@pagepocket/cli 0.6.0 → 0.6.1

package/dist/cli.js

@@ -6,6 +6,7 @@ Object.defineProperty(exports, "__esModule", { value: true });
 const core_1 = require("@oclif/core");
 const lib_1 = require("@pagepocket/lib");
 const lighterceptor_adapter_1 = require("@pagepocket/lighterceptor-adapter");
+const puppeteer_adapter_1 = require("@pagepocket/puppeteer-adapter");
 const chalk_1 = __importDefault(require("chalk"));
 const prepare_output_1 = require("./stages/prepare-output");
 const write_snapshot_1 = require("./stages/write-snapshot");
@@ -15,6 +16,7 @@ class PagepocketCommand extends core_1.Command {
         const { args, flags } = await this.parse(PagepocketCommand);
         const targetUrl = args.url;
         const outputFlag = flags.output ? flags.output.trim() : undefined;
+        const runtime = flags.runtime;
         const headers = {
             "user-agent": "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/122.0.0.0 Safari/537.36",
             accept: "text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,image/webp,*/*;q=0.8",
@@ -22,7 +24,9 @@ class PagepocketCommand extends core_1.Command {
             referer: targetUrl
         };
         const snapshot = await (0, with_spinner_1.withSpinner)(async () => {
-            const interceptor = new lighterceptor_adapter_1.LighterceptorAdapter({ headers });
+            const interceptor = runtime === "puppeteer"
+                ? new puppeteer_adapter_1.PuppeteerAdapter({ gotoOptions: { waitUntil: "networkidle0" } })
+                : new lighterceptor_adapter_1.LighterceptorAdapter({ headers });
             const pagepocket = lib_1.PagePocket.fromURL(targetUrl);
             return pagepocket.capture({
                 interceptor,
@@ -55,6 +59,12 @@ PagepocketCommand.flags = {
     output: core_1.Flags.string({
         char: "o",
         description: "Output path for the snapshot HTML file"
+    }),
+    runtime: core_1.Flags.string({
+        char: "r",
+        description: "Interceptor runtime to use",
+        options: ["lighterceptor", "puppeteer"],
+        default: "lighterceptor"
     })
 };
 exports.default = PagepocketCommand;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@pagepocket/cli",
-  "version": "0.6.0",
+  "version": "0.6.1",
   "description": "CLI for capturing offline snapshots of web pages.",
   "main": "dist/index.js",
   "bin": {
@@ -16,8 +16,9 @@
     "@oclif/core": "^4.0.9",
     "chalk": "^4.1.2",
     "ora": "^9.0.0",
-    "@pagepocket/lib": "0.6.0",
-    "@pagepocket/lighterceptor-adapter": "0.6.0"
+    "@pagepocket/lib": "0.6.1",
+    "@pagepocket/puppeteer-adapter": "0.6.1",
+    "@pagepocket/lighterceptor-adapter": "0.6.1"
   },
   "devDependencies": {
     "@types/node": "^20.11.30",
@@ -34,6 +35,6 @@
   "scripts": {
     "build": "tsc",
     "start": "node dist/index.js",
-    "test": "tsx --test specs/**/*.test.ts"
+    "test": "tsx --test specs/*.spec.ts"
   }
 }

package/dist/lib/content-type.js

@@ -1,36 +0,0 @@
-"use strict";
-Object.defineProperty(exports, "__esModule", { value: true });
-exports.isTextResponse = exports.extensionFromContentType = void 0;
-const extensionFromContentType = (contentType) => {
-    if (!contentType) {
-        return "";
-    }
-    if (contentType.includes("text/css"))
-        return ".css";
-    if (contentType.includes("javascript"))
-        return ".js";
-    if (contentType.includes("image/png"))
-        return ".png";
-    if (contentType.includes("image/jpeg"))
-        return ".jpg";
-    if (contentType.includes("image/gif"))
-        return ".gif";
-    if (contentType.includes("image/svg"))
-        return ".svg";
-    if (contentType.includes("font/woff2"))
-        return ".woff2";
-    if (contentType.includes("font/woff"))
-        return ".woff";
-    return "";
-};
-exports.extensionFromContentType = extensionFromContentType;
-const isTextResponse = (contentType) => {
-    const lowered = contentType.toLowerCase();
-    return (lowered.startsWith("text/") ||
-        lowered.includes("json") ||
-        lowered.includes("javascript") ||
-        lowered.includes("xml") ||
-        lowered.includes("svg") ||
-        lowered.includes("html"));
-};
-exports.isTextResponse = isTextResponse;

package/dist/lib/css-rewrite.js

@@ -1,62 +0,0 @@
-"use strict";
-var __importDefault = (this && this.__importDefault) || function (mod) {
-    return (mod && mod.__esModule) ? mod : { "default": mod };
-};
-Object.defineProperty(exports, "__esModule", { value: true });
-exports.rewriteCssUrls = exports.buildDataUrlMap = void 0;
-const promises_1 = __importDefault(require("node:fs/promises"));
-const getHeaderValue = (headers, name) => {
-    // Normalize header names to find the requested value.
-    for (const key in headers) {
-        if (key.toLowerCase() === name.toLowerCase()) {
-            return headers[key];
-        }
-    }
-    return undefined;
-};
-const buildDataUrlMap = (records) => {
-    // Build a lookup of recorded binary responses so CSS url() can be inlined.
-    const map = new Map();
-    for (const record of records) {
-        if (!record || !record.url || !record.responseBodyBase64) {
-            continue;
-        }
-        const headers = record.responseHeaders || {};
-        const contentType = getHeaderValue(headers, "content-type") || "application/octet-stream";
-        map.set(record.url, `data:${contentType};base64,${record.responseBodyBase64}`);
-    }
-    return map;
-};
-exports.buildDataUrlMap = buildDataUrlMap;
-const rewriteCssUrls = async (filePath, cssUrl, dataUrlMap) => {
-    // Replace external URLs in CSS with data URLs to preserve rendering offline.
-    const css = await promises_1.default.readFile(filePath, "utf-8");
-    const urlPattern = /url\(\s*(['"]?)([^'")]+)\1\s*\)/g;
-    const rewritten = css.replace(urlPattern, (match, quote, rawUrl) => {
-        const trimmed = String(rawUrl || "").trim();
-        if (!trimmed || trimmed.startsWith("data:") || trimmed.startsWith("blob:")) {
-            return match;
-        }
-        const absolute = (() => {
-            try {
-                return new URL(trimmed, cssUrl).toString();
-            }
-            catch {
-                return null;
-            }
-        })();
-        if (!absolute) {
-            return match;
-        }
-        const dataUrl = dataUrlMap.get(absolute);
-        if (!dataUrl) {
-            return match;
-        }
-        const safeQuote = quote || "";
-        return `url(${safeQuote}${dataUrl}${safeQuote})`;
-    });
-    if (rewritten !== css) {
-        await promises_1.default.writeFile(filePath, rewritten, "utf-8");
-    }
-};
-exports.rewriteCssUrls = rewriteCssUrls;

package/dist/lib/hackers/capture-network.js

@@ -1,64 +0,0 @@
-"use strict";
-Object.defineProperty(exports, "__esModule", { value: true });
-exports.captureNetworkRecorder = void 0;
-const content_type_1 = require("../content-type");
-exports.captureNetworkRecorder = {
-    id: "capture-network-recorder",
-    stage: "capture",
-    apply: async ({ page, networkRecords }) => {
-        await page.setRequestInterception(true);
-        page.on("request", (request) => {
-            request.continue().catch(() => undefined);
-        });
-        page.on("response", async (response) => {
-            const request = response.request();
-            const url = response.url();
-            const headers = response.headers();
-            const requestHeaders = request.headers();
-            const requestBody = request.postData() || "";
-            // Capture the response body while preserving encoding for replay.
-            const { responseBody, responseBodyBase64, responseEncoding, error } = await (async () => {
-                try {
-                    const buffer = await response.buffer();
-                    const contentType = headers["content-type"] || "";
-                    if ((0, content_type_1.isTextResponse)(contentType)) {
-                        return {
-                            responseBody: buffer.toString("utf-8"),
-                            responseBodyBase64: undefined,
-                            responseEncoding: "text",
-                            error: undefined
-                        };
-                    }
-                    return {
-                        responseBody: undefined,
-                        responseBodyBase64: buffer.toString("base64"),
-                        responseEncoding: "base64",
-                        error: undefined
-                    };
-                }
-                catch (err) {
-                    return {
-                        responseBody: undefined,
-                        responseBodyBase64: undefined,
-                        responseEncoding: undefined,
-                        error: String(err)
-                    };
-                }
-            })();
-            networkRecords.push({
-                url,
-                method: request.method(),
-                requestHeaders,
-                requestBody,
-                status: response.status(),
-                statusText: response.statusText(),
-                responseHeaders: headers,
-                responseBody,
-                responseBodyBase64,
-                responseEncoding,
-                error,
-                timestamp: Date.now()
-            });
-        });
-    }
-};

package/dist/lib/hackers/index.js

@@ -1,22 +0,0 @@
-"use strict";
-Object.defineProperty(exports, "__esModule", { value: true });
-exports.replayHackers = exports.preloadHackers = void 0;
-const preload_fetch_1 = require("./preload-fetch");
-const preload_xhr_1 = require("./preload-xhr");
-const replay_beacon_1 = require("./replay-beacon");
-const replay_dom_rewrite_1 = require("./replay-dom-rewrite");
-const replay_eventsource_1 = require("./replay-eventsource");
-const replay_fetch_1 = require("./replay-fetch");
-const replay_svg_image_1 = require("./replay-svg-image");
-const replay_websocket_1 = require("./replay-websocket");
-const replay_xhr_1 = require("./replay-xhr");
-exports.preloadHackers = [preload_fetch_1.preloadFetchRecorder, preload_xhr_1.preloadXhrRecorder];
-exports.replayHackers = [
-    replay_fetch_1.replayFetchResponder,
-    replay_xhr_1.replayXhrResponder,
-    replay_dom_rewrite_1.replayDomRewriter,
-    replay_svg_image_1.replaySvgImageRewriter,
-    replay_beacon_1.replayBeaconStub,
-    replay_websocket_1.replayWebSocketStub,
-    replay_eventsource_1.replayEventSourceStub
-];

package/dist/lib/hackers/preload-fetch.js

@@ -1,56 +0,0 @@
-"use strict";
-Object.defineProperty(exports, "__esModule", { value: true });
-exports.preloadFetchRecorder = void 0;
-exports.preloadFetchRecorder = {
-    id: "preload-fetch-recorder",
-    stage: "preload",
-    build: () => `
-  // Record fetch calls and responses for replay.
-  const originalFetch = window.fetch.bind(window);
-  const recordFetch = async (input, init) => {
-    trackPendingStart();
-    const url = typeof input === "string" ? toAbsoluteUrl(input) : toAbsoluteUrl(input.url);
-    const method = init && init.method ? init.method : (typeof input === "string" ? "GET" : input.method || "GET");
-    const requestBody = normalizeBody(init && init.body ? init.body : (typeof input === "string" ? undefined : input.body));
-
-    try {
-      const response = await originalFetch(input, init);
-      const clone = response.clone();
-      const responseBody = await clone.text();
-      const headers = {};
-      response.headers.forEach((value, key) => {
-        headers[key] = value;
-      });
-      records.push({
-        kind: "fetch",
-        url,
-        method,
-        requestBody,
-        status: response.status,
-        statusText: response.statusText,
-        responseHeaders: headers,
-        responseBody,
-        timestamp: Date.now()
-      });
-      return response;
-    } catch (error) {
-      records.push({
-        kind: "fetch",
-        url,
-        method,
-        requestBody,
-        error: String(error),
-        timestamp: Date.now()
-      });
-      throw error;
-    } finally {
-      trackPendingEnd();
-    }
-  };
-
-  window.fetch = (input, init) => {
-    return recordFetch(input, init);
-  };
-  window.fetch.__pagepocketOriginal = originalFetch;
-`
-};

package/dist/lib/hackers/preload-image.js

@@ -1,61 +0,0 @@
-"use strict";
-Object.defineProperty(exports, "__esModule", { value: true });
-exports.preloadImageRecorder = void 0;
-exports.preloadImageRecorder = {
-    id: "preload-image-recorder",
-    stage: "preload",
-    build: () => `
-  // Intercept image src assignments and record them without triggering network requests.
-  const NativeImage = window.Image;
-  const placeholderSrc =
-    "data:image/gif;base64,R0lGODlhAQABAIAAAAAAAP///ywAAAAAAQABAAACAUwAOw==";
-  const nativeDescriptor = Object.getOwnPropertyDescriptor(HTMLImageElement.prototype, "src");
-  const nativeSet = nativeDescriptor && nativeDescriptor.set ? nativeDescriptor.set : null;
-  const nativeGet = nativeDescriptor && nativeDescriptor.get ? nativeDescriptor.get : null;
-
-  const recordImage = (input) => {
-    if (!input) return;
-    const url = toAbsoluteUrl(input);
-    records.push({
-      kind: "image",
-      url,
-      method: "GET",
-      requestBody: "",
-      timestamp: Date.now()
-    });
-  };
-
-  const shouldBypass = (value) => {
-    const raw = String(value || "");
-    if (!raw) return true;
-    return raw.startsWith("data:") || raw.startsWith("blob:");
-  };
-
-  Object.defineProperty(HTMLImageElement.prototype, "src", {
-    set(value) {
-      console.info(value)
-
-      if (shouldBypass(value)) {
-        if (nativeSet) nativeSet.call(this, value);
-        return;
-      }
-      const url = toAbsoluteUrl(value);
-      this.__webechoImageSrc = url;
-      recordImage(url);
-      if (nativeSet) nativeSet.call(this, placeholderSrc);
-    },
-    get() {
-      if (this.__webechoImageSrc) return this.__webechoImageSrc;
-      return nativeGet ? nativeGet.call(this) : "";
-    },
-    configurable: true,
-    enumerable: true
-  });
-
-  window.Image = function(...args) {
-    return new NativeImage(...args);
-  };
-  window.Image.prototype = NativeImage.prototype;
-  Object.defineProperty(window.Image, "name", { value: "Image" });
-`
-};

package/dist/lib/hackers/preload-xhr.js

@@ -1,59 +0,0 @@
-"use strict";
-Object.defineProperty(exports, "__esModule", { value: true });
-exports.preloadXhrRecorder = void 0;
-exports.preloadXhrRecorder = {
-    id: "preload-xhr-recorder",
-    stage: "preload",
-    build: () => `
-  // Record XHR calls and responses for replay.
-  const originalOpen = XMLHttpRequest.prototype.open;
-  const originalSend = XMLHttpRequest.prototype.send;
-
-  XMLHttpRequest.prototype.open = function(method, url, ...rest) {
-    this.__pagepocketMethod = method;
-    this.__pagepocketUrl = toAbsoluteUrl(url);
-    return originalOpen.call(this, method, url, ...rest);
-  };
-
-  XMLHttpRequest.prototype.send = function(body) {
-    const xhr = this;
-    const requestBody = normalizeBody(body);
-
-    const onLoadEnd = () => {
-      const responseHeadersRaw = xhr.getAllResponseHeaders();
-      const headers = {};
-      responseHeadersRaw
-        .trim()
-        .split(/\\r?\\n/)
-        .filter(Boolean)
-        .forEach((line) => {
-          const index = line.indexOf(":");
-          if (index > -1) {
-            const key = line.slice(0, index).trim().toLowerCase();
-            const value = line.slice(index + 1).trim();
-            headers[key] = value;
-          }
-        });
-
-      records.push({
-        kind: "xhr",
-        url: xhr.__pagepocketUrl || "",
-        method: xhr.__pagepocketMethod || "GET",
-        requestBody,
-        status: xhr.status,
-        statusText: xhr.statusText,
-        responseHeaders: headers,
-        responseBody: xhr.responseText || "",
-        timestamp: Date.now()
-      });
-
-      xhr.removeEventListener("loadend", onLoadEnd);
-      trackPendingEnd();
-    };
-
-    trackPendingStart();
-    xhr.addEventListener("loadend", onLoadEnd);
-    return originalSend.call(xhr, body);
-  };
-`
-};

package/dist/lib/hackers/replay-beacon.js

@@ -1,21 +0,0 @@
-"use strict";
-Object.defineProperty(exports, "__esModule", { value: true });
-exports.replayBeaconStub = void 0;
-exports.replayBeaconStub = {
-    id: "replay-beacon-stub",
-    stage: "replay",
-    build: () => `
-  // Stub beacon calls so analytics doesn't leak outside the snapshot.
-  if (navigator.sendBeacon) {
-    const originalBeacon = navigator.sendBeacon.bind(navigator);
-    navigator.sendBeacon = (url, data) => {
-      const record = findRecord("POST", url, data);
-      if (record) {
-        return true;
-      }
-      return true;
-    };
-    navigator.sendBeacon.__pagepocketOriginal = originalBeacon;
-  }
-`
-};