import { test } from '@playwright/test';
import fs from 'fs';
import path from 'path';

test.use({ storageState: path.resolve(__dirname, '.auth', 'ig-session.json') });
function randomWait(min = 2500, max = 5000) {
    return Math.floor(Math.random() * (max - min + 1)) + min;
}

function extractPostsFromMedia(m) {
    return {
        id: m.pk,
        code: m.code,
        username: m.user?.username,
        full_name: m.user?.full_name,
        profile_pic_url:
            m.user?.hd_profile_pic_url_info?.url ||
            m.user?.profile_pic_url ||
            null,
        caption: m.caption?.text || null,
        accessibility_caption: m.accessibility_caption || null,
        is_video: m.media_type === 2,
        is_carousel: m.media_type === 8,
        like_count: m.like_count,
        comment_count: m.comment_count,
        taken_at: m.taken_at,
        image_url: m.image_versions2?.candidates?.[0]?.url || null,
        video_url: m.video_versions?.[0]?.url || null,
        carousel_images:
            m.carousel_media?.map((item) => ({
                url: item.image_versions2?.candidates?.[0]?.url || null,
                video_url: item.video_versions?.[0]?.url || null,
                accessibility_caption: item.accessibility_caption || null,
            })) || null,
    };
}

test('scrape following feed', async ({ page }) => {
    test.setTimeout(120_000);
    const posts = [];

    page.on('response', async (response) => {
        const url = response.url();
        if (!url.includes('/graphql/query')) return;
        try {
            const json = await response.json();
            const edges =
                json?.data?.xdt_api__v1__feed__timeline__connection?.edges;
            if (!edges) return;
            for (const edge of edges) {
                const m = edge.node?.media;
                if (m) posts.push(extractPostsFromMedia(m));
            }
        } catch (e) {
            /* not JSON, ignore */
        }
    });

    await page.goto('https://www.instagram.com/?variant=following&hl=en');
    await page.waitForTimeout(4000);

    const preloadedEdges = await page.evaluate(() => {
        function deepFind(obj, targetKey) {
            if (obj === null || typeof obj !== 'object') return null;
            if (targetKey in obj) return obj[targetKey];
            for (const value of Object.values(obj)) {
                const found = deepFind(value, targetKey);
                if (found) return found;
            }
            return null;
        }

        const results = [];
        document
            .querySelectorAll('script[type="application/json"]')
            .forEach((script) => {
                try {
                    const json = JSON.parse(script.textContent);
                    const connection = deepFind(
                        json,
                        'xdt_api__v1__feed__timeline__connection',
                    );
                    if (connection?.edges) {
                        results.push(...connection.edges);
                    }
                } catch (e) {
                    /* not parseable JSON, skip */
                }
            });
        return results;
    });

    for (const edge of preloadedEdges) {
        const m = edge.node?.media;
        if (m) posts.push(extractPostsFromMedia(m));
    }

    for (let i = 0; i < 10; i++) {
        await page.mouse.wheel(0, 3000);
        await page.waitForTimeout(randomWait());
    }

    const outputPath = path.resolve(__dirname, 'output', 'scrape-feed.json');
    fs.mkdirSync(path.dirname(outputPath), { recursive: true });
    fs.writeFileSync(outputPath, JSON.stringify(posts));
});