code-yeongyu/oh-my-openagent / packages/shared-skills/skills/ulw-research
ulw-research
UnverifiedRuns maximum-saturation research with a cooperating team, claim-graph gating, and a cited, QA'd deliverable. Use when the user explicitly asks for research or a deep investigation, including any 'ulw' research wording.
scripts/source-snapshot.test.ts
View on GitHubimport { describe, expect, onTestFinished, test } from "bun:test"
import { createHash } from "node:crypto"
import { existsSync, mkdtempSync, readFileSync, rmSync } from "node:fs"
import { tmpdir } from "node:os"
import { join } from "node:path"
import { snapshotFromHtml, writeSnapshot } from "./source-snapshot.mjs"
const FIXTURES = join(import.meta.dir, "tests", "fixtures", "fidelity", "sources")
const S4_HTML = readFileSync(join(FIXTURES, "S4.html"), "utf8")
const S9_TEXT = readFileSync(join(FIXTURES, "S9.txt"), "utf8")
function tempDir(): string {
const dir = mkdtempSync(join(tmpdir(), "ulw-snapshot-"))
onTestFinished(() => rmSync(dir, { recursive: true, force: true }))
return dir
}
type Sentence = { sid: string; start: number; end: number; text: string }
function readSentences(dir: string): Sentence[] {
return readFileSync(join(dir, "sentences.jsonl"), "utf8")
.split("\n")
.filter((line) => line !== "")
.map((line) => JSON.parse(line))
}
describe("snapshotFromHtml", () => {
test("#given a page with prose, a table, a script and a comment #when snapshotted #then blocks are lines and each table cell is its own line", () => {
const lines = snapshotFromHtml(S4_HTML).split("\n")
expect(lines).toContain("Span length: 320 m")
expect(lines).toContain("After the delay, we estimate completion no earlier than 2027.")
const header = lines.indexOf("Metric")
expect(lines.slice(header, header + 4)).toEqual(["Metric", "Length", "Span", "320 m"])
const text = lines.join("\n")
expect(text).not.toContain("not prose")
expect(text).not.toContain("internal draft comment")
expect(text).not.toContain("p { margin")
})
})
describe("writeSnapshot", () => {
test("#given html #when written #then raw.bin preserves the body and sha256 covers its bytes", () => {
const sources = tempDir()
const result = writeSnapshot(sources, "S4", {
url: "https://example.test/bridge",
html: S4_HTML,
extractor: "tool:source-snapshot@1",
observed_at: "2026-10-23T00:00:00Z",
})
expect(result.dir).toBe(join(sources, "S4"))
for (const name of ["raw.bin", "text.txt", "source.html", "meta.json", "sentences.jsonl", "acronyms.json"])
expect(existsSync(join(result.dir, name))).toBe(true)
const raw = readFileSync(join(result.dir, "raw.bin"))
expect(raw.toString("utf8")).toBe(S4_HTML)
expect(result.sha256).toBe(createHash("sha256").update(raw).digest("hex"))
expect(readFileSync(join(result.dir, "source.html"), "utf8")).toBe(S4_HTML)
const meta = JSON.parse(readFileSync(join(result.dir, "meta.json"), "utf8"))
expect(meta).toEqual({
source_id: "S4",
url: "https://example.test/bridge",
observed_at: "2026-10-23T00:00:00Z",
sha256: result.sha256,
extractor: "tool:source-snapshot@1",
revision: null,
kind: "page",
provenance: "unknown",
text_source: "html",
charset: "utf-8",
})
})
test("#given a written snapshot #when sentences.jsonl is read #then every row slices text.txt and carries an S<n>:s<k> id", () => {
const sources = tempDir()
const { dir } = writeSnapshot(sources, "S4", {
url: "https://example.test/bridge",
html: S4_HTML,
extractor: "tool:source-snapshot@1",
})
const text = readFileSync(join(dir, "text.txt"), "utf8")
const sentences = readSentences(dir)
expect(sentences.length).toBeGreaterThanOrEqual(4)
sentences.forEach((row, index) => {
expect(row.sid).toBe(`S4:s${index + 1}`)
expect(text.slice(row.start, row.end)).toBe(row.text)
})
})
test("#given both acronym forms #when written #then acronyms.json lists each expansion with its sentence and span", () => {
const sources = tempDir()
const text =
"The deck uses High Bandwidth Memory (HBM) sensors.\nA second note cites TSV (Through Silicon Via) bonding."
const { dir, sha256 } = writeSnapshot(sources, "S2", { path: "notes/memo.txt", text, extractor: "model" })
const acronyms = JSON.parse(readFileSync(join(dir, "acronyms.json"), "utf8"))
expect(acronyms.source_id).toBe("S2")
expect(acronyms.snapshot_sha256).toBe(sha256)
const entries = acronyms.entries.map(
(entry: { acronym: string; expansion: string; sentence_id: string; span: number[] }) => [
entry.acronym,
entry.expansion,
entry.sentence_id,
text.slice(entry.span[0], entry.span[1]),
],
)
expect(entries).toEqual([
["HBM", "High Bandwidth Memory", "S2:s1", "High Bandwidth Memory (HBM)"],
["TSV", "Through Silicon Via", "S2:s2", "TSV (Through Silicon Via)"],
])
expect(JSON.parse(readFileSync(join(dir, "meta.json"), "utf8")).kind).toBe("document")
expect(existsSync(join(dir, "source.html"))).toBe(false)
})
test("#given pages #when written #then pages are joined by a form feed and no sentence crosses a page", () => {
const sources = tempDir()
const { dir } = writeSnapshot(sources, "S7", {
path: "report.pdf",
pages: ["First page ends here", "second page starts here."],
extractor: "extractor:pdftotext",
})
expect(readFileSync(join(dir, "text.txt"), "utf8")).toBe("First page ends here\fsecond page starts here.")
expect(readSentences(dir).map((row) => row.text)).toEqual(["First page ends here", "second page starts here."])
})
test("#given Korean text #when written #then the sentence index covers it", () => {
const sources = tempDir()
const { dir } = writeSnapshot(sources, "S9", { path: "S9.txt", text: S9_TEXT, extractor: "tool:source-snapshot@1" })
expect(readSentences(dir).map((row) => row.sid)).toEqual(["S9:s1"])
})
test("#given an invalid extractor, source id, or a different text over an existing snapshot #when written #then it throws", () => {
const sources = tempDir()
expect(() => writeSnapshot(sources, "S1", { path: "a.txt", text: "x", extractor: "guess" })).toThrow()
expect(() => writeSnapshot(sources, "source1", { path: "a.txt", text: "x", extractor: "model" })).toThrow()
writeSnapshot(sources, "S1", { path: "a.txt", text: "first", extractor: "model" })
expect(() => writeSnapshot(sources, "S1", { path: "a.txt", text: "first", extractor: "model" })).not.toThrow()
expect(() => writeSnapshot(sources, "S1", { path: "a.txt", text: "second", extractor: "model" })).toThrow(
/snapshot_exists/,
)
})
})
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| finishing-a-development-branchobra/superpowersUse when implementation is complete, all tests pass, and you need to decide how to integrate the work | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| systematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixes | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| test-driven-developmentobra/superpowersUse when implementing any feature or bugfix, before writing implementation code | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| skill-creatoranthropics/skillsCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |