gecko-dev/toolkit/components/narrate/Narrator.jsm

/* This Source Code Form is subject to the terms of the Mozilla Public
 * License, v. 2.0. If a copy of the MPL was not distributed with this file,
 * You can obtain one at http://mozilla.org/MPL/2.0/. */

"use strict";

const { interfaces: Ci, utils: Cu } = Components;

Cu.import("resource://gre/modules/XPCOMUtils.jsm");

XPCOMUtils.defineLazyModuleGetter(this, "LanguageDetector",
  "resource:///modules/translation/LanguageDetector.jsm");
XPCOMUtils.defineLazyModuleGetter(this, "Services",
  "resource://gre/modules/Services.jsm");

this.EXPORTED_SYMBOLS = [ "Narrator" ];

// Maximum time into paragraph when pressing "skip previous" will go
// to previous paragraph and not the start of current one.
const PREV_THRESHOLD = 2000;
// All text-related style rules that we should copy over to the highlight node.
const kTextStylesRules = ["font-family", "font-kerning", "font-size",
  "font-size-adjust", "font-stretch", "font-variant", "font-weight",
  "line-height", "letter-spacing", "text-orientation",
  "text-transform", "word-spacing"];

function Narrator(win) {
  this._winRef = Cu.getWeakReference(win);
  this._inTest = Services.prefs.getBoolPref("narrate.test");
  this._speechOptions = {};
  this._startTime = 0;
  this._stopped = false;

  this.languagePromise = new Promise(resolve => {
    let detect = () => {
      win.document.removeEventListener("AboutReaderContentReady", detect);
      let sampleText = this._doc.getElementById(
        "moz-reader-content").textContent.substring(0, 60 * 1024);
      LanguageDetector.detectLanguage(sampleText).then(result => {
        resolve(result.confident ? result.language : null);
      });
    };

    if (win.document.body.classList.contains("loaded")) {
      detect();
    } else {
      win.document.addEventListener("AboutReaderContentReady", detect);
    }
  });
}

Narrator.prototype = {
  get _doc() {
    return this._winRef.get().document;
  },

  get _win() {
    return this._winRef.get();
  },

  get _treeWalker() {
    if (!this._treeWalkerRef) {
      let wu = this._win.QueryInterface(
        Ci.nsIInterfaceRequestor).getInterface(Ci.nsIDOMWindowUtils);
      let nf = this._win.NodeFilter;

      let filter = {
        _matches: new Set(),

        // We want high-level elements that have non-empty text nodes.
        // For example, paragraphs. But nested anchors and other elements
        // are not interesting since their text already appears in their
        // parent's textContent.
        acceptNode: function(node) {
          if (this._matches.has(node.parentNode)) {
            // Reject sub-trees of accepted nodes.
            return nf.FILTER_REJECT;
          }

          if (!/\S/.test(node.textContent)) {
            // Reject nodes with no text.
            return nf.FILTER_REJECT;
          }

          let bb = wu.getBoundsWithoutFlushing(node);
          if (!bb.width || !bb.height) {
            // Skip non-rendered nodes. We don't reject because a zero-sized
            // container can still have visible, "overflowed", content.
            return nf.FILTER_SKIP;
          }

          for (let c = node.firstChild; c; c = c.nextSibling) {
            if (c.nodeType == c.TEXT_NODE && /\S/.test(c.textContent)) {
              // If node has a non-empty text child accept it.
              this._matches.add(node);
              return nf.FILTER_ACCEPT;
            }
          }

          return nf.FILTER_SKIP;
        }
      };

      this._treeWalkerRef = new WeakMap();

      // We can't hold a weak reference on the treewalker, because there
      // are no other strong references, and it will be GC'ed. Instead,
      // we rely on the window's lifetime and use it as a weak reference.
      this._treeWalkerRef.set(this._win,
        this._doc.createTreeWalker(this._doc.getElementById("container"),
          nf.SHOW_ELEMENT, filter, false));
    }

    return this._treeWalkerRef.get(this._win);
  },

  get _timeIntoParagraph() {
    let rv = Date.now() - this._startTime;
    return rv;
  },

  get speaking() {
    return this._win.speechSynthesis.speaking ||
      this._win.speechSynthesis.pending;
  },

  _getVoice: function(voiceURI) {
    if (!this._voiceMap || !this._voiceMap.has(voiceURI)) {
      this._voiceMap = new Map(
        this._win.speechSynthesis.getVoices().map(v => [v.voiceURI, v]));
    }

    return this._voiceMap.get(voiceURI);
  },

  _isParagraphInView: function(paragraph) {
    if (!paragraph) {
      return false;
    }

    let bb = paragraph.getBoundingClientRect();
    return bb.top >= 0 && bb.top < this._win.innerHeight;
  },

  _sendTestEvent: function(eventType, detail) {
    let win = this._win;
    win.dispatchEvent(new win.CustomEvent(eventType,
      { detail: Cu.cloneInto(detail, win.document) }));
  },

  _speakInner: function() {
    this._win.speechSynthesis.cancel();
    let tw = this._treeWalker;
    let paragraph = tw.currentNode;
    if (paragraph == tw.root) {
      this._sendTestEvent("paragraphsdone", {});
      return Promise.resolve();
    }

    let utterance = new this._win.SpeechSynthesisUtterance(
      paragraph.textContent);
    utterance.rate = this._speechOptions.rate;
    if (this._speechOptions.voice) {
      utterance.voice = this._speechOptions.voice;
    } else {
      utterance.lang = this._speechOptions.lang;
    }

    this._startTime = Date.now();

    let highlighter = new Highlighter(paragraph);

    if (this._inTest) {
      let onTestSynthEvent = e => {
        if (e.detail.type == "boundary") {
          let args = Object.assign({ utterance }, e.detail.args);
          let evt = new this._win.SpeechSynthesisEvent(e.detail.type, args);
          utterance.dispatchEvent(evt);
        }
      };

      let removeListeners = () => {
        this._win.removeEventListener("testsynthevent", onTestSynthEvent);
      };

      this._win.addEventListener("testsynthevent", onTestSynthEvent);
      utterance.addEventListener("end", removeListeners);
      utterance.addEventListener("error", removeListeners);
    }

    return new Promise((resolve, reject) => {
      utterance.addEventListener("start", () => {
        paragraph.classList.add("narrating");
        let bb = paragraph.getBoundingClientRect();
        if (bb.top < 0 || bb.bottom > this._win.innerHeight) {
          paragraph.scrollIntoView({ behavior: "smooth", block: "start"});
        }

        if (this._inTest) {
          this._sendTestEvent("paragraphstart", {
            voice: utterance.chosenVoiceURI,
            rate: utterance.rate,
            paragraph: paragraph.textContent,
            tag: paragraph.localName
          });
        }
      });

      utterance.addEventListener("end", () => {
        if (!this._win) {
          // page got unloaded, don't do anything.
          return;
        }

        highlighter.remove();
        paragraph.classList.remove("narrating");
        this._startTime = 0;
        if (this._inTest) {
          this._sendTestEvent("paragraphend", {});
        }

        if (this._stopped) {
          // User pressed stopped.
          resolve();
        } else {
          tw.currentNode = tw.nextNode() || tw.root;
          this._speakInner().then(resolve, reject);
        }
      });

      utterance.addEventListener("error", () => {
        reject("speech synthesis failed");
      });

      utterance.addEventListener("boundary", e => {
        if (e.name != "word") {
          // We are only interested in word boundaries for now.
          return;
        }

        // Match non-whitespace. This isn't perfect, but the most universal
        // solution for now.
        let reWordBoundary = /\S+/g;
        // Match the first word from the boundary event offset.
        reWordBoundary.lastIndex = e.charIndex;
        let firstIndex = reWordBoundary.exec(paragraph.textContent);
        if (firstIndex) {
          highlighter.highlight(firstIndex.index, reWordBoundary.lastIndex);
          if (this._inTest) {
            this._sendTestEvent("wordhighlight", {
              start: firstIndex.index,
              end: reWordBoundary.lastIndex
            });
          }
        }
      });

      this._win.speechSynthesis.speak(utterance);
    });
  },

  start: function(speechOptions) {
    this._speechOptions = {
      rate: speechOptions.rate,
      voice: this._getVoice(speechOptions.voice)
    };

    this._stopped = false;
    return this.languagePromise.then(language => {
      if (!this._speechOptions.voice) {
        this._speechOptions.lang = language;
      }

      let tw = this._treeWalker;
      if (!this._isParagraphInView(tw.currentNode)) {
        tw.currentNode = tw.root;
        while (tw.nextNode()) {
          if (this._isParagraphInView(tw.currentNode)) {
            break;
          }
        }
      }
      if (tw.currentNode == tw.root) {
        tw.nextNode();
      }

      return this._speakInner();
    });
  },

  stop: function() {
    this._stopped = true;
    this._win.speechSynthesis.cancel();
  },

  skipNext: function() {
    this._win.speechSynthesis.cancel();
  },

  skipPrevious: function() {
    this._goBackParagraphs(this._timeIntoParagraph < PREV_THRESHOLD ? 2 : 1);
  },

  setRate: function(rate) {
    this._speechOptions.rate = rate;
    /* repeat current paragraph */
    this._goBackParagraphs(1);
  },

  setVoice: function(voice) {
    this._speechOptions.voice = this._getVoice(voice);
    /* repeat current paragraph */
    this._goBackParagraphs(1);
  },

  _goBackParagraphs: function(count) {
    let tw = this._treeWalker;
    for (let i = 0; i < count; i++) {
      if (!tw.previousNode()) {
        tw.currentNode = tw.root;
      }
    }
    this._win.speechSynthesis.cancel();
  }
};

/**
 * The Highlighter class is used to highlight a range of text in a container.
 *
 * @param {nsIDOMElement} container a text container
 */
function Highlighter(container) {
  this.container = container;
}

Highlighter.prototype = {
  /**
   * Highlight the range within offsets relative to the container.
   *
   * @param {Number} startOffset the start offset
   * @param {Number} endOffset the end offset
   */
  highlight: function(startOffset, endOffset) {
    let containerRect = this.container.getBoundingClientRect();
    let range = this._getRange(startOffset, endOffset);
    let rangeRects = range.getClientRects();
    let win = this.container.ownerDocument.defaultView;
    let computedStyle = win.getComputedStyle(range.endContainer.parentNode);
    let nodes = this._getFreshHighlightNodes(rangeRects.length);

    let textStyle = {};
    for (let textStyleRule of kTextStylesRules) {
      textStyle[textStyleRule] = computedStyle[textStyleRule];
    }

    for (let i = 0; i < rangeRects.length; i++) {
      let r = rangeRects[i];
      let node = nodes[i];

      let style = Object.assign({
        "top": `${r.top - containerRect.top + r.height / 2}px`,
        "left": `${r.left - containerRect.left + r.width / 2}px`,
        "width": `${r.width}px`,
        "height": `${r.height}px`
      }, textStyle);

      // Enables us to vary the CSS transition on a line change.
      node.classList.toggle("newline", style.top != node.dataset.top);
      node.dataset.top = style.top;

      // Enables CSS animations.
      node.classList.remove("animate");
      win.requestAnimationFrame(() => {
        node.classList.add("animate");
      });

      // Enables alternative word display with a CSS pseudo-element.
      node.dataset.word = range.toString();

      // Apply style
      node.style = Object.entries(style).map(
        s => `${s[0]}: ${s[1]};`).join(" ");
    }
  },

  /**
   * Releases reference to container and removes all highlight nodes.
   */
  remove: function() {
    for (let node of this._nodes) {
      node.remove();
    }

    this.container = null;
  },

  /**
   * Returns specified amount of highlight nodes. Creates new ones if necessary
   * and purges any additional nodes that are not needed.
   *
   * @param {Number} count number of nodes needed
   */
  _getFreshHighlightNodes: function(count) {
    let doc = this.container.ownerDocument;
    let nodes = Array.from(this._nodes);

    // Remove nodes we don't need anymore (nodes.length - count > 0).
    for (let toRemove = 0; toRemove < nodes.length - count; toRemove++) {
      nodes.shift().remove();
    }

    // Add additional nodes if we need them (count - nodes.length > 0).
    for (let toAdd = 0; toAdd < count - nodes.length; toAdd++) {
      let node = doc.createElement("div");
      node.className = "narrate-word-highlight";
      this.container.appendChild(node);
      nodes.push(node);
    }

    return nodes;
  },

  /**
   * Create and return a range object with the start and end offsets relative
   * to the container node.
   *
   * @param {Number} startOffset the start offset
   * @param {Number} endOffset the end offset
   */
  _getRange: function(startOffset, endOffset) {
    let doc = this.container.ownerDocument;
    let i = 0;
    let treeWalker = doc.createTreeWalker(
      this.container, doc.defaultView.NodeFilter.SHOW_TEXT);
    let node = treeWalker.nextNode();

    function _findNodeAndOffset(offset) {
      do {
        let length = node.data.length;
        if (offset >= i && offset <= i + length) {
          return [node, offset - i];
        }
        i += length;
      } while ((node = treeWalker.nextNode()));

      // Offset is out of bounds, return last offset of last node.
      node = treeWalker.lastChild();
      return [node, node.data.length];
    }

    let range = doc.createRange();
    range.setStart(..._findNodeAndOffset(startOffset));
    range.setEnd(..._findNodeAndOffset(endOffset));

    return range;
  },

  /*
   * Get all existing highlight nodes for container.
   */
  get _nodes() {
    return this.container.querySelectorAll(".narrate-word-highlight");
  }
};