’)
processed_html = processed_html.replace(’

’)
# Use a placeholder class to identify vc_column_text divs for later attribute removal
processed_html = re.sub(r’\

’)

# Unescape HTML entities for attributes (e.g., ” for ”)
processed_html = processed_html.replace(’”’, ’”’).replace(’″’, ’”’)

# Remove WPBakery shortcodes for images as per rules (no valid URLs provided in source HTML)
processed_html = re.sub(r’\[vc_single_image[^\]]*\]’, ”, processed_html)
processed_html = re.sub(r’\[vc_media_grid[^\]]*\]’, ”, processed_html)

soup = BeautifulSoup(processed_html, ’html.parser’)

# Task 1 – CLEAN HTML FORMATTING

# 1.1 Remove all Elementor/WPBakery classes, inline styles, and specific attributes
for tag in soup.find_all(True):
attrs_to_remove = [] for attr in tag.attrs:
if attr.startswith((’elementor-’, ’vc_’)) or attr in [’class’, ’style’, ’gap’, ’width’, ’alignment’, ’img_size’, ’onclick’, ’include’, ’grid_id’]:
attrs_to_remove.append(attr)
for attr in attrs_to_remove:
del tag.attrs[attr]

# 1.2 Remove H1 tags -> convert to H2
for h1_tag in soup.find_all(’h1’):
h1_tag.name = ’h2’

# 1.3 Remove empty tags and unnecessary nesting recursively
def remove_empty_tags_and_unwrap_recursively(current_soup):
changed = True
while changed:
changed = False
for tag in current_soup.find_all(True):
# Remove empty tags that are not self-closing and don’t contain significant content
if not tag.contents and not tag.is_empty_element and tag.name not in [’img’, ’br’, ’hr’, ’iframe’]:
tag.decompose()
changed = True
continue

# Unwrap divs that have no attributes and contain only one child or only whitespace
if tag.name == ’div’ and not tag.attrs:
if len(tag.contents) == 1 and isinstance(tag.contents[0], Tag):
if tag.contents[0].name not in [’div’, ’p’, ’ul’, ’ol’, ’table’]: # Avoid unwrapping content that needs parent block
tag.unwrap()
changed = True
elif not tag.get_text(strip=True) and not any(isinstance(c, Tag) for c in tag.contents): # Only whitespace or empty
tag.decompose()
changed = True
elif all(isinstance(c, NavigableString) and not c.strip() for c in tag.contents): # Only whitespace NavigableStrings
tag.decompose()
changed = True
return current_soup

soup = remove_empty_tags_and_unwrap_recursively(soup)

# Re-structure loose text content into paragraphs
final_body_contents = [] current_p_text = []

for content in soup.body.contents:
if isinstance(content, NavigableString) and content.strip():
current_p_text.append(content.strip())
elif isinstance(content, Tag):
if current_p_text:
new_p = soup.new_tag(”p”)
new_p.string = ” ”.join(current_p_text)
final_body_contents.append(new_p)
current_p_text = [] final_body_contents.append(content)
else: # Handle whitespace NavigableString
pass

if current_p_text:
new_p = soup.new_tag(”p”)
new_p.string = ” ”.join(current_p_text)
final_body_contents.append(new_p)

# Replace body contents with the newly structured list
soup.body.clear()
for item in final_body_contents:
soup.body.append(item)

# Create a new soup for the final output structure (summary, ToC, then content)
final_output_soup = BeautifulSoup(””, ”html.parser”)

# Task 3 – ADD SUMMARY AND KEY TAKEAWAYS (two-column layout)
summary_text = ”Tämä sivu esittelee erilaisia kattoremonttiprojekteja kuvagallerian muodossa. Se tarjoaa esimerkkejä huopakattoremonteista, piipunpellityksistä ja vesikourujen asennuksista, näyttäen työn kulkua ennen ja jälkeen -kuvien avulla.”
# Article text content is very short (under 300 words), so Avainkohdat will be skipped.

article_intro_div = final_output_soup.new_tag(”div”, class_=”article-intro”)
article_summary_div = final_output_soup.new_tag(”div”, class_=”article-summary”)
h2_summary = final_output_soup.new_tag(”h2”)
h2_summary.append(”📋 Tiivistelmä”)
article_summary_div.append(h2_summary)
p_summary = final_output_soup.new_tag(”p”)
p_summary.append(summary_text)
article_summary_div.append(p_summary)
article_intro_div.append(article_summary_div)
# Avainkohdat section is intentionally omitted as per rules for short articles.

final_output_soup.append(article_intro_div)

# Task 4 & 5 – ADD TABLE OF CONTENTS AND IDS TO HEADINGS
toc_data = [] headings = soup.find_all([’h2’, ’h3’])
previous_h2 = None

for heading in headings:
heading_text = heading.get_text().strip()
if not heading_text: # Skip empty headings
continue

heading_id = generate_id(heading_text)
heading[’id’] = heading_id

if heading.name == ’h2’:
current_h2_entry = {’text’: heading_text, ’id’: heading_id, ’subheadings’: []}
toc_data.append(current_h2_entry)
previous_h2 = current_h2_entry
elif heading.name == ’h3’ and previous_h2:
previous_h2[’subheadings’].append({’text’: heading_text, ’id’: heading_id})
elif heading.name == ’h3’ and not previous_h2: # Handle H3s that might appear before any H2
toc_data.append({’text’: heading_text, ’id’: heading_id, ’subheadings’: []})

if toc_data:
article_toc_div = final_output_soup.new_tag(”div”, class_=”article-toc”)
h2_toc = final_output_soup.new_tag(”h2”)
h2_toc.append(”📑 Sisällysluettelo”)
article_toc_div.append(h2_toc)
ul_toc = final_output_soup.new_tag(”ul”)

for entry in toc_data:
li_h2 = final_output_soup.new_tag(”li”)
a_h2 = final_output_soup.new_tag(”a”, href=”#” + entry[’id’])
a_h2.append(entry[’text’])
li_h2.append(a_h2)

if entry[’subheadings’]:
ul_h3 = final_output_soup.new_tag(”ul”)
for sub_entry in entry[’subheadings’]:
li_h3 = final_output_soup.new_tag(”li”)
a_h3 = final_output_soup.new_tag(”a”, href=”#” + sub_entry[’id’])
a_h3.append(sub_entry[’text’])
li_h3.append(a_h3)
ul_h3.append(li_h3)
li_h2.append(ul_h3)
ul_toc.append(li_h2)
article_toc_div.append(ul_toc)
final_output_soup.append(article_toc_div)

# Task 6 – IMPROVE READABILITY
# Bolding important key phrases
for h_tag in soup.find_all([’h2’, ’h3′]): # Only headings have content in this example
if ”Kattoremontti Pro hoitaa homman” in h_tag.get_text():
for content_node in list(h_tag.contents):
if isinstance(content_node, NavigableString):
text_parts = re.split(r'(Kattoremontti Pro)’, str(content_node))
for part in text_parts:
if part == ”Kattoremontti Pro”:
strong_tag = soup.new_tag(”strong”)
strong_tag.string = part
content_node.insert_before(strong_tag)
else:
content_node.insert_before(part)
content_node.decompose() # Remove original NavigableString
break

# Append all original content (now with IDs and bolding) to the final_output_soup
for content_element in soup.body.contents:
final_output_soup.append(content_element)

# Final cleanup pass for the *entire* final_output_soup to remove any residual empty elements
final_output_soup = remove_empty_tags_and_unwrap_recursively(final_output_soup)

return str(final_output_soup)

# — Original Content —
original_html_content = ”””

cleaned_html_output = clean_html_content(original_html_content)

# Task 7 – GENERATE METADATA
metadata_final = {
”category”: ”Kattoremontti”,
”subcategory”: ”Referenssit”,
”keywords”: [”kattoremontti”, ”huopakattoremontti”, ”piipunpellitys”, ”vesikourut”, ”kattohuolto”, ”kattotyöt”],
”isLocalSeo”: False
}
metadata_json_output = json.dumps(metadata_final, ensure_ascii=False)

# Final output format: HTML content followed by metadata JSON
final_output = cleaned_html_output + ”\n”

print(final_output)
”`
”`html

Kattoremontti galleria

Huopakattoremontti

Huopakattoremontti ennen

Huopakattoremontti jälkeen

Näin huopakattoremontti meillä tehdään

Joskus kattoremontti voi yllättää, mutta Kattoremontti Pro hoitaa homman.

Piipunpellitys on ammattimiehen puuhaa

Piipunpellitys ennen

Piipunpellitys jälkeen

Näin uudet vesikourut asentuvat paikoilleen

”`