make titles in markdown not be splited with following content (#2971)

### What problem does this PR solve?

#2970
### Type of change

- [ ] Bug Fix (non-breaking change which fixes an issue)
- [x] New Feature (non-breaking change which adds functionality)

rag/app/manual.py

@@ -67,9 +67,11 @@ class Pdf(PdfParser):
         return [(b["text"], b.get("layout_no", ""), self.get_position(b, zoomin))
                 for i, b in enumerate(self.boxes)], tbls
 
+
 class Docx(DocxParser):
     def __init__(self):
         pass
+
     def get_picture(self, document, paragraph):
         img = paragraph._element.xpath('.//pic:pic')
         if not img:
@@ -80,6 +82,7 @@ class Docx(DocxParser):
         image = related_part.image
         image = Image.open(BytesIO(image.blob))
         return image
+
     def concat_img(self, img1, img2):
         if img1 and not img2:
             return img1
@@ -160,6 +163,7 @@ class Docx(DocxParser):
             tbls.append(((None, html), ""))
         return ti_list, tbls
 
+
 def chunk(filename, binary=None, from_page=0, to_page=100000,
           lang="Chinese", callback=None, **kwargs):
     """
@@ -244,6 +248,7 @@ def chunk(filename, binary=None, from_page=0, to_page=100000,
         res = tokenize_table(tbls, doc, eng)
         res.extend(tokenize_chunks(chunks, doc, eng, pdf_parser))
         return res
+
     if re.search(r"\.docx$", filename, re.IGNORECASE):
         docx_parser = Docx()
         ti_list, tbls = docx_parser(filename, binary,
@@ -259,8 +264,6 @@ def chunk(filename, binary=None, from_page=0, to_page=100000,
         raise NotImplementedError("file type not supported yet(pdf and docx supported)")
     
 
-
-
 if __name__ == "__main__":
     import sys
 

rag/app/naive.py

@@ -168,8 +168,12 @@ class Markdown(MarkdownParser):
                 sections.append((sec[:int(len(sec) / 2)], ""))
                 sections.append((sec[int(len(sec) / 2):], ""))
             else:
-                sections.append((sec, ""))
-        print(tables)
+                if sections and sections[-1][0].strip().find("#") == 0:
+                    sec_, _ = sections.pop(-1)
+                    sections.append((sec_+"\n"+sec, ""))
+                else:
+                    sections.append((sec, ""))
+
         for table in tables:
             tbls.append(((None, markdown(table, extensions=['markdown.extensions.tables'])), ""))
         return sections, tbls

rag/app/qa.py

@@ -393,6 +393,7 @@ def chunk(filename, binary=None, lang="Chinese", callback=None, **kwargs):
             if sum_question:
                 res.append(beAdoc(deepcopy(doc), sum_question, markdown(last_answer, extensions=['markdown.extensions.tables']), eng))
         return res
+
     elif re.search(r"\.docx$", filename, re.IGNORECASE):
         docx_parser = Docx()
         qai_list, tbls = docx_parser(filename, binary,