Mercurial > hg > python
view pdfCrawl.py @ 45:7d4da4e72d37
fix argv handling
author | Henry S. Thompson <ht@inf.ed.ac.uk> |
---|---|
date | Tue, 05 Jul 2022 10:22:50 +0100 |
parents | 2d7c91f89f6b |
children |
line wrap: on
line source
import PyPDF2 as pyPdf, sys f = open(sys.argv[1],'rb') pdf = pyPdf.PdfFileReader(f) pgs = pdf.getNumPages() key = '/Annots' uri = '/URI' ank = '/A' #print pdf.getNamedDestinations() for pg in range(pgs): print '#',pg p = pdf.getPage(pg) o = p.getObject() print o if o.has_key(key): ann = o[key] print key,len(ann),ann i=0 for a in ann: u = a.getObject() if u[ank].has_key(uri): try: print i,u[ank][uri] except UnicodeEncodeError: print i,map(ord,u[ank][uri]) i+=1