mapcomplete/scripts/generateStats.ts

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

230 lines
8.7 KiB
TypeScript
Raw Normal View History

import known_layers from "../src/assets/generated/known_layers.json"
import { LayerConfigJson } from "../src/Models/ThemeConfig/Json/LayerConfigJson"
import { TagUtils } from "../src/Logic/Tags/TagUtils"
import { Utils } from "../src/Utils"
2024-05-13 17:21:40 +02:00
import { existsSync, readFileSync, writeFileSync } from "fs"
2022-03-13 01:27:19 +01:00
import ScriptUtils from "./ScriptUtils"
import TagRenderingConfig from "../src/Models/ThemeConfig/TagRenderingConfig"
import { And } from "../src/Logic/Tags/And"
2024-05-13 17:21:40 +02:00
import Script from "./Script"
import NameSuggestionIndex from "../src/Logic/Web/NameSuggestionIndex"
2024-06-01 12:48:22 +02:00
import TagInfo from "../src/Logic/Web/TagInfo"
2022-03-13 01:27:19 +01:00
2024-05-13 17:21:40 +02:00
class Utilities {
2024-06-16 16:06:26 +02:00
static mapValues<X extends string | number, T, TOut>(
record: Record<X, T>,
f: (t: T) => TOut
): Record<X, TOut> {
2024-05-13 17:21:40 +02:00
const newR = <Record<X, TOut>>{}
for (const x in record) {
newR[x] = f(record[x])
}
return newR
}
}
2024-05-23 04:42:26 +02:00
2024-05-13 17:21:40 +02:00
class GenerateStats extends Script {
async createOptimizationFile(includeTags = true) {
ScriptUtils.fixUtils()
const layers = <LayerConfigJson[]>known_layers.layers
2022-03-13 01:27:19 +01:00
2024-05-13 17:21:40 +02:00
const keysAndTags = new Map<string, Set<string>>()
2022-03-13 01:27:19 +01:00
2024-05-13 17:21:40 +02:00
for (const layer of layers) {
if (layer.source["geoJson"] !== undefined && !layer.source["isOsmCache"]) {
continue
}
if (layer.source == null || typeof layer.source === "string") {
continue
}
2022-03-13 01:27:19 +01:00
2024-05-13 17:21:40 +02:00
const sourcesList = [TagUtils.Tag(layer.source["osmTags"])]
if (layer?.title) {
sourcesList.push(...new TagRenderingConfig(layer.title).usedTags())
}
2024-05-13 17:21:40 +02:00
const sources = new And(sourcesList)
const allKeys = sources.usedKeys()
for (const key of allKeys) {
if (!keysAndTags.has(key)) {
keysAndTags.set(key, new Set<string>())
}
2022-03-13 01:27:19 +01:00
}
2024-05-13 17:21:40 +02:00
const allTags = includeTags ? sources.usedTags() : []
for (const tag of allTags) {
if (!keysAndTags.has(tag.key)) {
keysAndTags.set(tag.key, new Set<string>())
}
keysAndTags.get(tag.key).add(tag.value)
2022-03-13 01:27:19 +01:00
}
}
2024-05-13 17:21:40 +02:00
const keyTotal = new Map<string, number>()
const tagTotal = new Map<string, Map<string, number>>()
await Promise.all(
Array.from(keysAndTags.keys()).map(async (key) => {
const values = keysAndTags.get(key)
const data = await TagInfo.global.getStats(key)
2024-05-13 17:21:40 +02:00
const count = data.data.find((item) => item.type === "all").count
keyTotal.set(key, count)
console.log(key, "-->", count)
if (values.size > 0) {
tagTotal.set(key, new Map<string, number>())
await Promise.all(
Array.from(values).map(async (value) => {
2024-06-16 16:06:26 +02:00
const tagData = await TagInfo.global.getStats(key, value)
const count = tagData.data.find((item) => item.type === "all").count
tagTotal.get(key).set(value, count)
console.log(key + "=" + value, "-->", count)
})
2024-05-13 17:21:40 +02:00
)
}
2024-06-16 16:06:26 +02:00
})
)
2024-05-13 17:21:40 +02:00
writeFileSync(
"./src/assets/key_totals.json",
JSON.stringify(
{
"#": "Generated with generateStats.ts",
date: new Date().toISOString(),
keys: Utils.MapToObj(keyTotal, (t) => t),
2024-06-16 16:06:26 +02:00
tags: Utils.MapToObj(tagTotal, (v) => Utils.MapToObj(v, (t) => t)),
2024-05-13 17:21:40 +02:00
},
null,
" "
2022-03-13 01:27:19 +01:00
)
2024-05-13 17:21:40 +02:00
)
}
2022-03-13 01:27:19 +01:00
private summarizeNSI(sourcefile: string, pathNoExtension: string): void {
2024-06-16 16:06:26 +02:00
const data = <Record<string, Record<string, number>>>(
JSON.parse(readFileSync(sourcefile, "utf8"))
)
const allCountries: Set<string> = new Set()
for (const brand in data) {
const perCountry = data[brand]
for (const country in perCountry) {
allCountries.add(country)
const count = perCountry[country]
if (count === 0) {
delete perCountry[country]
}
}
}
const pathOut = pathNoExtension + ".summarized.json"
2024-06-16 16:06:26 +02:00
writeFileSync(pathOut, JSON.stringify(data, null, " "), "utf8")
console.log("Written", pathOut)
const allBrands = Object.keys(data)
allBrands.sort()
for (const country of allCountries) {
const summary = <Record<string, number>>{}
for (const brand of allBrands) {
const count = data[brand][country]
2024-06-16 16:06:26 +02:00
if (count > 2) {
// Eéntje is geentje
// We ignore count == 1 as they are rather exceptional
summary[brand] = data[brand][country]
}
}
const countryPath = pathNoExtension + "." + country + ".json"
writeFileSync(countryPath, JSON.stringify(summary), "utf8")
console.log("Written", countryPath)
}
}
2024-05-23 04:42:26 +02:00
async createNameSuggestionIndexFile(basepath: string, type: "brand" | "operator" | string) {
const path = basepath + type + ".json"
2024-05-13 17:21:40 +02:00
let allBrands = <Record<string, Record<string, number>>>{}
if (existsSync(path)) {
allBrands = JSON.parse(readFileSync(path, "utf8"))
2024-06-16 16:06:26 +02:00
console.log(
"Loaded",
Object.keys(allBrands).length,
" previously loaded " + type,
"from",
path
)
2024-05-13 17:21:40 +02:00
}
let skipped = 0
2024-06-16 16:06:26 +02:00
const allBrandNames: string[] = Utils.Dedup(
NameSuggestionIndex.allPossible(type).map((item) => item.tags[type])
)
const missingBrandNames: string[] = []
2024-05-23 04:42:26 +02:00
for (let i = 0; i < allBrandNames.length; i++) {
const brand = allBrandNames[i]
2024-05-23 04:42:26 +02:00
if (!!allBrands[brand] && Object.keys(allBrands[brand]).length == 0) {
delete allBrands[brand]
console.log("Deleted", brand, "as no entries at all")
}
2024-05-23 04:42:26 +02:00
if (allBrands[brand] !== undefined) {
const max = Math.max(...Object.values(allBrands[brand]))
skipped++
if (skipped % 100 == 0) {
console.warn("Busy; ", i + "/" + allBrandNames.length, "; skipped", skipped)
}
2024-05-23 04:42:26 +02:00
if (max < 0) {
console.log("HMMMM:", allBrands[brand])
delete allBrands[brand]
2024-05-23 04:42:26 +02:00
} else {
continue
}
2022-09-08 21:40:48 +02:00
}
missingBrandNames.push(brand)
}
const batchSize = 101
for (let i = 0; i < missingBrandNames.length; i += batchSize) {
2024-06-16 16:06:26 +02:00
console.warn(
"Downloading",
batchSize,
"items: ",
i + "/" + missingBrandNames.length,
"; skipped",
skipped,
"total:",
allBrandNames.length
)
2024-06-16 16:06:26 +02:00
const distributions = await Promise.all(
Utils.TimesT(batchSize, async (j) => {
await ScriptUtils.sleep(j * 250)
return TagInfo.getGlobalDistributionsFor(type, missingBrandNames[i + j])
})
)
for (let j = 0; j < distributions.length; j++) {
const brand = missingBrandNames[i + j]
2024-06-16 16:06:26 +02:00
const distribution: Record<string, number> = Utilities.mapValues(
distributions[j],
(s) => s.data.find((t) => t.type === "all").count
)
allBrands[brand] = distribution
2024-05-13 17:21:40 +02:00
}
writeFileSync(path, JSON.stringify(allBrands), "utf8")
console.log("Checkpointed", path)
2024-05-13 17:21:40 +02:00
}
writeFileSync(path, JSON.stringify(allBrands), "utf8")
}
constructor() {
2024-06-16 16:06:26 +02:00
super(
"Downloads stats on osmSource-tags and keys from tagInfo. There are two usecases with separate outputs:\n 1. To optimize the query before sending it to overpass (generates ./src/assets/key_totals.json) \n 2. To amend the Name Suggestion Index "
)
2024-05-13 17:21:40 +02:00
}
async main(_: string[]) {
const basepath = "./src/assets/generated/stats/"
2024-05-23 11:38:29 +02:00
await this.createOptimizationFile()
for (const type of ["operator", "brand"]) {
2024-05-23 04:42:26 +02:00
await this.createNameSuggestionIndexFile(basepath, type)
this.summarizeNSI(basepath + type + ".json", "./public/assets/data/nsi/stats/" + type)
}
2024-05-13 17:21:40 +02:00
}
2022-03-13 01:27:19 +01:00
}
2024-05-13 17:21:40 +02:00
new GenerateStats().run()