Scaffold WebExtension: manifest, content collector, background aria2 RPC, popup UI, filters, and scripts

.gitignore

@@ -0,0 +1,4 @@
+dist/
+.web-extension-id
+node_modules/
+

manifest.json

@@ -0,0 +1,31 @@
+{
+  "manifest_version": 3,
+  "name": "Archive.org Link Grabber",
+  "version": "0.1.0",
+  "description": "Filter and export archive.org /download links; copy or send to aria2 RPC.",
+  "action": {
+    "default_title": "Archive.org Link Grabber",
+    "default_popup": "src/popup/index.html"
+  },
+  "permissions": [
+    "storage",
+    "clipboardWrite",
+    "activeTab"
+  ],
+  "host_permissions": [
+    "https://archive.org/*",
+    "http://localhost:6800/*",
+    "https://localhost:6800/*"
+  ],
+  "background": {
+    "service_worker": "src/background/index.js",
+    "type": "module"
+  },
+  "content_scripts": [
+    {
+      "matches": ["https://archive.org/download/*"],
+      "js": ["src/content/collect.js"]
+    }
+  ]
+}
+

package.json

@@ -0,0 +1,14 @@
+{
+  "name": "archive-org-link-grabber",
+  "version": "0.1.0",
+  "private": true,
+  "description": "Firefox WebExtension to filter archive.org /download links and copy/send to aria2.",
+  "scripts": {
+    "dev": "web-ext run --start-url https://archive.org --source-dir .",
+    "build": "web-ext build -o -a dist",
+    "lint": "echo 'Add ESLint config to enable' && exit 0",
+    "format": "echo 'Add Prettier config to enable' && exit 0",
+    "test": "echo 'No tests yet' && exit 0"
+  }
+}
+

src/background/index.js

@@ -0,0 +1,19 @@
+import { addUrisBatch } from "../lib/aria2.js";
+
+browser.runtime.onMessage.addListener(async (msg, sender) => {
+  if (!msg || !msg.type) return;
+
+  if (msg.type === 'aria2.send') {
+    const { endpoint, secret, uris, options } = msg.payload || {};
+    if (!endpoint || !Array.isArray(uris) || uris.length === 0) {
+      return { ok: false, error: 'Missing endpoint or URIs' };
+    }
+    try {
+      const res = await addUrisBatch({ endpoint, secret, uris, options });
+      return { ok: true, result: res };
+    } catch (err) {
+      return { ok: false, error: String(err && err.message || err) };
+    }
+  }
+});
+

src/content/collect.js

@@ -0,0 +1,59 @@
+(() => {
+  function extFromName(name) {
+    const base = name.split('?')[0];
+    const idx = base.lastIndexOf('.');
+    return idx > -1 ? base.slice(idx + 1).toLowerCase() : '';
+  }
+
+  function toItem(a) {
+    try {
+      const url = a.href;
+      const u = new URL(url);
+      const name = decodeURIComponent(u.pathname.split('/').pop() || '');
+      // Attempt to find size/date hints in the DOM (best-effort, optional)
+      let sizeText = '';
+      let dateText = '';
+      const row = a.closest('tr');
+      if (row) {
+        const tds = Array.from(row.querySelectorAll('td'));
+        // Heuristics: look for cells containing size-like patterns or dates
+        for (const td of tds) {
+          const t = td.textContent.trim();
+          if (!sizeText && /\b\d+(?:[\.,]\d+)?\s?(?:[KMG]B|bytes?)\b/i.test(t)) sizeText = t;
+          if (!dateText && /\d{4}-\d{2}-\d{2}/.test(t)) dateText = t.match(/\d{4}-\d{2}-\d{2}/)[0];
+        }
+      }
+      return {
+        url,
+        name,
+        type: extFromName(name),
+        sizeText,
+        dateText
+      };
+    } catch (e) {
+      return null;
+    }
+  }
+
+  async function collectLinks() {
+    const anchors = Array.from(document.querySelectorAll('a[href]'));
+    const items = [];
+    for (const a of anchors) {
+      const href = a.getAttribute('href') || '';
+      // Only include likely file links under /download/
+      if (href.includes('/download/')) {
+        const item = toItem(a);
+        if (item && item.name) items.push(item);
+      }
+    }
+    return items;
+  }
+
+  // Respond to messages from the popup to collect links
+  browser.runtime.onMessage.addListener((msg, sender) => {
+    if (msg && msg.type === 'collectLinks') {
+      return collectLinks();
+    }
+  });
+})();
+

src/lib/aria2.js

@@ -0,0 +1,36 @@
+export async function addUri({ endpoint, secret, uri, options = {} }) {
+  const body = {
+    jsonrpc: "2.0",
+    id: "archive-org-link-grabber:" + Date.now(),
+    method: "aria2.addUri",
+    params: [
+      secret ? `token:${secret}` : undefined,
+      [uri],
+      options
+    ].filter(v => v !== undefined)
+  };
+
+  const res = await fetch(endpoint, {
+    method: 'POST',
+    headers: { 'Content-Type': 'application/json' },
+    body: JSON.stringify(body)
+  });
+  if (!res.ok) {
+    const text = await res.text().catch(() => String(res.status));
+    throw new Error(`aria2 RPC error ${res.status}: ${text}`);
+  }
+  return res.json();
+}
+
+export async function addUrisBatch({ endpoint, secret, uris, options = {} }) {
+  const results = [];
+  for (const uri of uris) {
+    // Simple sequential send; keep it obvious and robust
+    // Consider concurrency/batching later if needed
+    // eslint-disable-next-line no-await-in-loop
+    const r = await addUri({ endpoint, secret, uri, options });
+    results.push(r);
+  }
+  return results;
+}
+

src/lib/filters.js

@@ -0,0 +1,78 @@
+export function buildMatcher(input, { useRegex = false, caseSensitive = false } = {}) {
+  const value = (input || '').trim();
+  if (!value) return () => true;
+  if (useRegex || /^\/.+\/[gimsuy]*$/.test(value)) {
+    // Allow /pattern/flags syntax
+    let pattern = value;
+    let flags = caseSensitive ? '' : 'i';
+    const m = value.match(/^\/(.*)\/([gimsuy]*)$/);
+    if (m) { pattern = m[1]; flags = m[2] || flags; }
+    const re = new RegExp(pattern, flags);
+    return (s) => re.test(String(s || ''));
+  }
+  const needle = caseSensitive ? value : value.toLowerCase();
+  return (s) => {
+    const hay = caseSensitive ? String(s || '') : String(s || '').toLowerCase();
+    return hay.includes(needle);
+  };
+}
+
+export function parseSizeToBytes(v) {
+  if (v == null) return undefined;
+  const s = String(v).trim();
+  if (!s) return undefined;
+  const m = s.match(/^(\d+(?:[\.,]\d+)?)\s*(B|KB|MB|GB|TB)?$/i);
+  if (!m) return Number(s) || undefined;
+  const n = parseFloat(m[1].replace(',', '.'));
+  const unit = (m[2] || 'B').toUpperCase();
+  const mult = { B: 1, KB: 1024, MB: 1024 ** 2, GB: 1024 ** 3, TB: 1024 ** 4 }[unit];
+  return Math.round(n * mult);
+}
+
+export function extFromName(name) {
+  const base = String(name || '').split('?')[0];
+  const idx = base.lastIndexOf('.');
+  return idx > -1 ? base.slice(idx + 1).toLowerCase() : '';
+}
+
+export function applyFilters(items, filters) {
+  const {
+    name = '',
+    type = '',
+    nameRegex = false,
+    typeRegex = false,
+    caseSensitive = false,
+    sizeMin = '',
+    sizeMax = '',
+    dateFrom = '',
+    dateTo = ''
+  } = filters || {};
+
+  const nameMatch = buildMatcher(name, { useRegex: nameRegex, caseSensitive });
+  const typeMatch = buildMatcher(type, { useRegex: typeRegex, caseSensitive });
+  const minBytes = parseSizeToBytes(sizeMin);
+  const maxBytes = parseSizeToBytes(sizeMax);
+  const fromTime = dateFrom ? new Date(dateFrom).getTime() : undefined;
+  const toTime = dateTo ? new Date(dateTo).getTime() : undefined;
+
+  return (items || []).filter((it) => {
+    const okName = nameMatch(it.name);
+    const ext = it.type || extFromName(it.name);
+    const okType = typeMatch(ext);
+    let okSize = true;
+    if (minBytes != null || maxBytes != null) {
+      // Best-effort: parse from sizeText if provided
+      const s = it.sizeBytes != null ? it.sizeBytes : parseSizeToBytes(it.sizeText);
+      if (minBytes != null && (s == null || s < minBytes)) okSize = false;
+      if (maxBytes != null && (s != null && s > maxBytes)) okSize = false;
+    }
+    let okDate = true;
+    if (fromTime != null || toTime != null) {
+      const t = it.timeMs != null ? it.timeMs : (it.dateText ? new Date(it.dateText).getTime() : undefined);
+      if (fromTime != null && (t == null || t < fromTime)) okDate = false;
+      if (toTime != null && (t != null && t > toTime)) okDate = false;
+    }
+    return okName && okType && okSize && okDate;
+  });
+}
+

src/popup/index.html

@@ -0,0 +1,58 @@
+<!doctype html>
+<html>
+  <head>
+    <meta charset="utf-8" />
+    <title>Archive.org Link Grabber</title>
+    <link rel="stylesheet" href="styles.css" />
+  </head>
+  <body>
+    <header>
+      <h1>Archive.org Link Grabber</h1>
+    </header>
+    <section class="filters">
+      <div class="row">
+        <label>Type/Ext <input id="f-type" placeholder="mp4|mkv" /></label>
+        <label class="chk"><input type="checkbox" id="f-type-re" /> Regex</label>
+      </div>
+      <div class="row">
+        <label>Name <input id="f-name" placeholder="/^(movie|clip).*\\.mp4$/i" /></label>
+        <label class="chk"><input type="checkbox" id="f-name-re" /> Regex</label>
+      </div>
+      <div class="row">
+        <label>Size ≥ <input id="f-size-min" placeholder="100MB" size="8" /></label>
+        <label>≤ <input id="f-size-max" placeholder="2GB" size="8" /></label>
+      </div>
+      <div class="row">
+        <label>Date from <input id="f-date-from" type="date" /></label>
+        <label>to <input id="f-date-to" type="date" /></label>
+      </div>
+      <div class="row">
+        <label class="chk"><input type="checkbox" id="f-case" /> Case sensitive</label>
+      </div>
+      <div class="row actions">
+        <button id="btn-refresh">Load Links</button>
+        <span id="count">0</span> matches
+      </div>
+    </section>
+
+    <section class="aria2">
+      <h2>aria2 RPC</h2>
+      <div class="row">
+        <label>Endpoint <input id="rpc-endpoint" placeholder="http://localhost:6800/jsonrpc" /></label>
+      </div>
+      <div class="row">
+        <label>Secret <input id="rpc-secret" type="password" placeholder="your-token" /></label>
+      </div>
+      <div class="row actions">
+        <button id="btn-copy">Copy Links</button>
+        <button id="btn-send">Send to aria2</button>
+      </div>
+      <div class="row">
+        <output id="status"></output>
+      </div>
+    </section>
+
+    <script type="module" src="index.js"></script>
+  </body>
+  </html>
+

src/popup/index.js

@@ -0,0 +1,117 @@
+import { applyFilters } from "../lib/filters.js";
+
+const els = {
+  fType: document.getElementById('f-type'),
+  fTypeRe: document.getElementById('f-type-re'),
+  fName: document.getElementById('f-name'),
+  fNameRe: document.getElementById('f-name-re'),
+  fSizeMin: document.getElementById('f-size-min'),
+  fSizeMax: document.getElementById('f-size-max'),
+  fDateFrom: document.getElementById('f-date-from'),
+  fDateTo: document.getElementById('f-date-to'),
+  fCase: document.getElementById('f-case'),
+  btnRefresh: document.getElementById('btn-refresh'),
+  btnCopy: document.getElementById('btn-copy'),
+  btnSend: document.getElementById('btn-send'),
+  rpcEndpoint: document.getElementById('rpc-endpoint'),
+  rpcSecret: document.getElementById('rpc-secret'),
+  count: document.getElementById('count'),
+  status: document.getElementById('status'),
+};
+
+let allItems = [];
+
+function getFilters() {
+  return {
+    type: els.fType.value,
+    name: els.fName.value,
+    typeRegex: els.fTypeRe.checked,
+    nameRegex: els.fNameRe.checked,
+    sizeMin: els.fSizeMin.value,
+    sizeMax: els.fSizeMax.value,
+    dateFrom: els.fDateFrom.value,
+    dateTo: els.fDateTo.value,
+    caseSensitive: els.fCase.checked,
+  };
+}
+
+async function getActiveTabId() {
+  const tabs = await browser.tabs.query({ active: true, currentWindow: true });
+  return tabs[0]?.id;
+}
+
+async function refresh() {
+  els.status.textContent = '';
+  try {
+    const tabId = await getActiveTabId();
+    if (!tabId) throw new Error('No active tab');
+    const items = await browser.tabs.sendMessage(tabId, { type: 'collectLinks' });
+    allItems = items || [];
+    updateCount();
+  } catch (e) {
+    els.status.textContent = String(e?.message || e);
+  }
+}
+
+function filtered() {
+  return applyFilters(allItems, getFilters());
+}
+
+function updateCount() {
+  els.count.textContent = String(filtered().length);
+}
+
+async function copyLinks() {
+  const urls = filtered().map(i => i.url).join('\n');
+  await navigator.clipboard.writeText(urls);
+  els.status.textContent = `Copied ${filtered().length} link(s) to clipboard.`;
+}
+
+async function sendToAria2() {
+  const endpoint = els.rpcEndpoint.value.trim();
+  const secret = els.rpcSecret.value.trim();
+  const uris = filtered().map(i => i.url);
+  if (!endpoint) { els.status.textContent = 'Please set aria2 endpoint.'; return; }
+  if (uris.length === 0) { els.status.textContent = 'No links to send.'; return; }
+
+  // Persist settings
+  await browser.storage.local.set({ aria2: { endpoint, secret } });
+  els.status.textContent = 'Sending to aria2…';
+
+  const res = await browser.runtime.sendMessage({
+    type: 'aria2.send',
+    payload: { endpoint, secret, uris }
+  });
+  if (res?.ok) {
+    els.status.textContent = `Sent ${uris.length} link(s) to aria2.`;
+  } else {
+    els.status.textContent = `Error: ${res?.error || 'unknown'}`;
+  }
+}
+
+async function restoreSettings() {
+  const { aria2 } = await browser.storage.local.get('aria2');
+  if (aria2) {
+    if (aria2.endpoint) els.rpcEndpoint.value = aria2.endpoint;
+    if (aria2.secret) els.rpcSecret.value = aria2.secret;
+  } else {
+    els.rpcEndpoint.value = 'http://localhost:6800/jsonrpc';
+  }
+}
+
+function wire() {
+  els.btnRefresh.addEventListener('click', refresh);
+  els.btnCopy.addEventListener('click', copyLinks);
+  els.btnSend.addEventListener('click', sendToAria2);
+  [
+    els.fType, els.fTypeRe, els.fName, els.fNameRe, els.fSizeMin, els.fSizeMax,
+    els.fDateFrom, els.fDateTo, els.fCase
+  ].forEach(el => el.addEventListener('input', updateCount));
+}
+
+document.addEventListener('DOMContentLoaded', async () => {
+  wire();
+  await restoreSettings();
+  await refresh();
+});
+

src/popup/styles.css

@@ -0,0 +1,11 @@
+body { font-family: system-ui, -apple-system, Segoe UI, Roboto, sans-serif; margin: 10px; }
+header h1 { font-size: 14px; margin: 0 0 8px; }
+section { margin-bottom: 10px; }
+.row { display: flex; align-items: center; gap: 8px; margin: 6px 0; flex-wrap: wrap; }
+label { font-size: 12px; }
+label input { margin-left: 6px; }
+.chk { align-items: center; }
+button { font-size: 12px; padding: 4px 10px; }
+#count { font-weight: bold; margin: 0 4px; }
+#status { font-size: 12px; white-space: pre-wrap; }
+