get_alio_annexes

Instructions

Implementation Reference

• src/tools/alio/annexes.ts:38-105 (handler)

  Main handler function for get_alio_annexes. Loads the institution index, finds the matching regulation, reads its markdown body, extracts [별표 N] (annex) sections using regex, filters by annex number if specified, and returns either a list or full annex content.

  export async function getAlioAnnexes(
    _api: LawApiClient,
    input: GetAlioAnnexesInput
  ): Promise<ToolResponse> {
    try {
      const idx = await loadIndex()
      const inst = findInstitution(idx, input.institution)
      if (!inst) {
        return {
          content: [{ type: "text", text: `기관을 찾을 수 없습니다: ${input.institution}` }],
          isError: true,
        }
      }
      const manifest = idx.manifests.get(inst.apbaId)
      const entry = manifest?.regulations.find((r) =>
        input.regId ? r.regId === input.regId : input.title ? r.title.includes(input.title) : false
      )
      if (!entry) {
        return {
          content: [
            { type: "text", text: `규정을 찾을 수 없습니다. institution=${inst.apbaId}` },
          ],
          isError: true,
        }
      }
  
      const md = await readRegulationMd(inst.apbaId, entry.regId)
      if (!md) {
        return {
          content: [{ type: "text", text: `본문 파일을 읽을 수 없습니다: ${entry.mdPath}` }],
          isError: true,
        }
      }
  
      const annexes = extractAnnexes(md)
      const filtered = input.annexNumber
        ? annexes.filter((a) => a.number === input.annexNumber)
        : annexes
  
      const lines: string[] = []
      lines.push(`# [${inst.apbaId}] ${inst.apbaNa} — ${entry.title}`)
      lines.push(`> 별표 ${annexes.length}개 발견${input.annexNumber ? ` (#${input.annexNumber}만 표시)` : ""}`)
      lines.push("")
  
      if (filtered.length === 0) {
        lines.push("- 별표가 없거나 해당 번호의 별표가 없습니다.")
        if (annexes.length > 0) {
          lines.push("")
          lines.push("발견된 별표 목록:")
          annexes.forEach((a) => lines.push(`- ${a.heading}`))
        }
      } else if (input.listOnly) {
        filtered.forEach((a) => lines.push(`- ${a.heading}`))
      } else {
        for (const a of filtered) {
          lines.push("---")
          lines.push(`## ${a.heading}`)
          lines.push("")
          lines.push(a.body.trim())
          lines.push("")
        }
      }
  
      return { content: [{ type: "text", text: truncateResponse(lines.join("\n")) }] }
    } catch (err) {
      return formatToolError(err, "get_alio_annexes")
    }
  }

• src/tools/alio/annexes.ts:14-28 (schema)

  Zod schema defining the input: institution (required), regId or title (one required), optional annexNumber, and listOnly flag (defaults false) with a refine validator to ensure regId or title is provided.

  export const GetAlioAnnexesSchema = z
    .object({
      institution: z.string().describe("기관코드 또는 기관명"),
      regId: z.string().optional().describe("규정 ID"),
      title: z.string().optional().describe("규정 제목 부분일치 (regId 대신)"),
      annexNumber: z
        .number()
        .optional()
        .describe("특정 별표 번호만 (예: 1 → [별표 1] 만). 생략 시 전체 별표"),
      listOnly: z.boolean().default(false).describe("본문 없이 별표 목록만"),
    })
    .refine((v) => !!(v.regId || v.title), {
      message: "regId 또는 title 중 하나는 필수",
      path: ["regId"],
    })

• src/tool-registry.ts:764-769 (registration)

  Registration entry in the tool registry: imports getAlioAnnexes and GetAlioAnnexesSchema from src/tools/alio/annexes.ts, registers the tool with name 'get_alio_annexes', description, schema, and handler.

  {
    name: "get_alio_annexes",
    description: "[ALIO] 규정 본문에서 [별표 N] 섹션만 추출. annexNumber 로 특정 별표 한정 가능. listOnly 로 목록만 조회.",
    schema: GetAlioAnnexesSchema,
    handler: getAlioAnnexes
  },

• src/tools/alio/annexes.ts:108-135 (helper)

  Helper function extractAnnexes that parses the markdown body using a regex to find [별표 N] headings and splits the document into annex sections with number, heading, and body.

  function extractAnnexes(md: string): AnnexSection[] {
    // 헤딩 패턴: "## [별표 1]" 또는 "[별표 1] 제목" 또는 "[별표1]" / "[별 표 1]"
    const headRe = /^#{0,6}\s*\[\s*별\s*표\s*(\d+)?\s*\]([^\n]*)$/gm
    const matches: Array<{ idx: number; num: number | null; head: string }> = []
    let m: RegExpExecArray | null
    while ((m = headRe.exec(md)) !== null) {
      matches.push({
        idx: m.index,
        num: m[1] ? Number(m[1]) : null,
        head: m[0].replace(/^#+\s*/, "").trim(),
      })
    }
    if (matches.length === 0) return []
  
    const sections: AnnexSection[] = []
    for (let i = 0; i < matches.length; i++) {
      const cur = matches[i]
      const next = matches[i + 1]
      const start = cur.idx + matches[i].head.length // heading 직후
      const end = next ? next.idx : md.length
      sections.push({
        number: cur.num,
        heading: cur.head,
        body: md.slice(start, end).trim(),
      })
    }
    return sections
  }