docling-project/docling

Extracting user comments from office documents like (.docx, .pptx, .xlsx) and embedding them into output format (.md, .html).

Open

#2,119 opened on 2025年8月22日

GitHub で見る
 (1 comment) (0 reactions) (0 assignees)Python (59,751 stars) (4,140 forks)batch import
enhancementgood first issuepptxxlsx

説明

Requested feature

Add functionality to extract user comments and discussion threads from Office documents (.docx, .pptx, .xlsx) and embed them into the output formats (.md, .html).

Feature Discription

When users collaborate on Office documents, they often add comments to review content, ask questions, or have discussions. This commentary is valuable context that is directly related to specific passages in the document.

Currently, docling processes the main content but ignores these embedded comments. By discarding this information, the output loses important context, discussions

Proposed Solution

  • Parse Comments: Identify and extract all comments and their corresponding reply threads from the source .docx, .pptx, and .xlsx files.
  • Associate Content: Link each comment thread to the specific text or element it refers to in the document.
  • Embed in Output: Intelligently embed the extracted comments into the final .md or .html output. For example, they could be formatted as footnotes, side notes, or blockquotes adjacent to the relevant content.

example : Input Documnet:

Output:

Python Programming Language Overview

Python is a powerful programming language used for data analysis and web development. Many developers choose {{COMMENT_START:thread_id = 0}}Python{{COMMENT_END:thread_id =0}} because of its simplicity and readability. The Python community is very active and supportive.

{{COMMENT_START:thread_id =1}}Python offers excellent libraries for machine learning{{COMMENT_END:thread_id =2}} and artificial intelligence. Companies like Google and Netflix use Python extensively in their operations. Learning Python can open many career opportunities in technology.

Comments

  • Thread_id : 0: Passage: "Python"

    • Comment 1 by Omkar Musale (2025-07-21 11:01:52 UTC) Why Python specifically over other languages?

  • Thread_id : 1: Passage: "Python offers excellent libraries for machine learning"

    • Comment 1 by Omkar Musale (2025-07-21 11:02:11 UTC) Which libraries are you referring to?

    • Reply 2 by Omkar Musale (2025-07-21 11:02:19 UTC) Are these open source libraries?

    • Reply 3 by Omkar Musale (2025-07-21 11:02:25 UTC) not sure

コントリビューターガイド

技術スタック
python
領域
backenddata
Issue 種別
feature
難度新しい貢献者にとっての実装難度です。1 は非常に小さな変更、5 は専門的な作業です。
3
推定時間経験ある貢献者が調査、実装、テスト、pull request 準備にかけるおおよその時間範囲です。
half day
活動状況Issue が今どれだけ取り組みやすいかを示します。新しい、活発、古い、ブロック中、またはメンテナー入力待ちなどです。
fresh
明確さ期待される変更、受け入れ条件、次の手順がどれだけ明確かを示します。
clear
前提条件
Pythonpython docxpython pptxopenpyxl
初心者向け度初回貢献者にどれだけ取り組みやすいかを 1-100 で推定したスコアです。
50
調査方針
Start by examining the existing document parsers in the docling codebase (likely under a 'parsers' directory). For each format, understand how the XML structure stores comments. Use python docx for .docx, python pptx for .pptx, and openpyxl for .xlsx. Look for existing comment extraction code or similar features. Then design a method to link comments to specific text positions and embed them in Markdown/HTML output as shown in the example.