app/api/document/route.js

import { HF_DATASET_BASE_URL, getCorpus, getDocRepoPath, getDocLocalPath } from '../../../utils/config.js';
import fs from 'fs';

const isHFSpace = () => {
    return process.env.HF_TOKEN && process.env.NODE_ENV !== 'development';
};

export async function GET(request) {
    const { searchParams } = new URL(request.url);
    const index = searchParams.get('index');
    const page = searchParams.get('page');
    const corpusId = searchParams.get('corpus');

    if (index === null || page === null) {
        return new Response(
            JSON.stringify({ error: "Missing index or page parameter" }),
            { status: 400, headers: { 'Content-Type': 'application/json' } }
        );
    }

    const indexNum = parseInt(index, 10);
    const pageNum = parseInt(page, 10);

    if (isNaN(indexNum) || isNaN(pageNum) || indexNum < 0 || pageNum < 0) {
        return new Response(
            JSON.stringify({ error: "index and page must be non-negative integers" }),
            { status: 400, headers: { 'Content-Type': 'application/json' } }
        );
    }

    const corpus = getCorpus(corpusId);

    try {
        let pagesData;

        if (isHFSpace()) {
            const docRepoPath = getDocRepoPath(corpus, indexNum);
            const docUrl = `${HF_DATASET_BASE_URL}/raw/main/${docRepoPath}`;
            const res = await fetch(docUrl, {
                headers: { 'Authorization': `Bearer ${process.env.HF_TOKEN}` }
            });

            if (!res.ok) {
                return new Response(
                    JSON.stringify({ error: `doc_${indexNum} not found on HF (${corpus.id})` }),
                    { status: res.status, headers: { 'Content-Type': 'application/json' } }
                );
            }
            pagesData = await res.json();
        } else {
            const filePath = getDocLocalPath(corpus, indexNum);

            if (!fs.existsSync(filePath)) {
                return new Response(
                    JSON.stringify({ error: `doc_${indexNum} not found locally (${corpus.id})` }),
                    { status: 404, headers: { 'Content-Type': 'application/json' } }
                );
            }
            pagesData = JSON.parse(fs.readFileSync(filePath, 'utf-8'));
        }

        const pageData = pagesData.find(p => p.document?.pages?.[0] === pageNum);

        if (!pageData) {
            return new Response(
                JSON.stringify({ error: `Page ${pageNum} not found in doc ${indexNum} (${corpus.id})` }),
                { status: 404, headers: { 'Content-Type': 'application/json' } }
            );
        }

        return new Response(JSON.stringify(pageData), {
            status: 200,
            headers: { 'Content-Type': 'application/json' }
        });
    } catch (error) {
        console.error(error);
        return new Response(
            JSON.stringify({ error: "Failed to fetch document page" }),
            { status: 500, headers: { 'Content-Type': 'application/json' } }
        );
    }
}