Pdf subsampler

[rom1504]

https://pymupdf.readthedocs.io/en/latest/app1.html#performance

[rom1504]

https://github.com/sandorkonya/forms-dataset/blob/main/PDF_Info_retrieval.ipynb

[sandorkonya]

PdfSubsampler

import fitz
from urllib.parse import quote

class Subsampler:
    def __init__(self, timeout=5, maxlength=10**6):
        self.timeout = timeout
        self.maxlength = maxlength
    
    def __call__(self, r):
        """
        input: a request
        output: error,doc_size,doc_numberofpages,doc_form, characters, links,images,words,blocks
        """

        error,doc_size,doc_numberofpages,doc_form, characters, links,images,words,blocks = 0,0,0,0,0,0,0,0,0
        if r.status_code == 200:
            doc_size = r.headers["Content-Length"]
            if ((r.headers["Content-Type"] in ["application/pdf","stream/pdf"]) & (int(r.headers["Content-Length"]) < self.maxlength)):
                try:
                    doc = fitz.open(stream=r.content, filetype="pdf")
                    #gets the number of pages in the pdf
                    doc_numberofpages = len(doc)

                    #0 if no form element, otherwise the number of form elements in the pdf
                    doc_form = doc.is_form_pdf
                    doc_form = 0 if not doc_form else doc_form

                    #count together characters of readable text, links, images
                    for page in doc:
                        characters += len(page.get_text())
                        links += len(page.get_links())
                        images += len(page.get_images())
                        blocks += len(page.get_textpage().extractBLOCKS())
                        words += len(page.get_textpage().extractWORDS())

                except:
                    error = "OpenError"
            else:
                error = "HeaderError"
        else:
            error =  "StatusCodeError"

        return [error, quote(r.url, safe='/:?&') ,doc_numberofpages, doc_form, doc_size, characters, links,images,blocks,words]

call(self, r) where r is the request and not the file (since we wan to make some tests also on the request header).