def enrich_metadata(pdf: StructuredPdfDocument, source: Source):
"""
add some metadata to parsed PDF if possible
# add filename
# todo create document summary
# todo extract document-title from best titles
@return:
"""
# try to capture filename
try:
filename = Path(source.uri).name
pdf.update_metadata("filename", filename)
except Exception as e:
pass
def parse_pdf(self, source: Source) -> StructuredPdfDocument:
"""
Analysises and parses a PDF document from a given @Source containing its natural hierarchy.
@param source:
@return:
"""
# 1. iterate once through PDF and analyse style distribution
distribution = count_sizes(source.read())
size_mapper = PivotLogMapper(distribution)
style_annotator = StyleAnnotator(sizemapper=size_mapper,
style_info=distribution)
# 2. iterate second time trough pdf
# - annotate each paragraph with mapped Style
elements_with_style = style_annotator.process(
source.read(override_la_params=LAParams(
boxes_flow=None,
detect_vertical=False,
line_margin=distribution.line_margin)))
# - create nested document structure on the fly
structured_elements = self.create_hierarchy(elements_with_style,
distribution)
# 3. create wrapped document and capture some metadata
pdf_document = StructuredPdfDocument(elements=structured_elements,
style_info=distribution)
enrich_metadata(pdf_document, source)
return pdf_document
def test_full_content(self):
with open(str(Path("resources/parsed/interview_cheatsheet.json").absolute()), "r") as fp:
json_string = json.load(fp)
document = StructuredPdfDocument.from_json(json_string)
text = document.text
expected_newline_merged_subsections_excerpt = "Greedy Algorithm\nDefinition:\nAn algorithm that, while"
self.assertTrue(expected_newline_merged_subsections_excerpt in text)
def test_print_json_string(self):
printer = JsonStringPrinter()
jsonString = printer.print(self.testDocument)
decoded_document = StructuredPdfDocument.from_json(
json.loads(jsonString))
self.assertEqual(self.testDocument.elements[1].heading.text,
decoded_document.elements[1].heading.text)
self.assertEqual(self.testDocument.elements[-1].heading.text,
decoded_document.elements[-1].heading.text)
self.assertEqual("Array",
decoded_document.elements[5].children[0].heading.text)
self.assertEqual(
"Time Complexity:",
decoded_document.elements[5].children[0].children[2].heading.text)
def test_print_json_file(self):
printer = JsonFilePrinter()
file_path = Path("resources/parsed/interview_cheatsheet.json")
printer.print(self.testDocument, file_path=str(file_path.absolute()))
with open(file_path, "r") as file:
decoded_document = StructuredPdfDocument.from_json(json.load(file))
self.assertEqual(self.testDocument.elements[1].heading.text,
decoded_document.elements[1].heading.text)
self.assertEqual(self.testDocument.elements[-1].heading.text,
decoded_document.elements[-1].heading.text)
self.assertEqual(
"Array", decoded_document.elements[5].children[0].heading.text)
self.assertEqual(
"Time Complexity:", decoded_document.elements[5].children[0].
children[2].heading.text)
# from pdfstructure.hierarchy.parser import HierarchyParser
# from pdfstructure.source import FileSource
# import pathlib
# path = "./Nurse.pdf"
# parser = HierarchyParser()
# source = FileSource(path)
# document = parser.parse_pdf(source)
# from pdfstructure.printer import JsonFilePrinter
# printer = JsonFilePrinter()
# file_path = pathlib.Path("test-pdf.json")
# printer.print(document, file_path=str(file_path.absolute()))
from pdfstructure.model.document import StructuredPdfDocument
file = './test-pdf.json'
jsonString = json.load(test-pdf.json)
document = StructuredPdfDocument.from_json(jsonString)