clawdbot/src/discord/chunk.test.ts

import { describe, expect, it } from "vitest";

import { chunkDiscordText } from "./chunk.js";

function countLines(text: string) {
  return text.split("\n").length;
}

function hasBalancedFences(chunk: string) {
  let open: { markerChar: string; markerLen: number } | null = null;
  for (const line of chunk.split("\n")) {
    const match = line.match(/^( {0,3})(`{3,}|~{3,})(.*)$/);
    if (!match) continue;
    const marker = match[2];
    if (!open) {
      open = { markerChar: marker[0], markerLen: marker.length };
      continue;
    }
    if (open.markerChar === marker[0] && marker.length >= open.markerLen) {
      open = null;
    }
  }
  return open === null;
}

describe("chunkDiscordText", () => {
  it("splits tall messages even when under 2000 chars", () => {
    const text = Array.from({ length: 45 }, (_, i) => `line-${i + 1}`).join("\n");
    expect(text.length).toBeLessThan(2000);

    const chunks = chunkDiscordText(text, { maxChars: 2000, maxLines: 20 });
    expect(chunks.length).toBeGreaterThan(1);
    for (const chunk of chunks) {
      expect(countLines(chunk)).toBeLessThanOrEqual(20);
    }
  });

  it("keeps fenced code blocks balanced across chunks", () => {
    const body = Array.from({ length: 30 }, (_, i) => `console.log(${i});`).join("\n");
    const text = `Here is code:\n\n\`\`\`js\n${body}\n\`\`\`\n\nDone.`;

    const chunks = chunkDiscordText(text, { maxChars: 2000, maxLines: 10 });
    expect(chunks.length).toBeGreaterThan(1);

    for (const chunk of chunks) {
      expect(hasBalancedFences(chunk)).toBe(true);
      expect(chunk.length).toBeLessThanOrEqual(2000);
    }

    expect(chunks[0]).toContain("```js");
    expect(chunks.at(-1)).toContain("Done.");
  });

  it("reserves space for closing fences when chunking", () => {
    const body = "a".repeat(120);
    const text = `\`\`\`txt\n${body}\n\`\`\``;

    const chunks = chunkDiscordText(text, { maxChars: 50, maxLines: 50 });
    expect(chunks.length).toBeGreaterThan(1);
    for (const chunk of chunks) {
      expect(chunk.length).toBeLessThanOrEqual(50);
      expect(hasBalancedFences(chunk)).toBe(true);
    }
  });

  it("preserves whitespace when splitting long lines", () => {
    const text = Array.from({ length: 40 }, () => "word").join(" ");
    const chunks = chunkDiscordText(text, { maxChars: 20, maxLines: 50 });
    expect(chunks.length).toBeGreaterThan(1);
    expect(chunks.join("")).toBe(text);
  });

  it("preserves mixed whitespace across chunk boundaries", () => {
    const text = "alpha  beta\tgamma   delta epsilon  zeta";
    const chunks = chunkDiscordText(text, { maxChars: 12, maxLines: 50 });
    expect(chunks.length).toBeGreaterThan(1);
    expect(chunks.join("")).toBe(text);
  });

  it("keeps leading whitespace when splitting long lines", () => {
    const text = "    indented line with words that force splits";
    const chunks = chunkDiscordText(text, { maxChars: 14, maxLines: 50 });
    expect(chunks.length).toBeGreaterThan(1);
    expect(chunks.join("")).toBe(text);
  });

  it("keeps reasoning italics balanced across chunks", () => {
    const body = Array.from({ length: 25 }, (_, i) => `${i + 1}. line`).join("\n");
    const text = `Reasoning:\n_${body}_`;

    const chunks = chunkDiscordText(text, { maxLines: 10, maxChars: 2000 });
    expect(chunks.length).toBeGreaterThan(1);

    for (const chunk of chunks) {
      // Each chunk should have balanced italics markers (even count).
      const count = (chunk.match(/_/g) || []).length;
      expect(count % 2).toBe(0);
    }

    // Ensure italics reopen on subsequent chunks
    expect(chunks[0]).toContain("_1. line");
    // Second chunk should reopen italics at the start
    expect(chunks[1].trimStart().startsWith("_")).toBe(true);
  });

  it("keeps reasoning italics balanced when chunks split by char limit", () => {
    const longLine = "This is a very long reasoning line that forces char splits.";
    const body = Array.from({ length: 5 }, () => longLine).join("\n");
    const text = `Reasoning:\n_${body}_`;

    const chunks = chunkDiscordText(text, { maxChars: 80, maxLines: 50 });
    expect(chunks.length).toBeGreaterThan(1);

    for (const chunk of chunks) {
      const underscoreCount = (chunk.match(/_/g) || []).length;
      expect(underscoreCount % 2).toBe(0);
    }
  });

  it("reopens italics while preserving leading whitespace on following chunk", () => {
    const body = [
      "1. line",
      "2. line",
      "3. line",
      "4. line",
      "5. line",
      "6. line",
      "7. line",
      "8. line",
      "9. line",
      "10. line",
      "  11. indented line",
      "12. line",
    ].join("\n");
    const text = `Reasoning:\n_${body}_`;

    const chunks = chunkDiscordText(text, { maxLines: 10, maxChars: 2000 });
    expect(chunks.length).toBeGreaterThan(1);

    const second = chunks[1];
    expect(second.startsWith("_")).toBe(true);
    expect(second).toContain("  11. indented line");
  });
});