[{"_1":2,"_3":-5,"_4":-5},"loaderData",{"_5":6,"_7":8},"actionData","errors","root",{"_106":107,"_108":109},"routes/embed.x.$handle.$name.$",{"_9":10,"_11":12,"_13":14,"_15":16,"_17":18,"_19":20,"_21":-5},"handle","pagosaoutside","name","black-hat-hydrology","path","backtest.ts","project",{"_22":45,"_11":12,"_33":46,"_47":-5,"_48":42,"_49":50,"_51":50,"_52":53,"_54":-5,"_55":56,"_57":58,"_59":32,"_60":-5,"_61":-5,"_62":63,"_64":65,"_66":67,"_68":69,"_70":71,"_72":-5,"_73":74,"_75":74,"_76":77,"_78":77,"_79":77,"_80":74,"_81":74,"_82":83,"_84":85,"_86":87,"_88":42},"file",{"_22":23,"_11":14,"_24":25,"_26":-5,"_27":28,"_29":30,"_31":32,"_33":34,"_35":36,"_37":38,"_39":40,"_41":42,"_43":-5,"_13":44},"permalink","https://www.val.town/x/pagosaoutside/black-hat-hydrology/code/backtest.ts","extension","id","321d8364-b173-49cf-b9a1-8d99dc5adbff","fileId","9e5e92f6-abcb-11f1-8a7a-1607ee4eb77e","parentId","branchId","70817041-1dfa-4778-885e-ac83be02b1d0","branchVersion",268,"latest",true,"createdAt",["D",1788963239382],"version",12,"fileType","script","content","import { CONFIG } from \"./config.ts\";\nimport {\n addDays,\n firstMondayOfSeptember,\n fixedIssueAt,\n isoDay,\n mountainCalendarDate,\n mountainToUtc,\n} from \"./dates.ts\";\nimport { QUESTIONS } from \"./questions.ts\";\nimport { RIGHT_NOW_MODELS } from \"./models.ts\";\nimport type { BasinSnapshot, FlowPoint, Milestones, Snotel } from \"./types.ts\";\n\nconst average = (values: unknown[]): number | null => {\n const valid = values.map(Number).filter(Number.isFinite);\n return valid.length\n ? valid.reduce((sum, value) => sum + value, 0) / valid.length\n : null;\n};\n\nasync function fetchJson(url: string): Promise {\n const response = await fetch(url, {\n headers: { Accept: \"application/json\" },\n });\n if (!response.ok) {\n throw new Error(`${new URL(url).hostname} returned ${response.status}`);\n }\n return response.json();\n}\n\nfunction parseUsgs(payload: any): FlowPoint[] {\n const series = payload?.value?.timeSeries?.[0]?.values?.[0]?.value || [];\n return series.map((item: any) => ({\n date: new Date(item.dateTime),\n value: Number(item.value),\n }))\n .filter((item: FlowPoint) =>\n !Number.isNaN(item.date.getTime()) && Number.isFinite(item.value) &&\n item.value >= 0\n )\n .sort((a: FlowPoint, b: FlowPoint) => a.date.getTime() - b.date.getTime());\n}\n\nfunction parseSnotel(payload: any): Snotel {\n const output: Snotel = {};\n for (const station of payload || []) {\n const stationId = String(station.stationTriplet);\n output[stationId] ||= {};\n for (const series of station.data || []) {\n const code = series.stationElement?.elementCode;\n if (!code) continue;\n output[stationId][code] = (series.values || []).map((item: any) => ({\n date: new Date(`${item.date}T12:00:00Z`),\n value: Number(item.value),\n median: item.median == null ? null : Number(item.median),\n })).filter((item: any) =>\n !Number.isNaN(item.date.getTime()) && Number.isFinite(item.value)\n );\n }\n }\n return output;\n}\n\nfunction latestAtOrBefore(\n points: T[],\n date: Date,\n): T | null {\n const eligible = points.filter((point) => point.date <= date);\n return eligible.length ? eligible[eligible.length - 1] : null;\n}\n\nfunction filteredSnotel(snotel: Snotel, cutoff: Date): Snotel {\n return Object.fromEntries(\n Object.entries(snotel).map((\n [id, station],\n ) => [\n id,\n Object.fromEntries(\n Object.entries(station).map((\n [code, points],\n ) => [\n code,\n points.filter((point) => point.date <= cutoff),\n ]),\n ),\n ]),\n );\n}\n\nfunction stationAverage(\n snotel: Snotel,\n element: string,\n date: Date,\n field: \"value\" | \"median\" = \"value\",\n): number | null {\n return average(\n Object.values(snotel).map((station) =>\n (latestAtOrBefore(station[element] || [], date) as any)?.[field]\n ),\n );\n}\n\nfunction waterYearDay(date: Date): number {\n const startYear = date.getUTCMonth() >= 9\n ? date.getUTCFullYear()\n : date.getUTCFullYear() - 1;\n return Math.round(\n (Date.UTC(date.getUTCFullYear(), date.getUTCMonth(), date.getUTCDate()) -\n Date.UTC(startYear, 9, 1)) / 86_400_000,\n );\n}\n\nfunction buildFeatures(\n flowDaily: FlowPoint[],\n currentFlow: number,\n snotel: Snotel,\n now: Date,\n): Record {\n const date7 = addDays(now, -7), date30 = addDays(now, -30);\n const recent7 = flowDaily.filter((p) => p.date >= date7 && p.date <= now).map(\n (p) => p.value\n );\n const recent30 = flowDaily.filter((p) => p.date >= date30 && p.date <= now)\n .map((p) => p.value);\n const flow7Ago = latestAtOrBefore(flowDaily, date7)?.value ?? currentFlow;\n const flowMean7 = average(recent7) ?? currentFlow,\n flowMean30 = average(recent30) ?? flowMean7;\n const swe = stationAverage(snotel, \"WTEQ\", now) ?? 0,\n sweMedian = stationAverage(snotel, \"WTEQ\", now, \"median\") ?? 0;\n const swe7Ago = stationAverage(snotel, \"WTEQ\", date7) ?? swe;\n const sweRatio = sweMedian > 0.05\n ? Math.min(3, Math.max(0, swe / sweMedian))\n : swe <= 0.05\n ? 0\n : 3;\n const stationIds = Object.keys(snotel), temperatures: number[] = [];\n for (const id of stationIds) {\n for (const point of snotel[id].TAVG || []) {\n if (point.date >= date7 && point.date <= now) {\n temperatures.push(point.value);\n }\n }\n }\n const temp7 = average(temperatures) ?? 32;\n const precipitationChange = (start: Date) =>\n average(stationIds.map((id) => {\n const current = latestAtOrBefore(snotel[id].PREC || [], now)?.value;\n const earlier = latestAtOrBefore(snotel[id].PREC || [], start)?.value;\n return Number.isFinite(current) && Number.isFinite(earlier)\n ? Math.max(0, Number(current) - Number(earlier))\n : null;\n })) ?? 0;\n const precip7 = precipitationChange(date7),\n precip30 = precipitationChange(date30),\n snowPresent = swe > 0.05 ? 1 : 0;\n const angle = 2 * Math.PI * waterYearDay(now) / 365.25;\n return {\n log_flow: Math.log1p(currentFlow),\n log_flow_mean_7d: Math.log1p(flowMean7),\n log_flow_mean_30d: Math.log1p(flowMean30),\n relative_flow_change_7d: (currentFlow - flow7Ago) /\n (Math.abs(flowMean7) + 30),\n swe_in: swe,\n swe_ratio: sweRatio,\n swe_change_7d: swe - swe7Ago,\n temperature_scaled: (temp7 - 32) / 20,\n prec_total_7d: precip7,\n prec_total_30d: precip30,\n snow_present: snowPresent,\n season_sin: Math.sin(angle),\n season_cos: Math.cos(angle),\n season_sin_2: Math.sin(2 * angle),\n season_cos_2: Math.cos(2 * angle),\n melt_pressure: swe * Math.max(temp7 - 32, 0) / 100,\n rain_without_snow: precip7 * (1 - snowPresent),\n current_flow: currentFlow,\n swe_median: sweMedian,\n temp_7d: temp7,\n precip_7d: precip7,\n };\n}\n\nfunction findTubingOpening(\n series: FlowPoint[],\n after: Date | null,\n): Date | null {\n if (!after) return null;\n const ordered = series.filter((item) => item.date > after).sort((a, b) =>\n a.date.getTime() - b.date.getTime()\n );\n let left = 0, outOfRange = 0;\n for (let index = 0; index < ordered.length; index += 1) {\n const endPoint = ordered[index],\n startTime = endPoint.date.getTime() - 24 * 3_600_000;\n if (endPoint.value < 30 || endPoint.value >= 400) outOfRange += 1;\n while (left <= index && ordered[left].date.getTime() < startTime) {\n if (ordered[left].value < 30 || ordered[left].value >= 400) {\n outOfRange -= 1;\n }\n left += 1;\n }\n if (\n index - left + 1 >= 4 &&\n endPoint.date.getTime() - ordered[left].date.getTime() >=\n 23 * 3_600_000 &&\n outOfRange === 0\n ) return endPoint.date;\n }\n return null;\n}\n\nfunction deriveMilestones(\n dailyFlow: FlowPoint[],\n instantaneous: FlowPoint[],\n snotel: Snotel,\n current: FlowPoint,\n): Milestones {\n const runoffYear = current.date.getUTCMonth() >= 9\n ? current.date.getUTCFullYear() + 1\n : current.date.getUTCFullYear();\n const waterYearStart = new Date(Date.UTC(runoffYear - 1, 9, 1, 12)),\n springStart = new Date(Date.UTC(runoffYear, 2, 1, 12));\n const peakWindowStart = new Date(Date.UTC(runoffYear, 3, 1, 12)),\n peakWindowEnd = new Date(Date.UTC(runoffYear, 6, 31, 23, 59));\n const recreationStart = new Date(Date.UTC(runoffYear, 4, 1, 12)),\n laborDay = firstMondayOfSeptember(runoffYear);\n laborDay.setUTCHours(23, 59, 59, 999);\n const waterYearDaily = dailyFlow.filter((p) =>\n p.date >= waterYearStart && p.date <= current.date\n ),\n springDaily = waterYearDaily.filter((p) => p.date >= springStart);\n const first400 = springDaily.find((p) => p.value >= 400) || null;\n const peakDaily = springDaily.filter((p) =>\n p.date >= peakWindowStart && p.date <= peakWindowEnd\n ).reduce(\n (best, p) => !best || p.value > best.value ? p : best,\n null,\n );\n const peakInstant = instantaneous.filter((p) =>\n p.date >= peakWindowStart && p.date <= peakWindowEnd &&\n p.date <= current.date\n ).reduce(\n (best, p) => !best || p.value > best.value ? p : best,\n null,\n );\n const daysSincePeak = peakDaily\n ? Math.floor(\n (current.date.getTime() - peakDaily.date.getTime()) / 86_400_000,\n )\n : 0;\n const dropFromPeak = peakDaily?.value\n ? 1 - current.value / peakDaily.value\n : 0;\n const peakFlowLikelyPassed = Boolean(\n peakDaily && daysSincePeak >= 6 && dropFromPeak >= 0.18,\n );\n let tubingForecastTrigger: FlowPoint | null = null;\n if (peakDaily) {\n const peakIndex = springDaily.findIndex((p) =>\n isoDay(p.date) === isoDay(peakDaily.date) && p.value === peakDaily.value\n );\n for (\n let index = Math.max(2, peakIndex + 1);\n index < springDaily.length;\n index += 1\n ) {\n const item = springDaily[index],\n fallingThreeDay = item.value < springDaily[index - 2].value;\n const eligible = peakDaily.value >= 800\n ? item.value < 800\n : peakFlowLikelyPassed;\n if (eligible && fallingThreeDay) {\n tubingForecastTrigger = item;\n break;\n }\n }\n }\n const dailyTubingFallback = peakDaily\n ? springDaily.find((p) =>\n p.date > peakDaily.date && p.value >= 30 && p.value < 400\n )?.date || null\n : null;\n const tubingOpening =\n findTubingOpening(instantaneous, peakDaily?.date || null) ||\n dailyTubingFallback;\n const seasonDaily = waterYearDaily.filter((p) =>\n p.date >= recreationStart && p.date <= laborDay && p.date <= current.date\n );\n const counts = {\n p300: seasonDaily.filter((p) => p.value >= 300).length,\n p600: seasonDaily.filter((p) => p.value >= 600).length,\n p800: seasonDaily.filter((p) => p.value >= 800).length,\n tubing: seasonDaily.filter((p) => p.value >= 30 && p.value < 400).length,\n };\n const strength = average(\n waterYearDaily.filter((p) =>\n p.date >= recreationStart &&\n p.date <= new Date(Date.UTC(runoffYear, 5, 30, 23, 59))\n ).map((p) => p.value),\n );\n const snowDates = [\n ...new Set(\n Object.values(snotel).flatMap((station) =>\n (station.WTEQ || []).map((p) => isoDay(p.date))\n ),\n ),\n ].map((day) => new Date(`${day}T12:00:00Z`)).filter((date) =>\n date >= waterYearStart && date <= current.date\n ).sort((a, b) => +a - +b);\n const snowSeries = snowDates.map((date) => ({\n date,\n value: stationAverage(snotel, \"WTEQ\", date),\n })).filter((p): p is { date: Date; value: number } =>\n Number.isFinite(p.value)\n );\n let persistentSnow = null;\n for (let i = 0; i <= snowSeries.length - 7; i += 1) {\n const window = snowSeries.slice(i, i + 7);\n if (window[0].value >= 1 && window.every((p) => p.value >= 0.5)) {\n persistentSnow = window[0];\n break;\n }\n }\n const peakSwe = snowSeries.reduce<{ date: Date; value: number } | null>(\n (best, p) => !best || p.value > best.value ? p : best,\n null,\n );\n const daysSincePeakSwe = peakSwe\n ? Math.floor(\n (current.date.getTime() - peakSwe.date.getTime()) / 86_400_000,\n )\n : 0,\n currentSwe = snowSeries.at(-1)?.value || 0;\n const peakSweLikelyPassed = Boolean(\n peakSwe && peakSwe.value >= 1 && daysSincePeakSwe >= 14 &&\n currentSwe <= peakSwe.value * 0.7,\n );\n return {\n runoffYear,\n laborDay,\n first400,\n peakDaily,\n peakInstant,\n peakFlowLikelyPassed,\n tubingForecastTrigger,\n tubingOpening,\n counts,\n strength,\n persistentSnow,\n peakSwe,\n peakSweLikelyPassed,\n };\n}\n\ntype YearData = { daily: FlowPoint[]; instant: FlowPoint[]; snotel: Snotel };\n\nfunction runTreeForBacktest(tree: any, values: number[]): number {\n let node = 0;\n while (Number(tree.feature[node]) >= 0) {\n const feature = Number(tree.feature[node]);\n node = values[feature] <= Number(tree.threshold[node])\n ? Number(tree.left[node])\n : Number(tree.right[node]);\n }\n return Number(tree.value[node]);\n}\n\nfunction calculateForBacktest(question: any, snapshot: BasinSnapshot) {\n if (question.number !== 3) return question.calculate(snapshot);\n const model: any = (RIGHT_NOW_MODELS as any).first400;\n const values = model.features.map((feature: string) =>\n Number(snapshot.features[feature])\n );\n if (values.some((value: number) => !Number.isFinite(value))) {\n throw new Error(\"First 400 backtest features are incomplete\");\n }\n const estimate =\n model.trees.reduce(\n (sum: number, tree: any) => sum + runTreeForBacktest(tree, values),\n 0,\n ) / model.trees.length;\n const days = Math.round(Math.min(80, Math.max(0, estimate)));\n return {\n date: addDays(mountainCalendarDate(snapshot.cutoffAt), days).toISOString(),\n };\n}\n\nfunction snapshotAt(data: YearData, cutoffAt: Date): BasinSnapshot {\n const dailyFlow = data.daily.filter((p) => p.date <= cutoffAt),\n instantaneous = data.instant.filter((p) => p.date <= cutoffAt),\n snotel = filteredSnotel(data.snotel, cutoffAt);\n const current = instantaneous.at(-1) || dailyFlow.at(-1);\n if (!current || !dailyFlow.length) {\n throw new Error(`No flow data at ${cutoffAt.toISOString()}`);\n }\n return {\n capturedAt: cutoffAt,\n cutoffAt,\n current,\n instantaneous,\n dailyFlow,\n snotel,\n features: buildFeatures(dailyFlow, current.value, snotel, current.date),\n milestones: deriveMilestones(dailyFlow, instantaneous, snotel, current),\n sourceStatus: {},\n };\n}\n\nasync function fetchYear(runoffYear: number): Promise {\n const start = `${runoffYear - 1}-08-15`, end = `${runoffYear}-09-12`;\n const site = CONFIG.usgsSite,\n stations = encodeURIComponent(CONFIG.snotelStations.join(\",\"));\n const dailyUrl =\n `https://waterservices.usgs.gov/nwis/dv/?format=json&sites=${site}&startDT=${start}&endDT=${end}¶meterCd=00060&statCd=00003&siteStatus=all`;\n const instantUrl =\n `https://waterservices.usgs.gov/nwis/iv/?format=json&sites=${site}&startDT=${start}&endDT=${end}¶meterCd=00060&siteStatus=all`;\n const snowUrl =\n `https://wcc.sc.egov.usda.gov/awdbRestApi/services/v1/data?stationTriplets=${stations}&elements=WTEQ%2CSNWD%2CPREC%2CTAVG&duration=DAILY&beginDate=${start}&endDate=${end}&periodRef=END¢ralTendencyType=MEDIAN&returnFlags=false&returnOriginalValues=false&returnSuspectData=false`;\n const [daily, instant, snow] = await Promise.all([\n fetchJson(dailyUrl),\n fetchJson(instantUrl),\n fetchJson(snowUrl),\n ]);\n return {\n daily: parseUsgs(daily),\n instant: parseUsgs(instant),\n snotel: parseSnotel(snow),\n };\n}\n\nasync function replayYear(runoffYear: number) {\n const data = await fetchYear(runoffYear),\n finalSnapshot = snapshotAt(data, mountainToUtc(runoffYear, 9, 10, 12));\n const eventSnapshots = new Map();\n // Event locks are day-scale rules. One noon-Mountain snapshot per day preserves\n // the historical cutoff while avoiding thousands of redundant full rebuilds.\n for (\n let t = mountainToUtc(runoffYear, 4, 1, 12).getTime();\n t <= mountainToUtc(runoffYear, 8, 1, 12).getTime();\n t += 24 * 3_600_000\n ) {\n const snapshot = snapshotAt(data, new Date(t));\n for (const q of QUESTIONS.filter((item) => item.issueKind === \"event\")) {\n if (!eventSnapshots.has(q.number) && q.trigger?.(snapshot)) {\n eventSnapshots.set(q.number, snapshotAt(data, q.trigger!(snapshot)!));\n }\n }\n }\n const cards = QUESTIONS.map((q) => {\n const issue = q.issueKind === \"fixed\"\n ? snapshotAt(data, fixedIssueAt(runoffYear, q.issueMonth!, q.issueDay!))\n : eventSnapshots.get(q.number);\n if (!issue) {\n return { number: q.number, title: q.title, status: \"not-issued\" };\n }\n const prediction = calculateForBacktest(q, issue),\n actual = q.actual(finalSnapshot);\n return {\n number: q.number,\n title: q.title,\n issue_at: issue.cutoffAt.toISOString(),\n prediction,\n actual,\n result: actual\n ? (q.score(prediction, actual) ? \"HIT\" : \"MISS\")\n : \"NO-ACTUAL\",\n };\n });\n const hits = cards.filter((c: any) => c.result === \"HIT\").length,\n misses = cards.filter((c: any) => c.result === \"MISS\").length;\n return {\n water_year: runoffYear,\n score: `Black Hat ${hits} – Mother Nature ${misses}`,\n hits,\n misses,\n cards,\n };\n}\n\nexport async function runHistoricalBacktest() {\n const results = [];\n for (const year of [2015, 2016, 2017]) results.push(await replayYear(year));\n console.log(\n \"BACKTEST_SUMMARY\",\n JSON.stringify(\n results.map(({ water_year, score, hits, misses }) => ({\n water_year,\n score,\n hits,\n misses,\n })),\n ),\n );\n for (const result of results) {\n for (const card of result.cards) {\n console.log(\n `CARD_${result.water_year}_${card.number}`,\n JSON.stringify(card),\n );\n }\n }\n}\n\nexport default runHistoricalBacktest;\n\n// Val Town's Run button evaluates the module; execute the read-only replay at module load.\nawait runHistoricalBacktest();","deleted",false,"label",[14],"f84d0eeb-ceea-4772-8eb3-fa7a8cedbc13",["D",1788718960405],"deletedAt","legacyEnvVars","privacy","public","httpPrivacy","description","San Juan River seasonal prediction tracker","imageUrl","userId","66be13f3-dd08-4f8a-aca1-8697006c9a46","pinnedFileId","0210c8ee-aa20-11f1-a516-1607ee4eb77e","pinPreview","pinMigrated","screenshotVersion","likeCount",0,"orgId","01a077f2-9ad7-764c-8f6b-583a36c43414","tursoName","f84d0eebceea47728eb3fa7a8cedbc13","tursoHostname","f84d0eebceea47728eb3fa7a8cedbc13-val-town.aws-us-east-2.turso.io","tursoMeasuredAt",["D",1789488693443],"pinned","userPerm","r","valPerm","blobPerm","rw","emailPerm","sqlitePerm","projectPerm","telemetryPerm","user",{"_9":10,"_105":101},"org",{"_9":10,"_98":99,"_100":101,"_55":56,"_102":103,"_104":42},"branch",{"_35":30,"_11":89,"_22":28,"_90":45,"_91":-5,"_92":-5,"_93":-5,"_94":-5,"_95":-5,"_96":12,"_97":10},"isOwned","main","projectId","forkedBranchId","forkedBranchVersion","forkedProjectPrivacy","forkedProjectId","forkedProjectOrgId","projectName","userHandle","displayName","Pagosa Outside","avatar","https://lh3.googleusercontent.com/a/ACg8ocLkPUw3N28PFDWb5so2_Azzb6wOyGiti-gB56AMhHTn6R7IXJA6Ug=s96-c","tier","free","canRestrictVals","profileImageUrl","ctx",{"_82":-5,"_139":140,"_141":142,"_143":144,"_145":42,"_146":42,"_147":42,"_148":32,"_149":-5,"_150":-7,"_151":-7,"_152":63,"_153":42,"_154":42,"_155":-5,"_156":42,"_157":42,"_158":42,"_159":-5},"ENV",{"_110":111,"_112":113,"_114":89,"_115":116,"_117":118,"_119":120,"_121":122,"_123":124,"_125":126,"_127":128,"_129":130,"_131":132,"_133":134,"_135":136,"_137":138},"REACT_APP_SENTRY_DSN","https://586020d86f05489db63a554690acf1e9@o1398359.ingest.sentry.io/6724499","RENDER_GIT_COMMIT","e87eb04e707cf64e7e532c8e8ec395f209e17f15","RENDER_GIT_BRANCH","POSTHOG_API_TOKEN","phc_nmoyuAkStEZuUp69rYlg1DZgrIt19Gx0Y4Hwf8tGgJP","POSTHOG_API_HOST","https://wart.val.town","WEB_API_TEMPLATE","https://{val}.web.val.run","APP_URL","https://www.val.town","MODULE_URL","https://esm.town","VANITY_WEB_API_TEMPLATE","https://{label}.val.run/","DENOLS_URL","https://lsp.val.town","NODE_ENV","production","CLOUDFLARE_TURNSTILE_SITE_KEY","0x4AAAAAACtaIZJH7s9rBMZ6","GOOGLE_CLIENT_ID","875989696899-jtsqkhlbdv5jrlsk8631p7ipjg3c6o62.apps.googleusercontent.com","STRIPE_PUBLISHABLE_KEY","pk_live_51N2zvWEpZTTLipIkVEwXrUG9AT0yeOOEYxszarZT1jpsZV58Ly6wlg3SltZpHN26CUalOBmUSY7adQRQiSkCZz2500qSXNkf8Y","API_URL","https://api.val.town","orgs",[],"preferencesKeybindings","default","preferencesTheme","light","preferencesFixers","preferencesCopilot","preferencesSelectionAi","preferencesCloseBrackets","preferencesChatPanelState","preferencesDprint","preferencesDenofmt","notificationCount","flagDenols","isBusinessOrgMember","businessOrgHandle","isBusinessOrgAdmin","isHigherTownieTier","navCollapsed","currentOrgId"]