feat/Infer the hierarchical heading/title levels such as H1, H2, H3, H4 for PDFs

**Is your feature request related to a problem? Please describe.** This was supposedly raised and addressed as part of #1354 But I can't seem to get any metadata regarding the inferred heading of any content I test with. All `Title` elements all have the same level at the moment which isn't an accurate representation of the parsed document. **Describe the solution you'd like** The parser should be able to infer PDF hierarchy in a couple ways: * Through the document outline, e.g. example PDF linked on https://pdfa.org/techniques-for-accessible-pdf/headings-with-different-levels-correctly-tagged/UA1_Tpdf-H_03/ * Calculate and infer the heading size relative to the size of the page (and other headings within the same page/document). **Describe alternatives you've considered** N/A **Additional context**