Feat: Refact pipeline (#13826)

### What problem does this PR solve?

### Type of change

- [x] New Feature (non-breaking change which adds functionality)
- [x] Refactoring

---------

Co-authored-by: Zhichang Yu <yuzhichang@gmail.com>
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>

deepdoc/parser/utils.py

@@ -14,6 +14,10 @@
 #  limitations under the License.
 #
 
+from io import BytesIO
+
+from pypdf import PdfReader as pdf2_read
+
 from rag.nlp import find_codec
 
 
@@ -30,3 +34,21 @@ def get_text(fnm: str, binary=None) -> str:
                     break
                 txt += line
     return txt
+
+
+def extract_pdf_outlines(source):
+    try:
+        with pdf2_read(source if isinstance(source, str) else BytesIO(source)) as pdf:
+            outlines = []
+
+            def dfs(nodes, depth):
+                for node in nodes:
+                    if isinstance(node, list):
+                        dfs(node, depth + 1)
+                    else:
+                        outlines.append((node["/Title"], depth, pdf.get_destination_page_number(node) + 1))
+
+            dfs(pdf.outline, 0)
+            return outlines
+    except Exception:
+        return []