fix: When parsing the bold content in PDF, the result is duplicated. (#1729)

### What problem does this PR solve?

_fix: When parsing the bold content in PDF, the result is duplicated._

the detail: [When using OCR to recognize Chinese titles, the structure
appears to be
duplicated](https://github.com/infiniflow/ragflow/issues/1718)

### Type of change

- [x] Bug Fix (non-breaking change which fixes an issue)

deepdoc/parser/pdf_parser.py

@@ -952,7 +952,7 @@ class RAGFlowPdfParser:
                 fnm, str) else pdfplumber.open(BytesIO(fnm))
             self.page_images = [p.to_image(resolution=72 * zoomin).annotated for i, p in
                                 enumerate(self.pdf.pages[page_from:page_to])]
-            self.page_chars = [[{**c, 'top': max(0, c['top'] - 10), 'bottom': max(0, c['bottom'] - 10)} for c in page.chars if self._has_color(c)] for page in
+            self.page_chars = [[{**c, 'top': max(0, c['top'] - 10), 'bottom': max(0, c['bottom'] - 10)} for c in page.dedupe_chars().chars if self._has_color(c)] for page in
                                self.pdf.pages[page_from:page_to]]
             self.total_page = len(self.pdf.pages)
         except Exception as e: