Merge pull request #13375 from Snuffleupagus/refactor-getContentStream

Improve the `Page.content` and `Page.getContentStream` methods

src/core/document.js

@@ -54,6 +54,7 @@ import {
 } from "./core_utils.js";
 import { NullStream, Stream } from "./stream.js";
 import { AnnotationFactory } from "./annotation.js";
+import { BaseStream } from "./base_stream.js";
 import { calculateMD5 } from "./crypto.js";
 import { Catalog } from "./catalog.js";
 import { Linearization } from "./parser.js";
@@ -136,7 +137,7 @@ class Page {
   }
 
   get content() {
-    return this.pageDict.get("Contents");
+    return this.pageDict.getArray("Contents");
   }
 
   get resources() {
@@ -229,25 +230,20 @@ class Page {
     return shadow(this, "rotate", rotate);
   }
 
+  /**
+   * @returns {Promise<BaseStream>}
+   */
   getContentStream() {
-    const content = this.content;
+    return this.pdfManager.ensure(this, "content").then(content => {
-    let stream;
+      if (content instanceof BaseStream) {
-
+        return content;
-    if (Array.isArray(content)) {
+      }
-      // Fetching the individual streams from the array.
+      if (Array.isArray(content)) {
-      const xref = this.xref;
+        return new StreamsSequenceStream(content);
-      const streams = [];
-      for (const subStream of content) {
-        streams.push(xref.fetchIfRef(subStream));
       }
-      stream = new StreamsSequenceStream(streams);
-    } else if (isStream(content)) {
-      stream = content;
-    } else {
       // Replace non-existent page content with empty content.
-      stream = new NullStream();
+      return new NullStream();
-    }
+    });
-    return stream;
   }
 
   get xfaData() {
@@ -313,10 +309,7 @@ class Page {
     renderInteractiveForms,
     annotationStorage,
   }) {
-    const contentStreamPromise = this.pdfManager.ensure(
+    const contentStreamPromise = this.getContentStream();
-      this,
-      "getContentStream"
-    );
     const resourcesPromise = this.loadResources([
       "ColorSpace",
       "ExtGState",
@@ -420,10 +413,7 @@ class Page {
     sink,
     combineTextItems,
   }) {
-    const contentStreamPromise = this.pdfManager.ensure(
+    const contentStreamPromise = this.getContentStream();
-      this,
-      "getContentStream"
-    );
     const resourcesPromise = this.loadResources([
       "ExtGState",
       "Font",

src/core/primitives.js

@@ -118,16 +118,24 @@ class Dict {
 
   // Same as get(), but dereferences all elements if the result is an Array.
   getArray(key1, key2, key3) {
-    let value = this.get(key1, key2, key3);
+    let value = this._map[key1];
-    if (!Array.isArray(value) || !this.xref) {
+    if (value === undefined && key2 !== undefined) {
-      return value;
+      value = this._map[key2];
-    }
+      if (value === undefined && key3 !== undefined) {
-    value = value.slice(); // Ensure that we don't modify the Dict data.
+        value = this._map[key3];
-    for (let i = 0, ii = value.length; i < ii; i++) {
+      }
-      if (!(value[i] instanceof Ref)) {
+    }
-        continue;
+    if (value instanceof Ref && this.xref) {
+      value = this.xref.fetch(value, this.suppressEncryption);
+    }
+
+    if (Array.isArray(value)) {
+      value = value.slice(); // Ensure that we don't modify the Dict data.
+      for (let i = 0, ii = value.length; i < ii; i++) {
+        if (value[i] instanceof Ref && this.xref) {
+          value[i] = this.xref.fetch(value[i], this.suppressEncryption);
+        }
       }
-      value[i] = this.xref.fetch(value[i], this.suppressEncryption);
     }
     return value;
   }