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/sentence-split.test.ts
View on GitHubimport { describe, expect, test } from "bun:test"
import { splitSentences } from "./sentence-split.mjs"
type Sentence = { sid: string; start: number; end: number; text: string }
function at(records: Sentence[], index: number): Sentence {
const record = records[index]
if (record === undefined) throw new Error(`no sentence at index ${index}`)
return record
}
function expectSlices(text: string, records: Sentence[]): void {
for (const record of records) expect(text.slice(record.start, record.end)).toBe(record.text)
}
describe("splitSentences", () => {
test("#given a Korean paragraph with da/yo/nida endings #when split #then sids and offsets match each sentence", () => {
// given
const text = "매출은 100억 원입니다. 비용도 늘었어요. 이익은 줄었다.다만 회복이 기대된다."
// when
const records = splitSentences(text) as Sentence[]
// then
expect(records.map((r) => r.sid)).toEqual(["s1", "s2", "s3", "s4"])
expect(records.map((r) => r.text)).toEqual([
"매출은 100억 원입니다.",
"비용도 늘었어요.",
"이익은 줄었다.",
"다만 회복이 기대된다.",
])
expect(at(records, 1).start).toBe(text.indexOf("비용도"))
expectSlices(text, records)
})
test("#given an English paragraph with abbreviations and decimals #when split #then e.g. and 3.2 do not end a sentence", () => {
// given
const text = "Use units, e.g. 3.2 m per bar. The next one follows! Is it fine? Yes."
// when
const records = splitSentences(text) as Sentence[]
// then
expect(records.map((r) => r.sid)).toEqual(["s1", "s2", "s3", "s4"])
expect(records.map((r) => r.text)).toEqual([
"Use units, e.g. 3.2 m per bar.",
"The next one follows!",
"Is it fine?",
"Yes.",
])
expect(at(records, 3).start).toBe(text.indexOf("Yes."))
expect(at(records, 3).end).toBe(text.length)
expectSlices(text, records)
})
test("#given line breaks and blank lines #when split #then each non-empty line yields its own record", () => {
// given
const text = "매출은 100억 원이다. 다만 아직 확정되지 않았다.\n\n Second line without stop\nThird?"
// when
const records = splitSentences(text) as Sentence[]
// then
expect(records.map((r) => r.text)).toEqual([
"매출은 100억 원이다.",
"다만 아직 확정되지 않았다.",
"Second line without stop",
"Third?",
])
expect(records.map((r) => r.sid)).toEqual(["s1", "s2", "s3", "s4"])
expectSlices(text, records)
})
test("#given Korean quoted speech closed right after a da ending #when split #then the quote stays inside one sentence", () => {
// given
const single = '그는 "매출이 늘었다."라고 말했다.'
const pair = '그는 "매출이 늘었다."라고 말했다. 다음 날 떨어졌다.'
const bracket = "그는 「좋아요.」라고 답했다."
// when
const singleRecords = splitSentences(single) as Sentence[]
const pairRecords = splitSentences(pair) as Sentence[]
const bracketRecords = splitSentences(bracket) as Sentence[]
// then
expect(singleRecords.map((r) => r.text)).toEqual([single])
expect(pairRecords.map((r) => r.text)).toEqual(['그는 "매출이 늘었다."라고 말했다.', "다음 날 떨어졌다."])
expect(pairRecords.map((r) => r.sid)).toEqual(["s1", "s2"])
expect(at(pairRecords, 0).start).toBe(0)
expect(at(pairRecords, 1).start).toBe(pair.indexOf("다음 날"))
expect(at(pairRecords, 1).end).toBe(pair.length)
expect(bracketRecords.map((r) => r.text)).toEqual([bracket])
expectSlices(pair, pairRecords)
})
test("#given the word no #when split #then it is an abbreviation only before a number", () => {
// given
const answer = "The answer was no. Then we left."
const numbered = "See No. 5 for details. Then stop."
// when
const answerRecords = splitSentences(answer) as Sentence[]
const numberedRecords = splitSentences(numbered) as Sentence[]
// then
expect(answerRecords.map((r) => r.text)).toEqual(["The answer was no.", "Then we left."])
expect(at(answerRecords, 1).start).toBe(answer.indexOf("Then"))
expect(numberedRecords.map((r) => r.text)).toEqual(["See No. 5 for details.", "Then stop."])
expectSlices(answer, answerRecords)
expectSlices(numbered, numberedRecords)
})
test("#given empty or whitespace-only text #when split #then no records are produced", () => {
// when / then
expect(splitSentences("")).toEqual([])
expect(splitSentences(" \n\t\n")).toEqual([])
})
})
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 |