Merge pull request #13171 from brendandahl/struct-tree

[api-minor] Add support for basic structure tree for accessibility.

src/core/document.js

@@ -58,6 +58,7 @@ import { calculateMD5 } from "./crypto.js";
 import { Linearization } from "./parser.js";
 import { OperatorList } from "./operator_list.js";
 import { PartialEvaluator } from "./evaluator.js";
+import { StructTreePage } from "./struct_tree.js";
 import { XFAFactory } from "./xfa/factory.js";
 
 const DEFAULT_USER_UNIT = 1.0;
@@ -104,6 +105,10 @@ class Page {
       static createObjId() {
         return `p${pageIndex}_${++idCounters.obj}`;
       }
+
+      static getPageObjId() {
+        return `page${ref.toString()}`;
+      }
     };
   }
 
@@ -406,6 +411,7 @@ class Page {
     handler,
     task,
     normalizeWhitespace,
+    includeMarkedContent,
     sink,
     combineTextItems,
   }) {
@@ -437,12 +443,22 @@ class Page {
         task,
         resources: this.resources,
         normalizeWhitespace,
+        includeMarkedContent,
         combineTextItems,
         sink,
       });
     });
   }
 
+  async getStructTree() {
+    const structTreeRoot = await this.pdfManager.ensureCatalog(
+      "structTreeRoot"
+    );
+    const tree = new StructTreePage(structTreeRoot, this.pageDict);
+    tree.parse();
+    return tree;
+  }
+
   getAnnotationsData(intent) {
     return this._parsedAnnotations.then(function (annotations) {
       const annotationsData = [];
@@ -604,6 +620,10 @@ class PDFDocument {
       static createObjId() {
         unreachable("Abstract method `createObjId` called.");
       }
+
+      static getPageObjId() {
+        unreachable("Abstract method `getPageObjId` called.");
+      }
     };
   }
 

src/core/evaluator.js

@@ -1913,7 +1913,10 @@ class PartialEvaluator {
               return;
             }
             // Other marked content types aren't supported yet.
-            args = [args[0].name];
+            args = [
+              args[0].name,
+              args[1] instanceof Dict ? args[1].get("MCID") : null,
+            ];
 
             break;
           case OPS.beginMarkedContent:
@@ -1973,6 +1976,7 @@ class PartialEvaluator {
     stateManager = null,
     normalizeWhitespace = false,
     combineTextItems = false,
+    includeMarkedContent = false,
     sink,
     seenStyles = new Set(),
   }) {
@@ -2573,6 +2577,7 @@ class PartialEvaluator {
                     stateManager: xObjStateManager,
                     normalizeWhitespace,
                     combineTextItems,
+                    includeMarkedContent,
                     sink: sinkWrapper,
                     seenStyles,
                   })
@@ -2650,6 +2655,38 @@ class PartialEvaluator {
               })
             );
             return;
+          case OPS.beginMarkedContent:
+            if (includeMarkedContent) {
+              textContent.items.push({
+                type: "beginMarkedContent",
+                tag: isName(args[0]) ? args[0].name : null,
+              });
+            }
+            break;
+          case OPS.beginMarkedContentProps:
+            if (includeMarkedContent) {
+              flushTextContentItem();
+              let mcid = null;
+              if (isDict(args[1])) {
+                mcid = args[1].get("MCID");
+              }
+              textContent.items.push({
+                type: "beginMarkedContentProps",
+                id: Number.isInteger(mcid)
+                  ? `${self.idFactory.getPageObjId()}_mcid${mcid}`
+                  : null,
+                tag: isName(args[0]) ? args[0].name : null,
+              });
+            }
+            break;
+          case OPS.endMarkedContent:
+            if (includeMarkedContent) {
+              flushTextContentItem();
+              textContent.items.push({
+                type: "endMarkedContent",
+              });
+            }
+            break;
         } // switch
         if (textContent.items.length >= sink.desiredSize) {
           // Wait for ready, if we reach highWaterMark.

src/core/obj.js

@@ -60,6 +60,7 @@ import { CipherTransformFactory } from "./crypto.js";
 import { ColorSpace } from "./colorspace.js";
 import { GlobalImageCache } from "./image_utils.js";
 import { MetadataParser } from "./metadata_parser.js";
+import { StructTreeRoot } from "./struct_tree.js";
 
 function fetchDestination(dest) {
   return isDict(dest) ? dest.get("D") : dest;
@@ -200,6 +201,32 @@ class Catalog {
     return markInfo;
   }
 
+  get structTreeRoot() {
+    let structTree = null;
+    try {
+      structTree = this._readStructTreeRoot();
+    } catch (ex) {
+      if (ex instanceof MissingDataException) {
+        throw ex;
+      }
+      warn("Unable read to structTreeRoot info.");
+    }
+    return shadow(this, "structTreeRoot", structTree);
+  }
+
+  /**
+   * @private
+   */
+  _readStructTreeRoot() {
+    const obj = this._catDict.get("StructTreeRoot");
+    if (!isDict(obj)) {
+      return null;
+    }
+    const root = new StructTreeRoot(obj);
+    root.init();
+    return root;
+  }
+
   get toplevelPagesDict() {
     const pagesObj = this._catDict.get("Pages");
     if (!isDict(pagesObj)) {
@@ -2626,4 +2653,4 @@ const ObjectLoader = (function () {
   return ObjectLoader;
 })();
 
-export { Catalog, FileSpec, ObjectLoader, XRef };
+export { Catalog, FileSpec, NumberTree, ObjectLoader, XRef };