
if (!isnum(scale)) {
scale = 1
}
scale = cdbl(scale)
//krigingModel
let krigingModel = get(opt, 'model')
if (krigingModel !== 'exponential' && krigingModel !== 'gaussian' && krigingModel !== 'spherical') {
krigingModel = 'exponential'
}
//krigingSigma2
let krigingSigma2 = get(opt, 'sigma2')
if (!isnum(krigingSigma2)) {
krigingSigma2 = 0
}
krigingSigma2 = cdbl(krigingSigma2)
//krigingAlpha
let krigingAlpha = get(opt, 'alpha')
if (!isnum(krigingAlpha)) {
krigingAlpha = 100
}
krigingAlpha = cdbl(krigingAlpha)
//returnWithVariogram
let returnWithVariogram = get(opt, 'returnWithVariogram')
if (!isbol(returnWithVariogram)) {
returnWithVariogram = false
}
// console.log('returnWithVariogram', returnWithVariogram)
//ptsXYZtoArr
psSrc = ptsXYZtoArr(psSrc, opt)
//check psSrc
if (size(psSrc) === 0) {
return {
err: 'psSrc has no effective data'
}
}
// console.log('ptsXYZtoArr psSrc', psSrc)
//ptsXYtoArr
psTar = ptsXYtoArr(psTar, opt)
//check psTar
if (size(psTar) === 0) {
return {
err: 'psTar has no effective data'
}
}
// console.log('ptsXYtoArr psTar', psTar)
//interp2Normalize
let itnm = interp2Normalize(psSrc, { scale })
psSrc = itnm.ps //複寫正規化數據
let nv = itnm.nv
let inv = itnm.inv
//kpKnown
let kpKnown = {}
each(psSrc, (v) => {
let k = `${v.x}-${v.y}` //已經被正規化至x,y
kpKnown[k] = v.z //已經被正規化至x,y,z
})
//x, y, t
let x = []
let y = []
let t = []
each(psSrc, (v) => {
x.push(v.x)
y.push(v.y)
t.push(v.z)
})
//variogram
let variogram = kriging.train(
t,
x,
y,
krigingModel,
krigingSigma2,
krigingAlpha
)
// console.log('variogram', variogram)
//kpdt
let kpdt = (nx, ny) => {
//check
let k = `${nx}-${ny}` //已經被正規化至x,y
if (haskey(kpKnown, k)) {
return kpKnown[k]
}
//predict
let nz = kriging.predict(nx, ny, variogram)
//save
kpKnown[k] = nz
return nz
}
//itv
let itv = (p) => {
//x, nx
let x = p.x //已經被正規化至x,y
let nx = nv(x, 0)
if (!isNumber(nx)) {
throw new Error(`invalid nx[${nx}]`) //x,y預期都是有值
}
//y, ny
let y = p.y //已經被正規化至x,y
let ny = nv(y, 1)
if (!isNumber(ny)) {
throw new Error(`invalid ny[${ny}]`) //x,y預期都是有值
}
//nz, z
let nz = kpdt(nx, ny)
let z = inv(nz, 2)
if (!isNumber(z)) {
z = null //z須支援可能無法內插
}
// console.log('x', x, 'y', y, 'z', z)
// console.log('nx', nx, 'ny', ny, 'nz', nz)
//r
let r = {
[keyX]: x,
[keyY]: y,
[keyZ]: z,
}
return r
}
//r
let r
if (isOne) {
let p = psTar[0]
r = itv(p)
}
else {
r = []
each(psTar, (p) => {
r.push(itv(p))
})
}
//returnWithVariogram
if (returnWithVariogram) {
r = {
result: r,
variogram,
}
}
return r
}
export default interp2Kriging