Skip to content

Latest commit

 

History

History
70 lines (52 loc) · 3.67 KB

definition.md

File metadata and controls

70 lines (52 loc) · 3.67 KB
Raw

Document Handling in appjsonify

Go back to top

appjsonify defines five classes (Token, Line, Paragraph, Page, and Document) to structure a given document(s).

Token

This class holds a token (word) and its position and surface information such as its bounding box coordinates, font size and font name.

__init__

Inputs

  • token (str): A token string.
  • pos (tuple[int]): A tuple of bounding box information ($x_0$, $y_0$, $x_1$, $y_1$). The bounding box values range from 0 to 1000.
  • font_size (float): Font size obtained using pdfplumber.
  • font_name (str): Font name obtained using pdfplumber.
  • meta (dict): A meta dictionary used to save supplementary information.

Line

This class holds a line element, its position and surface information such as its bounding box coordinates, font size and font name, and a list of tokens. Line consists of a list of Token, i.e. a line instance must be formed with the corresponding Token instances.

__init__

Inputs

  • line (str): A line string.
  • pos (tuple[int]): A tuple of bounding box information ($x_0$, $y_0$, $x_1$, $y_1$). The bounding box values range from 0 to 1000.
  • font_size (float): Font size obtained using pdfplumber.
  • font_name (str): Font name obtained using pdfplumber.
  • tokens (list[Token]): A list of Token instances that form a line.
  • meta (dict): A meta dictionary used to save supplementary information.

Paragraph

This class holds a paragraph element, its position and surface information such as its bounding box coordinates, font size and font name, and a list of lines. Paragraph consists of a list of Line, i.e. a paragraph instance must be formed with the corresponding Line instances.

__init__

Inputs

  • paragraph (str): A paragraph string.
  • pos (tuple[int]): A tuple of bounding box information ($x_0$, $y_0$, $x_1$, $y_1$). The bounding box values range from 0 to 1000.
  • font_size (float): Font size obtained using pdfplumber.
  • font_name (str): Font name obtained using pdfplumber.
  • lines (list[Line]): A list of Line instances that form a paragraph.
  • meta (dict): A meta dictionary used to save supplementary information.

Page

This class holds lists of Paragraph, Line, and Token instances per PDF page, as well as its meta dictionary.

__init__

Inputs

  • paragraphs (list[Paragraph]): A list of Paragraph instances that form a Page.
  • lines (list[Line]): A list of Line instances that form a Page.
  • tokens (list[Token]): A list of Token instances that form a Page.
  • meta (dict): A meta dictionary used to save supplementary information.

Document

This class holds the path to an original PDF file, its Page instances, formatted Paragraph instances, and meta dictionary. This is the core class used in BaseRunner, and every processing module needs to take the list of Document: list[Document] as its first input parameter and return the processed list[Document] in the execute method.

__init__

Inputs

  • input_path (Path): A path to a PDF file.
  • pages (list[Page]): A list of Page instances.
  • formatted_paragraphs (list[Paragraph]): A list of Paragraph instances.
  • meta (dict): A meta dictionary used to save supplementary information such as footers and captions.