fenix
/
cpdf-source
mirror of https://github.com/johnwhitington/cpdf-source.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Avoid tagging if no process_struct_tree

This commit is contained in:
John Whitington 2024-10-04 14:42:31 +01:00
parent e3beeae2a0
commit 3cf1f914c1
1 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
cpdftexttopdf.ml
View File

@ -86,7 +86,7 @@ let typeset ~process_struct_tree ?subformat ?title ~papersize ~font ~fontsize te
let instrs = of_utf8_with_newlines fontpack fontsize (Pdfio.string_of_bytes text) in let instrs = of_utf8_with_newlines fontpack fontsize (Pdfio.string_of_bytes text) in
(*flprint (Cpdftype.to_string instrs); (*flprint (Cpdftype.to_string instrs);
flprint "------------------------------";*) flprint "------------------------------";*)
let tagged = tag_paragraphs instrs in let tagged = if process_struct_tree then tag_paragraphs instrs else instrs in
(*flprint (Cpdftype.to_string tagged);*) (*flprint (Cpdftype.to_string tagged);*)
let margin = Pdfunits.points (Pdfpaper.width papersize) (Pdfpaper.unit papersize) /. 15. in let margin = Pdfunits.points (Pdfpaper.width papersize) (Pdfpaper.unit papersize) /. 15. in
let instrs = let instrs =
@ -122,12 +122,12 @@ let typeset ~process_struct_tree ?subformat ?title ~papersize ~font ~fontsize te
| [] -> rev (map rev a) | [] -> rev (map rev a)
in in
let nodes = find_nodes [[]] tagtriples in let nodes = find_nodes [[]] tagtriples in
Printf.printf "Paragraphs and their page and MCIDs\n"; (*Printf.printf "Paragraphs and their page and MCIDs\n";
iter iter
(fun parts_of_para -> (fun parts_of_para ->
Printf.printf "Paragraph:\n"; Printf.printf "Paragraph:\n";
iter (fun (para, page, mcid) -> Printf.printf "Para %i, Page %i, MCID %i\n" para page mcid) parts_of_para) iter (fun (para, page, mcid) -> Printf.printf "Para %i, Page %i, MCID %i\n" para page mcid) parts_of_para)
nodes; nodes;*)
let pages = let pages =
map2 map2
(fun pn p -> if process_struct_tree then {p with Pdfpage.rest = Pdf.add_dict_entry p.Pdfpage.rest "/StructParents" (Pdf.Integer pn)} else p) (fun pn p -> if process_struct_tree then {p with Pdfpage.rest = Pdf.add_dict_entry p.Pdfpage.rest "/StructParents" (Pdf.Integer pn)} else p)