fix: Fail to identify fonts (name and size) #629 (#630)

* fix: Fail to identify fonts (name and size) #629 * docs(fix): in sample code, get font from the Page rather than the PDF list, as font's IDs are different * added two tests to prove changes in #630 are working --------- Co-authored-by: Konrad Abicht <[email protected]>
smalot · Aug 11, 2023 · f3a5a3e · f3a5a3e
1 parent eef0263
commit f3a5a3e
Show file tree

Hide file tree

Showing 4 changed files with 46 additions and 6 deletions.
diff --git a/doc/Usage.md b/doc/Usage.md
@@ -76,8 +76,8 @@ $config->setDataTmFontInfoHasToBeIncluded(true);
 // use config and parse file
 $parser = new Smalot\PdfParser\Parser([], $config);
 $pdf = $parser->parseFile('document.pdf');
-
-$data = $pdf->getPages()[0]->getDataTm();
+$firstpage = $pdf->getPages()[0];
+$data = $firstpage->getDataTm();
 
 Array
 (
@@ -121,9 +121,8 @@ Text width should be calculated on text from dataTm to make sure all character w
 In next example we are using data from above.
 
 ```php
-$fonts = $pdf->getFonts();
 $font_id = $data[0][2]; //R7
-$font = $fonts[$font_id];
+$font = $firstpage->getFont($font_id);
 $text = $data[0][1];
 $width = $font->calculateTextWidth($text, $missing);
 ```

diff --git a/samples/bugs/Issue629.pdf b/samples/bugs/Issue629.pdf
diff --git a/src/Smalot/PdfParser/Page.php b/src/Smalot/PdfParser/Page.php
@@ -712,8 +712,6 @@ public function getDataTm(array $dataCommands = null): array
                     $Tl = $defaultTl;
                     $Tx = 0;
                     $Ty = 0;
-                    $fontId = $defaultFontId;
-                    $fontSize = $defaultFontSize;
                     break;
 
                     /*

diff --git a/tests/PHPUnit/Integration/PageTest.php b/tests/PHPUnit/Integration/PageTest.php
@@ -846,4 +846,47 @@ public function testIssue454(): void
         $textData = $page->getTextXY(67.5, 756.25);
         $this->assertStringContainsString('{signature:signer505906:Please+Sign+Here}', $textData[0][1]);
     }
+
+    /**
+     * Check that BT and ET do not reset the font.
+     *
+     * Data TM font info is included.
+     *
+     * @see https://github.com/smalot/pdfparser/pull/630
+     */
+    public function testIssue629WithDataTmFontInfo(): void
+    {
+        $config = new Config();
+        $config->setDataTmFontInfoHasToBeIncluded(true);
+
+        $filename = $this->rootDir.'/samples/bugs/Issue629.pdf';
+        $parser = $this->getParserInstance($config);
+        $document = $parser->parseFile($filename);
+        $pages = $document->getPages();
+        $page = end($pages);
+        $dataTm = $page->getDataTm();
+
+        $this->assertCount(4, $dataTm[0]);
+        $this->assertEquals('F2', $dataTm[0][2]);
+    }
+
+    /**
+     * Data TM font info is NOT included.
+     *
+     * @see https://github.com/smalot/pdfparser/pull/630
+     */
+    public function testIssue629WithoutDataTmFontInfo(): void
+    {
+        $config = new Config();
+
+        $filename = $this->rootDir.'/samples/bugs/Issue629.pdf';
+        $parser = $this->getParserInstance($config);
+        $document = $parser->parseFile($filename);
+        $pages = $document->getPages();
+        $page = end($pages);
+        $dataTm = $page->getDataTm();
+
+        $this->assertCount(2, $dataTm[0]);
+        $this->assertFalse(isset($dataTm[0][2]));
+    }
 }