71 lines
2.6 KiB
Plaintext
71 lines
2.6 KiB
Plaintext
Here's the pdf file that this storytest looks at: http://files/pdf/eg.pdf.
|
|
|
|
* It's worth reading this as it is a document that discusses the issues in accessing text from PDFs.
|
|
|
|
|''with PDF''|
|
|
|
|
|''open''|pdf/eg.pdf|
|
|
|
|
|''number of pages''|'''is'''|2|
|
|
|
|
|''select all pages''|
|
|
|
|
|'''show'''|''text''|
|
|
|
|
* We can check that the whole document contains some specific text:
|
|
|
|
|''text''|''contains''|PDF|
|
|
|
|
|''text''|''contains''|PDF Format|
|
|
|
|
|''text''|''contains''|PDF Format PDFs do not retain paragraph|
|
|
|
|
* We can use pattern matching if we want to ignore parts of the document:
|
|
|
|
|''text''|''matches''|PDF Format PDFs do not .* paragraph|
|
|
|
|
* The footer appears twice (and it is added in the middle of the main text):
|
|
|
|
|''text''|''matches''|Extracting Paragraphs from PDFs.*Extracting Paragraphs from PDFs|
|
|
|
|
* The flow of text from one column to the next is captured here (luckily, but not between pages):
|
|
|
|
|''text''|''matches''|If the space between that line and the previous one is larger than a space threshhold|
|
|
|
|
* We can use the paragraph structure to select a relative paragraph:
|
|
|
|
|''paragraph below heading''|pdfbox|'''contains'''|extracts the text elements|
|
|
|
|
|''paragraph after containing''|Limitations|'''contains'''|a simple approach|
|
|
|
|
* Show the text with breaks between the paragraphs, to make it easier to see the results:
|
|
|
|
|show|paragraphed text|
|
|
|
|
* We can match some of the paragraphs:
|
|
|
|
|paragraphs from|0|to|6|
|
|
|Extracting Paragraphs from PDF Files|
|
|
|PDF Format|
|
|
|PDFs do not retain paragraph, heading, etc information. Instead, a PDF encodes a rendered form of a document, rather like the individual characters that are rendered on a screen. A PDF file can be thought of as containing a sequence of pieces of information. Each piece of text is located at a particular (x,y,z) position, along with font information.|
|
|
|Depending on the application writing the PDF file, a word may be added as a single word, or it may be added as several substrings. Some applications tend to add each of the characters of a word separately.|
|
|
|Space characters are not (usually) added to the file, as they don't add anything to the rendering.|
|
|
|pdfbox|
|
|
|PDFDocument uses pdfbox to do much of the work. pdfbox extracts the text elements from a PDF file, taking account of columns of text and pages. pdfbox works out where to add spaces, and is configured by PDFDocument to add a space at the end of each line.|
|
|
|
|
|show|text|
|
|
|
|
|''select page''|2|
|
|
|
|
|show|paragraphed text|
|
|
|
|
* This shows the affect of customising the heuristics:
|
|
|
|
|''customise''|
|
|
|''basic paragraph drop''|4|
|
|
|''height space factor''|0.5|
|
|
|
|
|show|paragraphed text|
|
|
|
|
|''close PDF file''|
|