Merge pull request #16223 from calixteman/16221

Create a new chunk when the char is too rised compared to the previous one

src/core/evaluator.js

@@ -2341,6 +2341,12 @@ class PartialEvaluator {
     const SPACE_IN_FLOW_MIN_FACTOR = 0.102;
     const SPACE_IN_FLOW_MAX_FACTOR = 0.6;
 
+    // If a char is too high/too low compared to the previous we just create
+    // a new chunk.
+    // If the advance isn't in the +/-VERTICAL_SHIFT_RATIO * height range then
+    // a new chunk is created.
+    const VERTICAL_SHIFT_RATIO = 0.25;
+
     const self = this;
     const xref = this.xref;
     const showSpacedTextBuffer = [];
@@ -2649,6 +2655,10 @@ class PartialEvaluator {
           }
         }
 
+        if (Math.abs(advanceX) > textContentItem.width * VERTICAL_SHIFT_RATIO) {
+          flushTextContentItem();
+        }
+
         return true;
       }
 
@@ -2706,6 +2716,10 @@ class PartialEvaluator {
         }
       }
 
+      if (Math.abs(advanceY) > textContentItem.height * VERTICAL_SHIFT_RATIO) {
+        flushTextContentItem();
+      }
+
       return true;
     }
 

test/pdfs/.gitignore

@@ -581,3 +581,4 @@
 !issue16063.pdf
 !issue16067.pdf
 !bug1820909.1.pdf
+!issue16221.pdf

test/test_manifest.json

@@ -7510,5 +7510,12 @@
       "md5": "f71e89ebe3d6e75e0c83ce41cd72df1f",
       "link": true,
       "type": "other"
+   },
+   {
+      "id": "issue16221-text",
+      "file": "pdfs/issue16221.pdf",
+      "md5": "62d93c9b3aa4ba3af5446504632e78a5",
+      "rounds": 1,
+      "type": "text"
    }
 ]

test/unit/api_spec.js

@@ -2624,6 +2624,17 @@ Caron Broadcasting, Inc., an Ohio corporation (“Lessee”).`)
       await loadingTask.destroy();
     });
 
+    it("gets text content with a rised text", async function () {
+      const loadingTask = getDocument(buildGetDocumentParams("issue16221.pdf"));
+      const pdfDoc = await loadingTask.promise;
+      const pdfPage = await pdfDoc.getPage(1);
+      const { items } = await pdfPage.getTextContent();
+
+      expect(items.map(i => i.str)).toEqual(["Hello ", "World"]);
+
+      await loadingTask.destroy();
+    });
+
     it("gets empty structure tree", async function () {
       const tree = await page.getStructTree();