From b4421b076aa6eda4490b34dfbf5b3e635be49aea Mon Sep 17 00:00:00 2001
From: Calixte Denizet <calixte.denizet@gmail.com>
Date: Sat, 23 Jan 2021 18:21:48 +0100
Subject: [PATCH] Modifiy the way to compute baseline to have a better match
 between canvas and text layer  - use ascent of the fallback font instead of
 the one from pdf to position spans  - use TextMetrics.fontBoundingBoxAscent
 if available or  - use a basic heuristic to guess ascent in drawing char on a
 canvas  - compute ascent as a ratio of font height

---
 src/display/text_layer.js | 79 +++++++++++++++++++++++++++++++++++----
 1 file changed, 71 insertions(+), 8 deletions(-)

diff --git a/src/display/text_layer.js b/src/display/text_layer.js
index 19e750729..6903b1804 100644
--- a/src/display/text_layer.js
+++ b/src/display/text_layer.js
@@ -54,6 +54,9 @@ import {
  */
 const renderTextLayer = (function renderTextLayerClosure() {
   const MAX_TEXT_DIVS_TO_RENDER = 100000;
+  const DEFAULT_FONT_SIZE = 30;
+  const DEFAULT_FONT_ASCENT = 0.8;
+  const ascentCache = new Map();
 
   const NonWhitespaceRegexp = /\S/;
 
@@ -61,7 +64,70 @@ const renderTextLayer = (function renderTextLayerClosure() {
     return !NonWhitespaceRegexp.test(str);
   }
 
-  function appendText(task, geom, styles) {
+  function getAscent(fontFamily, ctx) {
+    const cachedAscent = ascentCache.get(fontFamily);
+    if (cachedAscent) {
+      return cachedAscent;
+    }
+
+    ctx.save();
+    ctx.font = `${DEFAULT_FONT_SIZE}px ${fontFamily}`;
+    const metrics = ctx.measureText("");
+
+    // Both properties aren't available by default in Firefox.
+    let ascent = metrics.fontBoundingBoxAscent;
+    let descent = Math.abs(metrics.fontBoundingBoxDescent);
+    if (ascent) {
+      ctx.restore();
+      const ratio = ascent / (ascent + descent);
+      ascentCache.set(fontFamily, ratio);
+      return ratio;
+    }
+
+    // Try basic heuristic to guess ascent/descent.
+    // Draw a g with baseline at 0,0 and then get the line
+    // number where a pixel has non-null red component (starting
+    // from bottom).
+    ctx.strokeStyle = "red";
+    ctx.clearRect(0, 0, DEFAULT_FONT_SIZE, DEFAULT_FONT_SIZE);
+    ctx.strokeText("g", 0, 0);
+    let pixels = ctx.getImageData(0, 0, DEFAULT_FONT_SIZE, DEFAULT_FONT_SIZE)
+      .data;
+    descent = 0;
+    for (let i = pixels.length - 1 - 3; i >= 0; i -= 4) {
+      if (pixels[i] > 0) {
+        descent = Math.ceil(i / 4 / DEFAULT_FONT_SIZE);
+        break;
+      }
+    }
+
+    // Draw an A with baseline at 0,DEFAULT_FONT_SIZE and then get the line
+    // number where a pixel has non-null red component (starting
+    // from top).
+    ctx.clearRect(0, 0, DEFAULT_FONT_SIZE, DEFAULT_FONT_SIZE);
+    ctx.strokeText("A", 0, DEFAULT_FONT_SIZE);
+    pixels = ctx.getImageData(0, 0, DEFAULT_FONT_SIZE, DEFAULT_FONT_SIZE).data;
+    ascent = 0;
+    for (let i = 0, ii = pixels.length; i < ii; i += 4) {
+      if (pixels[i] > 0) {
+        ascent = DEFAULT_FONT_SIZE - Math.floor(i / 4 / DEFAULT_FONT_SIZE);
+        break;
+      }
+    }
+
+    ctx.restore();
+
+    if (ascent) {
+      const ratio = ascent / (ascent + descent);
+      ascentCache.set(fontFamily, ratio);
+      return ratio;
+    }
+
+    ascentCache.set(fontFamily, DEFAULT_FONT_ASCENT);
+    return DEFAULT_FONT_ASCENT;
+  }
+
+  function appendText(task, geom, styles, ctx) {
     // Initialize all used properties to keep the caches monomorphic.
     const textDiv = document.createElement("span");
     const textDivProperties = {
@@ -90,12 +156,7 @@ const renderTextLayer = (function renderTextLayerClosure() {
       angle += Math.PI / 2;
     }
     const fontHeight = Math.hypot(tx[2], tx[3]);
-    let fontAscent = fontHeight;
-    if (style.ascent) {
-      fontAscent = style.ascent * fontAscent;
-    } else if (style.descent) {
-      fontAscent = (1 + style.descent) * fontAscent;
-    }
+    const fontAscent = fontHeight * getAscent(style.fontFamily, ctx);
 
     let left, top;
     if (angle === 0) {
@@ -578,7 +639,7 @@ const renderTextLayer = (function renderTextLayerClosure() {
     _processItems(items, styleCache) {
       for (let i = 0, len = items.length; i < len; i++) {
         this._textContentItemsStr.push(items[i].str);
-        appendText(this, items[i], styleCache);
+        appendText(this, items[i], styleCache, this._layoutTextCtx);
       }
     },
 
@@ -628,6 +689,8 @@ const renderTextLayer = (function renderTextLayerClosure() {
 
       // The temporary canvas is used to measure text length in the DOM.
       const canvas = this._document.createElement("canvas");
+      canvas.height = canvas.width = DEFAULT_FONT_SIZE;
+
       if (
         typeof PDFJSDev === "undefined" ||
         PDFJSDev.test("MOZCENTRAL || GENERIC")