docstruct 1.0.11

A package for representing documents as a tree of document, pages, paragraphs, lines, words, and characters

Docstruct

Overview

Docstruct is a package that parses the results of optical character recognition (OCR) algorithms, such as Tesseract (using the hOCR output) or Textract (AWS), into a tree structure. This tree structure allows for the visual representation of the document, with each node representing a document, page, paragraph, line, word, or character, along with its bounding box. The package also includes support for paragraph detection and text splitting that preserves logical units.

Installation

To install the Docstruct package, use the following pip command:

pip install docstruct

Usage

To use the Docstruct package, simply import it and call the parsing function on your OCR results. For example:

import docstruct

# Load the OCR results from a file
with open("ocr_results.hocr", "r") as f:
    ocr_results = f.read()

# Parse the OCR results into a tree structure
document = docstruct.parse(ocr_results)

Once you have a `document` object, you can access its individual elements, such as pages, paragraphs, and lines, using standard list indexing.

For example, to access the first page of the document, you can use the following code:

first_page = document[0]

You can also visualize each element in the tree structure by calling the `show` method on an element. This will display a visual representation of the bounding boxes for each object.

For example:

first_page.show()

For more information on how to use the Docstruct package, refer to the documentation and example code provided with the package.

Contributions

Contributions to the Docstruct package are always welcome. If you have a bug fix or a new feature, feel free to create a pull request on the GitHub repository.

License

The Docstruct package is licensed under the MIT License.