Chat with a page

/** * Chat with a page: open a page once in a session, then ask questions about it. Each answer comes from the page * that is open in the browser (session.extract), with the chat so far, and quotes the sentence it is based on. * *   npx tsx node/index.ts            (BOXLINE_API_KEY; PAGE_URL for another page) * * QUESTIONS="first?|second?" answers those and stops; without it, type questions and "exit" to stop. * Writes output/result.json and output/page.md (the page as the model read it). */import { mkdirSync, writeFileSync } from "node:fs";import { join } from "node:path";import { createInterface } from "node:readline/promises";import { Boxline } from "@boxline/sdk";const out = process.env.OUTPUT_DIR ?? "output";const url = process.env.PAGE_URL ?? "https://books.toscrape.com/catalogue/a-light-in-the-attic_1000/index.html";const preset = process.env.QUESTIONS?.split("|").map((q) => q.trim()).filter(Boolean);const bx = new Boxline();type Answer = { answer: string; quote: string; found: boolean };const SCHEMA = {  type: "object",  properties: {    answer: { type: "string", description: "a short answer, from this page only" },    quote: { type: "string", description: "the exact words on the page the answer is based on, copied as they are; empty when the page does not say" },    found: { type: "boolean", description: "false when the page does not answer the question" },  },  required: ["answer", "quote", "found"],};const session = await bx.sessions.create({ timeout: 900, idleTimeout: 300, userMetadata: { example: "chat-with-a-page" } });console.log(`Session: ${session.id}`);const turns: { question: string; answer: string; quote: string; found: boolean }[] = [];let modelUsd = 0;try {  const opened = await session.goto(url, { waitUntil: "load" });  const page = await session.content("markdown");  console.log(`Opened ${page.url} ("${page.title}"), ${page.content.length} characters. Ask about it (type "exit" to stop).\n`);  mkdirSync(out, { recursive: true });  writeFileSync(join(out, "page.md"), page.content);  const rl = preset ? null : createInterface({ input: process.stdin, output: process.stdout });  for (let i = 0; ; i++) {    const question = preset ? preset[i] : (await rl!.question("You: ")).trim();    if (!question || question.toLowerCase() === "exit") break;    if (preset) console.log(`You: ${question}`);    // The chat so far goes into the instruction, so follow-up questions ("and the one after that?") make sense.    const history = turns.map((t) => `Q: ${t.question}\nA: ${t.answer}`).join("\n");    const r = await session.extract<Answer>(      `${history ? `The conversation so far:\n${history}\n\n` : ""}Answer this question about the page: ${question}`,      { schema: SCHEMA },    );    modelUsd += r.usage.costUsd;    if (!r.data.found && !r.data.answer.trim()) r.data.answer = "The page does not say.";    turns.push({ question, ...r.data });    console.log(`AI: ${r.data.answer}${r.data.quote ? `\n    (the page: "${r.data.quote}")` : ""}\n`);  }  rl?.close();  writeFileSync(join(out, "result.json"), JSON.stringify({ url, finalUrl: page.url, title: page.title, status: opened.status, turns, usage: { modelUsd } }, null, 2));  console.log(`${turns.length} answers for $${modelUsd.toFixed(4)} of model use.`);} finally {  await session.stop();}

"""Chat with a page: open a page once in a session, then ask questions about it. Each answer comes from the pagethat is open in the browser (session.extract), with the chat so far, and quotes the sentence it is based on.    python python/main.py            (BOXLINE_API_KEY; PAGE_URL for another page)QUESTIONS="first?|second?" answers those and stops; without it, type questions and "exit" to stop.Writes output/result.json and output/page.md (the page as the model read it)."""import jsonimport osfrom pathlib import Pathfrom boxline import Boxlineout = Path(os.environ.get("OUTPUT_DIR", "output"))url = os.environ.get("PAGE_URL", "https://books.toscrape.com/catalogue/a-light-in-the-attic_1000/index.html")preset = [q.strip() for q in os.environ["QUESTIONS"].split("|") if q.strip()] if os.environ.get("QUESTIONS") else Nonebx = Boxline()SCHEMA = {    "type": "object",    "properties": {        "answer": {"type": "string", "description": "a short answer, from this page only"},        "quote": {"type": "string", "description": "the exact words on the page the answer is based on, copied as they are; empty when the page does not say"},        "found": {"type": "boolean", "description": "false when the page does not answer the question"},    },    "required": ["answer", "quote", "found"],}turns, model_usd = [], 0.0with bx.sessions.create(timeout=900, idle_timeout=300, user_metadata={"example": "chat-with-a-page"}) as session:    print(f"Session: {session.id}", flush=True)    opened = session.goto(url, wait_until="load")    page = session.content("markdown")    print(f'Opened {page["url"]} ("{page["title"]}"), {len(page["content"])} characters. Ask about it (type "exit" to stop).\n', flush=True)    out.mkdir(parents=True, exist_ok=True)    (out / "page.md").write_text(page["content"])    i = 0    while True:        if preset is not None:            question = preset[i] if i < len(preset) else ""            if question:                print(f"You: {question}")        else:            question = input("You: ").strip()        i += 1        if not question or question.lower() == "exit":            break        # The chat so far goes into the instruction, so follow-up questions ("and the one after that?") make sense.        history = "\n".join(f"Q: {t['question']}\nA: {t['answer']}" for t in turns)        r = session.extract(            (f"The conversation so far:\n{history}\n\n" if history else "") + f"Answer this question about the page: {question}",            schema=SCHEMA,        )        model_usd += r["usage"]["costUsd"]        if not r["data"]["found"] and not r["data"]["answer"].strip():            r["data"]["answer"] = "The page does not say."        turns.append({"question": question, **r["data"]})        quote = f'\n    (the page: "{r["data"]["quote"]}")' if r["data"]["quote"] else ""        print(f'AI: {r["data"]["answer"]}{quote}\n', flush=True)    result = {"url": url, "finalUrl": page["url"], "title": page["title"], "status": opened["status"], "turns": turns, "usage": {"modelUsd": model_usd}}    (out / "result.json").write_text(json.dumps(result, indent=2, ensure_ascii=False))    print(f"{len(turns)} answers for ${model_usd:.4f} of model use.")

# Node 18 or newer, in the example's folderexport BOXLINE_API_KEY=…   # your API keynpm install @boxline/sdk tsxnpx tsx node/index.ts

# Python 3.9 or newer, in the example's folderexport BOXLINE_API_KEY=…   # your API keypip install boxline-sdkpython python/main.py