"use strict";

// Downloads the Tesseract language data files needed by the OCR parser.
// Both files are fetched from the naptha/tessdata repository (the source used
// by tesseract.js) and saved into the project root so they ship with the repo
// and tesseract.js never has to download them from the network on first OCR.
//
//   npm run download
//
// Uses the "_fast" builds to match what tesseract.js fetches by default.

const fs = require("node:fs");
const path = require("node:path");
const https = require("node:https");
const { pipeline } = require("node:stream/promises");
const { createGunzip } = require("node:zlib");

const projectRoot = path.resolve(__dirname, "..");

// tesseract.js language code -> URL of the .traineddata.gz in naptha/tessdata
// (the repo's data files live on the gh-pages branch)
const ASSETS = {
    eng: "https://raw.githubusercontent.com/naptha/tessdata/gh-pages/4.0.0_fast/eng.traineddata.gz",
    rus: "https://raw.githubusercontent.com/naptha/tessdata/gh-pages/4.0.0_fast/rus.traineddata.gz",
};

async function download(url, dest) {
    try {
        const res = await new Promise((resolve, reject) => {
            https
                .get(url, resolve)
                .on("error", reject);
        });
        if (res.statusCode !== 200) {
            res.resume();
            throw new Error(`HTTP ${res.statusCode} for ${url}`);
        }
        const file = fs.createWriteStream(dest);
        await pipeline(res, file);
    } catch (err) {
        fs.rmSync(dest, { force: true });
        throw err;
    }
}

async function main() {
    const haveAny = Object.values(ASSETS).length > 0;
    if (!haveAny) return;

    let hadError = false;
    for (const [lang, url] of Object.entries(ASSETS)) {
        const gz = path.join(projectRoot, `${lang}.traineddata.gz`);
        const dest = path.join(projectRoot, `${lang}.traineddata`);
        console.log(`[download-languages] Downloading ${lang}...`);
        try {
            await download(url, gz);
            await pipeline(
                fs.createReadStream(gz),
                createGunzip(),
                fs.createWriteStream(dest),
            );
            fs.unlinkSync(gz);
            console.log(`[download-languages] Saved ${dest}`);
        } catch (err) {
            hadError = true;
            console.error(`[download-languages] Failed to download ${lang}:`, err.message);
        }
    }

    if (hadError) {
        console.error("[download-languages] One or more language files failed to download.");
        process.exitCode = 1;
    } else {
        console.log("[download-languages] All language files are up to date.");
    }
}

void main();

"use strict";

// Downloads the Tesseract language data files needed by the OCR parser.
// Both files are fetched from the naptha/tessdata repository (the source used
// by tesseract.js) and saved into the project root so they ship with the repo
// and tesseract.js never has to download them from the network on first OCR.
//
//   npm run download
//
// Uses the "_fast" builds to match what tesseract.js fetches by default.

const fs = require("node:fs");
const path = require("node:path");
const https = require("node:https");
const { pipeline } = require("node:stream/promises");
const { createGunzip } = require("node:zlib");

const projectRoot = path.resolve(__dirname, "..");

// tesseract.js language code -> URL of the .traineddata.gz in naptha/tessdata
// (the repo's data files live on the gh-pages branch)
const ASSETS = {
    eng: "https://raw.githubusercontent.com/naptha/tessdata/gh-pages/4.0.0_fast/eng.traineddata.gz",
    rus: "https://raw.githubusercontent.com/naptha/tessdata/gh-pages/4.0.0_fast/rus.traineddata.gz",
};

async function download(url, dest) {
    try {
        const res = await new Promise((resolve, reject) => {
            https
                .get(url, resolve)
                .on("error", reject);
        });
        if (res.statusCode !== 200) {
            res.resume();
            throw new Error(`HTTP ${res.statusCode} for ${url}`);
        }
        const file = fs.createWriteStream(dest);
        await pipeline(res, file);
    } catch (err) {
        fs.rmSync(dest, { force: true });
        throw err;
    }
}

async function main() {
    const haveAny = Object.values(ASSETS).length > 0;
    if (!haveAny) return;

    let hadError = false;
    for (const [lang, url] of Object.entries(ASSETS)) {
        const gz = path.join(projectRoot, `${lang}.traineddata.gz`);
        const dest = path.join(projectRoot, `${lang}.traineddata`);
        console.log(`[download-languages] Downloading ${lang}...`);
        try {
            await download(url, gz);
            await pipeline(
                fs.createReadStream(gz),
                createGunzip(),
                fs.createWriteStream(dest),
            );
            fs.unlinkSync(gz);
            console.log(`[download-languages] Saved ${dest}`);
        } catch (err) {
            hadError = true;
            console.error(`[download-languages] Failed to download ${lang}:`, err.message);
        }
    }

    if (hadError) {
        console.error("[download-languages] One or more language files failed to download.");
        process.exitCode = 1;
    } else {
        console.log("[download-languages] All language files are up to date.");
    }
}

void main();