Redundent data from HTML source are included

[kehiy]

Question

I have simple script that used docling to parse this webpage:

https://ramzinex.com/help/register-in-ramzinex

The issue is when I parse the document it contains ol tag and footer details as well. How can I exclude them?

The ol tag info:
صرافی رمزینکس
راهنما
ثبت نام و احراز هویت

The footer info:

This whole section will be included in final document.
Also, when I use hybrid chunker to chunk them, these are still included.

Is there any config to exclude redundant stuff? from links, PDFs or anything else?

[dosubot]

Hi @kehiy! I'm Dosu and I’m helping the docling team.

Docling doesn’t have a built-in config to exclude specific HTML tags like ol or footer during parsing or chunking. The two main ways to handle this are:

1. Preprocess the HTML: Before passing the HTML to Docling, use a tool like BeautifulSoup to remove unwanted tags. For example:

from bs4 import BeautifulSoup

with open("input.html", "r", encoding="utf-8") as f:
    soup = BeautifulSoup(f, "html.parser")
for tag in soup.find_all(["ol", "footer"]):
    tag.decompose()
clean_html = str(soup)
# Pass clean_html to Docling

2. Custom Serializer: Docling supports custom serializers, which let you filter out specific content types after parsing. You can define a serializer that skips elements with certain labels or tags. This is the recommended approach for excluding things like formulas, tables, or other unwanted content in the final output. See the discussion here.

There’s no high-level config for this, so you’ll need to add a preprocessing step or implement a custom serializer in your workflow. Both approaches work with chunkers, including the hybrid chunker, since filtering happens before or after chunking as needed.

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}