create a way to test against srj18 without committing srj18

.gitignore

@@ -37,5 +37,6 @@ hypergraph
 
 CPU.*
 results
+generated-datasets
 cosmos-export
 .vercel

package.json

@@ -11,6 +11,7 @@
     "benchmark1": "bun run --cpu-prof-md scripts/profiling/hg07-first10.ts",
     "benchmark:bus": "./benchmark-bus.sh",
     "benchmark:srj13": "./benchmark-srj13.sh",
+    "generate:srj18": "bun run scripts/benchmarking/generate-srj18.ts",
     "benchmark:section": "bun run scripts/benchmarking/hg07-section-pipeline.ts",
     "benchmark:section:profile": "bun run scripts/benchmarking/hg07-first40-section-profile.ts",
     "benchmark:port-point-pathing": "bun run scripts/benchmarking/port-point-pathing-section-pipeline.ts"
@@ -34,6 +35,8 @@
     "typescript": "^5"
   },
   "dependencies": {
-    "@tsci/seveibar.dataset-srj13": "https://github.com/tscircuit/dataset-srj13#161a835ed40ce72a2fa18ffe5ba6ba6392fcc176"
+    "@tsci/seveibar.dataset-srj13": "https://github.com/tscircuit/dataset-srj13#161a835ed40ce72a2fa18ffe5ba6ba6392fcc176",
+    "@tscircuit/capacity-autorouter": "^0.0.557",
+    "dataset-srj18": "git+https://github.com/tscircuit/dataset-srj18.git#48cd2f8f651bc9213a339580fdf125d2a1dd9254"
   }
 }

pages/dataset-srj18.page.tsx

@@ -0,0 +1,181 @@
+import type { SerializedHyperGraph } from "@tscircuit/hypergraph"
+import { type ChangeEvent, useEffect, useMemo, useState } from "react"
+import { TinyHyperGraphSectionPipelineSolver } from "lib/index"
+import { Debugger } from "./components/Debugger"
+
+type ImportGlob = <T>(
+  pattern: string,
+  options?: { import?: string },
+) => Record<string, () => Promise<T>>
+
+const SAMPLE_HASH_PARAM = "sample"
+const SAMPLE_NAMES = Array.from(
+  { length: 16 },
+  (_, index) => `sample${String(index + 1).padStart(3, "0")}`,
+)
+const srj18SampleLoaders = (
+  import.meta as ImportMeta & { glob: ImportGlob }
+).glob<SerializedHyperGraph>("../generated-datasets/srj18/*.hg.json", {
+  import: "default",
+})
+
+const clampSampleIndex = (sampleIndex: number) =>
+  Math.min(
+    Math.max(Number.isFinite(sampleIndex) ? sampleIndex : 0, 0),
+    SAMPLE_NAMES.length - 1,
+  )
+
+const getSampleIndexFromHash = () => {
+  if (typeof window === "undefined") return 0
+
+  const hashParams = new URLSearchParams(window.location.hash.slice(1))
+  const sampleNumber = Number(hashParams.get(SAMPLE_HASH_PARAM))
+
+  if (!Number.isFinite(sampleNumber)) return 0
+  return clampSampleIndex(sampleNumber - 1)
+}
+
+const setSampleIndexInHash = (sampleIndex: number) => {
+  if (typeof window === "undefined") return
+
+  const url = new URL(window.location.href)
+  const hashParams = new URLSearchParams(url.hash.slice(1))
+  hashParams.set(SAMPLE_HASH_PARAM, String(sampleIndex + 1))
+  url.hash = hashParams.toString()
+
+  window.history.replaceState(window.history.state, "", url)
+}
+
+const getSampleLoader = (sampleName: string) =>
+  srj18SampleLoaders[`../generated-datasets/srj18/${sampleName}.hg.json`]
+
+export default function DatasetSrj18Page() {
+  const [selectedSampleIndex, setSelectedSampleIndex] = useState(
+    getSampleIndexFromHash,
+  )
+  const [serializedHyperGraph, setSerializedHyperGraph] =
+    useState<SerializedHyperGraph | null>(null)
+  const [loadError, setLoadError] = useState<string | null>(null)
+
+  useEffect(() => {
+    const syncSelectedSampleFromHash = () => {
+      setSelectedSampleIndex(getSampleIndexFromHash())
+    }
+
+    window.addEventListener("hashchange", syncSelectedSampleFromHash)
+    return () =>
+      window.removeEventListener("hashchange", syncSelectedSampleFromHash)
+  }, [])
+
+  useEffect(() => {
+    setSampleIndexInHash(selectedSampleIndex)
+  }, [selectedSampleIndex])
+
+  const selectedSampleName =
+    SAMPLE_NAMES[selectedSampleIndex] ?? SAMPLE_NAMES[0]!
+  const generatedSampleNames = useMemo(
+    () =>
+      Object.keys(srj18SampleLoaders)
+        .map((samplePath) => /\/(sample\d+)\.hg\.json$/.exec(samplePath)?.[1])
+        .filter((sampleName): sampleName is string => Boolean(sampleName))
+        .sort(),
+    [],
+  )
+
+  useEffect(() => {
+    let cancelled = false
+    setSerializedHyperGraph(null)
+    setLoadError(null)
+
+    const loadSample = async () => {
+      const load = getSampleLoader(selectedSampleName)
+      if (!load) {
+        throw new Error(`Missing generated srj18 sample: ${selectedSampleName}`)
+      }
+
+      const nextSerializedHyperGraph = await load()
+      if (!cancelled) {
+        setSerializedHyperGraph(nextSerializedHyperGraph)
+      }
+    }
+
+    loadSample().catch((error) => {
+      if (!cancelled) {
+        setLoadError(error instanceof Error ? error.message : String(error))
+      }
+    })
+
+    return () => {
+      cancelled = true
+    }
+  }, [selectedSampleName])
+
+  return (
+    <div className="flex h-screen flex-col gap-3 p-3">
+      <div className="flex flex-wrap items-center gap-3 rounded border border-slate-300 bg-white p-3">
+        <label className="flex items-center gap-2 text-sm">
+          <span className="font-medium">Sample</span>
+          <input
+            className="w-24 rounded border border-slate-300 px-2 py-1"
+            type="number"
+            min={1}
+            max={SAMPLE_NAMES.length}
+            value={selectedSampleIndex + 1}
+            onChange={(event: ChangeEvent<HTMLInputElement>) => {
+              setSelectedSampleIndex(
+                clampSampleIndex(Number(event.currentTarget.value) - 1),
+              )
+            }}
+          />
+        </label>
+        <button
+          className="rounded border border-slate-300 px-3 py-1 text-sm disabled:opacity-50"
+          disabled={selectedSampleIndex === 0}
+          onClick={() =>
+            setSelectedSampleIndex((index) => Math.max(index - 1, 0))
+          }
+          type="button"
+        >
+          Previous
+        </button>
+        <button
+          className="rounded border border-slate-300 px-3 py-1 text-sm disabled:opacity-50"
+          disabled={selectedSampleIndex === SAMPLE_NAMES.length - 1}
+          onClick={() =>
+            setSelectedSampleIndex((index) =>
+              Math.min(index + 1, SAMPLE_NAMES.length - 1),
+            )
+          }
+          type="button"
+        >
+          Next
+        </button>
+        <div className="text-sm text-slate-700">
+          {selectedSampleName} • srj18 • generated {generatedSampleNames.length}
+          /{SAMPLE_NAMES.length}
+        </div>
+      </div>
+      <div className="min-h-0 flex-1">
+        {loadError ? (
+          <div className="rounded border border-amber-300 bg-amber-50 p-3 text-sm text-amber-900">
+            {loadError}
+          </div>
+        ) : serializedHyperGraph ? (
+          <Debugger
+            key={selectedSampleName}
+            serializedHyperGraph={serializedHyperGraph}
+            createSolver={(nextSerializedHyperGraph) =>
+              new TinyHyperGraphSectionPipelineSolver({
+                serializedHyperGraph: nextSerializedHyperGraph,
+              })
+            }
+          />
+        ) : (
+          <div className="rounded border border-slate-300 bg-white p-3 text-sm text-slate-700">
+            Loading {selectedSampleName}
+          </div>
+        )}
+      </div>
+    </div>
+  )
+}

scripts/benchmarking/benchmark.ts

@@ -4,7 +4,7 @@ import {
   stackGraphicsHorizontally,
   type GraphicsObject,
 } from "graphics-debug"
-import { mkdir, readdir, writeFile } from "node:fs/promises"
+import { mkdir, readFile, readdir, writeFile } from "node:fs/promises"
 import path from "node:path"
 import { loadSerializedHyperGraph } from "../../lib/compat/loadSerializedHyperGraph"
 import {
@@ -54,14 +54,16 @@ type BenchmarkSampleResult = {
 }
 
 type SolverVariant = "core" | "poly"
+type DatasetKey = "hg07" | "srj18"
 
 const IMPROVEMENT_EPSILON = 1e-9
 
 const HELP_TEXT = `Usage: ./benchmark.sh [options]
 
-Run the hg07 section-pipeline benchmark and write per-sample artifacts under ./results/runNNN/.
+Run the section-pipeline benchmark and write per-sample artifacts under ./results/runNNN/.
 
 Options:
+  --dataset NAME  Dataset to run: hg07 or 18/srj18. Defaults to hg07.
   --limit N       Run the first N samples from the dataset.
   --sample NUM    Run a specific sample by number or name (e.g. 2, 002, sample002).
   --solver NAME   Solver variant: core or poly. Defaults to core.
@@ -71,6 +73,7 @@ Options:
 
 Examples:
   ./benchmark.sh
+  ./benchmark.sh --dataset 18
   ./benchmark.sh --limit 20
   ./benchmark.sh --limit 20 --solver poly
   ./benchmark.sh --sample 2
@@ -163,6 +166,7 @@ const parseArgs = () => {
   let sampleName: string | null = null
   let candidateFamilies: TinyHyperGraphSectionCandidateFamily[] | null = null
   let solverVariant: SolverVariant = "core"
+  let datasetKey: DatasetKey = "hg07"
 
   for (let index = 0; index < process.argv.length; index += 1) {
     const arg = process.argv[index]
@@ -185,6 +189,20 @@ const parseArgs = () => {
       continue
     }
 
+    if (arg === "--dataset") {
+      const rawValue = process.argv[index + 1]
+      if (rawValue === "hg07" || rawValue === "7") {
+        datasetKey = "hg07"
+      } else if (rawValue === "18" || rawValue === "srj18") {
+        datasetKey = "srj18"
+      } else {
+        usageError(`Invalid --dataset value: ${rawValue ?? "<missing>"}`)
+      }
+
+      index += 1
+      continue
+    }
+
     if (arg === "--sample") {
       const rawValue = process.argv[index + 1]
       if (!rawValue) {
@@ -227,7 +245,7 @@ const parseArgs = () => {
     usageError("Use either --limit or --sample, not both")
   }
 
-  return { limit, sampleName, candidateFamilies, solverVariant }
+  return { limit, sampleName, candidateFamilies, solverVariant, datasetKey }
 }
 
 const formatSeconds = (durationMs: number) =>
@@ -299,7 +317,7 @@ const getNextRunDirectory = async (resultsDir: string) => {
   }
 }
 
-const loadDatasetModule = async (): Promise<DatasetModule> => {
+const loadHg07DatasetModule = async (): Promise<DatasetModule> => {
   console.log("loading dataset=hg07")
   const datasetModule = (await import("dataset-hg07")) as DatasetModule
   console.log(
@@ -308,6 +326,78 @@ const loadDatasetModule = async (): Promise<DatasetModule> => {
   return datasetModule
 }
 
+const loadSrj18DatasetModule = async (
+  cwd: string,
+  limit: number | null,
+  sampleName: string | null,
+): Promise<DatasetModule> => {
+  const {
+    ensureSrj18DatasetGenerated,
+    getSrj18DatasetDir,
+    getSrj18SampleNames,
+  } = await import("./generate-srj18")
+
+  const allSampleNames = getSrj18SampleNames()
+  if (sampleName && !allSampleNames.includes(sampleName)) {
+    usageError(`Unknown sample: ${sampleName}`)
+  }
+
+  const requestedSampleNames = sampleName
+    ? [sampleName]
+    : allSampleNames.slice(
+        0,
+        limit === null
+          ? allSampleNames.length
+          : Math.min(limit, allSampleNames.length),
+      )
+
+  await ensureSrj18DatasetGenerated(cwd, requestedSampleNames)
+
+  const datasetDir = getSrj18DatasetDir(cwd)
+  console.log(`loading dataset=srj18 dir=${datasetDir}`)
+
+  const datasetModule: DatasetModule = {
+    manifest: {
+      sampleCount: allSampleNames.length,
+      samples: allSampleNames.map((srj18SampleName) => ({
+        sampleName: srj18SampleName,
+        circuitKey: "srj18",
+        circuitId: srj18SampleName,
+        stepsToPortPointSolve: 0,
+      })),
+    },
+  }
+
+  for (const srj18SampleName of requestedSampleNames) {
+    const serializedHyperGraph = JSON.parse(
+      await readFile(
+        path.join(datasetDir, `${srj18SampleName}.hg.json`),
+        "utf8",
+      ),
+    ) as SerializedHyperGraph
+
+    datasetModule[srj18SampleName] = serializedHyperGraph
+  }
+
+  console.log(
+    `loaded dataset=srj18 samples=${datasetModule.manifest.sampleCount}`,
+  )
+  return datasetModule
+}
+
+const loadDatasetModule = async (
+  datasetKey: DatasetKey,
+  cwd: string,
+  limit: number | null,
+  sampleName: string | null,
+): Promise<DatasetModule> => {
+  if (datasetKey === "srj18") {
+    return loadSrj18DatasetModule(cwd, limit, sampleName)
+  }
+
+  return loadHg07DatasetModule()
+}
+
 const getSelectedSamples = (
   datasetModule: DatasetModule,
   limit: number | null,
@@ -360,9 +450,15 @@ const stringifyLogValue = (value: unknown) =>
   typeof value === "string" ? value : JSON.stringify(value, null, 2)
 
 const main = async () => {
-  const { limit, sampleName, candidateFamilies, solverVariant } = parseArgs()
-  const datasetModule = await loadDatasetModule()
   const cwd = process.cwd()
+  const { limit, sampleName, candidateFamilies, solverVariant, datasetKey } =
+    parseArgs()
+  const datasetModule = await loadDatasetModule(
+    datasetKey,
+    cwd,
+    limit,
+    sampleName,
+  )
   const resultsDir = path.join(cwd, "results")
   const { runName } = await getNextRunDirectory(resultsDir)
   const runDir = path.join(resultsDir, runName)
@@ -375,7 +471,7 @@ const main = async () => {
       : TinyHyperGraphSectionPipelineSolver
 
   console.log(
-    `dataset=hg07 samples=${sampleMetas.length}/${datasetModule.manifest.sampleCount} run=${runName} solver=${solverVariant} families=${candidateFamilies?.join(",") ?? "default"}`,
+    `dataset=${datasetKey} samples=${sampleMetas.length}/${datasetModule.manifest.sampleCount} run=${runName} solver=${solverVariant} families=${candidateFamilies?.join(",") ?? "default"}`,
   )
 
   for (const sampleMeta of sampleMetas) {