Comment convertir des PDF en Markdown à l'aide de PyMuPDFM et de son évaluation

PyMuPDF4LLM is a library designed to convert PDFs into Markdown format. Here, I’ll share my experience testing this library.

Installation

Start by installing the library using the following command:

pip install pymupdf4llm

Usage

The basic usage is quite simple, requiring just three lines of code to convert a PDF to Markdown:

import pymupdf4llm
md_text = pymupdf4llm.to_markdown("input.pdf")
print(md_text)

You can specify arguments to adjust how content is extracted.

Extracting Text by Page

By default, the entire PDF is converted into a single text output. However, you can extract text page by page by specifying page_chunks=True.

md_text = pymupdf4llm.to_markdown("input.pdf", page_chunks=True)

Extracting Images

To extract images as files, use the write_images=True option:

md_text = pymupdf4llm.to_markdown("input.pdf", write_images=True)

It’s also possible to embed images directly in the Markdown using base64 encoding:

md_text = pymupdf4llm.to_markdown("input.pdf", embed_images=True)

Evaluation of Conversion Results

For testing, various PDFs with different Markdown elements were used.

Header Conversion

Headers are correctly converted into Markdown format. Here is a portion of the result:

# Sample Markdown Guide

This is a sample markdown file that includes various features for quick reference.

## 1. Headers

...

## 3. Lists

Bold and Italic Text

Bold and italic formatting is also properly converted:

**Bold: **Bold Text****

_Italic: *Italic Text*_

**_Bold and Italic: ***Bold and Italic***_**

List Conversion

Ordered lists at the first level are converted without issues, but nested lists and unordered lists are not accurately converted.

## 3. Lists

### Unordered List

Item 1

Item 2

Sub-item 1

Sub-item 2

### Ordered List

1. First item

2. Second item

1. Sub-item A

2. Sub-item B

Link Conversion

The URLs of links are extracted, but the entire line containing the link becomes a hyperlink, deviating from the original format.

## 4. Links and Images

[You can add links using [Link Text](URL).](https://www.example.com/)

Image Extraction

Images are not extracted by default but can be saved locally with write_images=True.

md_text = pymupdf4llm.to_markdown("input.pdf", write_images=True)

The saved images are then referenced in the Markdown as follows:

<p>### Image Example</p>

<p>![](input.pdf-1-0.png)</p>

Table Conversion

Simple tables without vertical borders are not accurately converted (likely because ambiguous column boundaries result in tables being treated as plain text).

<p>## 5. Tables</p>

<p>**Column 1** **Column 2** **Column 3**</p>

<p>Row 1 Data A Data B</p>

<p>Row 2 Data C Data D</p>

Code Conversion

Code blocks are correctly converted, but language specification (e.g., python) is not retained. Inline code conversion also has issues.

<p>## 6. Code</p>

<p>### Inline Code</p>

<p>Use backticks for inline code: print("Hello, world!")</p>

<p>### Code Block</p>

<p>Use triple backticks for code blocks:</p>

<p>```<br>
def greet(name):<br>
  return f"Hello, {name}!"<br>
print(greet("Markdown"))<br>
```</p>

Multi-Line Text

For multi-line text, the line breaks are preserved as they appear in the original PDF.

<p>Markdown is a lightweight and versatile markup language favored by developers, writers, and bloggers alike</p>

<p>due to its simplicity in formatting text, enabling users to create readable and well-structured documents—</p>

<p>whether for documentation, blog posts, or articles—without the complexity of HTML, while also offering the</p>

<p>ability to convert content seamlessly into other formats like HTML, PDF, and even slideshows, making it an</p>

<p>ideal choice for projects that require both clarity and flexibility in presentation.</p>

Conclusion

Despite challenges in accurately converting lists and links, PyMuPDF4LLM is a useful tool for converting PDFs to Markdown. It can work locally without the need for external language models, making it suitable for environments where internet access is unavailable.

Ce qui précède est le contenu détaillé de. pour plus d'informations, suivez d'autres articles connexes sur le site Web de PHP en chinois!