fix marking paragraphs as headings

src/html_preprocessor.py

@@ -5,6 +5,7 @@ import re
 from shutil import copyfile
 
 from bs4 import BeautifulSoup, NavigableString
+
 from config import LawCartaConfig, BookLogger, BookApiWrapper
 
 
@@ -388,6 +389,16 @@ class HTMLPreprocessor:
         for div in divs:
             div.unwrap()
 
+    def _check_parent_link_exist_in_toc(self, tag_with_link):
+        toc_links = []
+        for a_tag in tag_with_link.find_all('a'):
+            link_name = a_tag.attrs['name']
+            toc_item = self.body_tag.find("a", {'href': '#' + link_name})
+            if toc_item:
+                toc_links.append(toc_item)
+
+        return len(toc_links) > 0
+
     def _process_toc_links(self):
         """
         Function to extract nodes which contains TOC links, remove links from file and detect headers.
@@ -399,12 +410,9 @@ class HTMLPreprocessor:
             if re.search(r"^h\d$", tag.name):
                 tag.a.unwrap()
                 # outline_level = tag.name[-1]  # TODO: add prediction of the outline level
-            # TODO: escape from recounting paragraphs every time
             elif tag.name == "p":
-                link_name = tag.a.attrs['name']
-                toc_item = self.body_tag.find("a", {'href': '#' + link_name})
-                # TODO: if it is needed, check existence of the link in toc
-                if tag in self.body_tag.find_all("p"):
+                exist_in_toc = self._check_parent_link_exist_in_toc(tag)
+                if tag in self.body_tag.find_all("p") and exist_in_toc:
                     new_tag = BeautifulSoup(features="lxml").new_tag("h" + outline_level)
                     text = tag.text
                     tag.replaceWith(new_tag)