Alexandre created PDFBOX-3804:
---------------------------------
Summary: Detect end of paragraphs
Key: PDFBOX-3804
URL: https://issues.apache.org/jira/browse/PDFBOX-3804
Project: PDFBox
Issue Type: Improvement
Components: Text extraction
Affects Versions: 2.0.6, 3.0.0
Reporter: Alexandre
Priority: Minor
Attachments: example.pdf
Hi,
To extract text by paragraphs is probably the most looking forward improvement
asked by PDFBox users.
*The current text extraction approach detects correctly end of lines. But it
does not detect correctly end of paragraph.* Indeed, a carriage return
character is added after each new line.
*What is a paragraph ?* A paragraph is a text that contain one or several
sentences. It can start by a tabulation but this is not mandatory. In a
paragraph, there is one or more lines but there is no carriage return (except
the one at the very end). A paragraph can end before the very end of a line,
but some paragraphs end at the very end.
*So, the last line of a paragraph ends before reaching the very end of the
line.* +And we could use that pattern detect properly paragraphs.+
In my opinion, the algorithm never adds carriage return except when a end of
paragraph is detect.
In my opinion, the algorithm should use the following information:
(*) the width of the block containing the paragraph ;
(*) detect the padding and margin of that block ;
(*) precomputed width of the next word.
The width of a block is either the width of the pdf minus left and right
margins if the pdf has one column. Or the width of a block is the width of a
column if the pdf has two columns for example.
The algorithm runs on every character and when it reach the last character of a
line, it pre computes the next line first word to have it's width. If this word
fits in the previous line, then we conclude we have a end of paragraph.
Best,
A.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
