2021-12-07 00:55:46 +01:00
|
|
|
open Pdfutil
|
2024-10-02 15:20:51 +02:00
|
|
|
open Cpdferror
|
2021-12-07 00:55:46 +01:00
|
|
|
|
2024-02-27 16:57:31 +01:00
|
|
|
let of_utf8_with_newlines fontpack fontsize t =
|
2021-12-07 00:55:46 +01:00
|
|
|
let items = ref [] in
|
2023-06-13 21:23:59 +02:00
|
|
|
let currfont = ref ~-1 in
|
2022-09-21 16:21:57 +02:00
|
|
|
let codepoints = Pdftext.codepoints_of_utf8 t in
|
2022-10-20 15:28:14 +02:00
|
|
|
let currtext = ref [] in
|
|
|
|
let process_codepoints cs =
|
|
|
|
iter
|
2022-09-21 16:21:57 +02:00
|
|
|
(fun u ->
|
2022-10-19 15:47:20 +02:00
|
|
|
match Cpdfembed.get_char fontpack u with
|
2022-10-20 15:28:14 +02:00
|
|
|
| Some (c, n, f) ->
|
|
|
|
begin if n <> !currfont then
|
|
|
|
begin
|
|
|
|
if !currtext <> [] then items := Cpdftype.Text (rev !currtext)::!items;
|
|
|
|
currtext := [];
|
|
|
|
currfont := n;
|
2023-07-20 14:35:06 +02:00
|
|
|
items := Cpdftype.Font (string_of_int n, f, fontsize)::!items;
|
2022-10-20 15:28:14 +02:00
|
|
|
currtext := char_of_int c::!currtext;
|
|
|
|
end
|
|
|
|
else
|
|
|
|
currtext := char_of_int c::!currtext
|
|
|
|
end
|
|
|
|
| None -> Printf.printf "No glyph for unicode U+%04X in this font\n" u)
|
|
|
|
cs;
|
|
|
|
items := Cpdftype.Text (rev !currtext)::!items
|
2022-09-21 16:21:57 +02:00
|
|
|
in
|
2022-10-20 15:28:14 +02:00
|
|
|
let buf = ref [] in
|
2022-09-21 16:21:57 +02:00
|
|
|
List.iter
|
2021-12-07 00:55:46 +01:00
|
|
|
(function
|
2022-09-21 16:21:57 +02:00
|
|
|
| 10 (*'\n'*) ->
|
|
|
|
let c = rev !buf in
|
2022-10-20 15:28:14 +02:00
|
|
|
if c <> [] then process_codepoints c;
|
2021-12-07 00:55:46 +01:00
|
|
|
items := Cpdftype.NewLine::!items;
|
2023-06-27 16:54:15 +02:00
|
|
|
currtext := [];
|
2022-09-21 16:21:57 +02:00
|
|
|
buf := []
|
|
|
|
| 13 (*'\r'*) -> ()
|
2021-12-07 00:55:46 +01:00
|
|
|
| x ->
|
2022-09-21 16:21:57 +02:00
|
|
|
buf := x::!buf)
|
|
|
|
codepoints;
|
2021-12-07 00:55:46 +01:00
|
|
|
(* Do last one *)
|
2022-09-21 16:21:57 +02:00
|
|
|
let c = rev !buf in
|
2022-10-20 15:28:14 +02:00
|
|
|
if c <> [] then process_codepoints c;
|
2022-09-21 16:21:57 +02:00
|
|
|
rev !items
|
2021-12-07 00:55:46 +01:00
|
|
|
|
2024-10-02 15:45:40 +02:00
|
|
|
(* Post process, adding Tag / EndTag around paragraphs *)
|
|
|
|
let rec tag_paragraphs = function
|
|
|
|
| Cpdftype.NewLine::Cpdftype.NewLine::t ->
|
2024-10-02 17:39:16 +02:00
|
|
|
Cpdftype.EndTag::Cpdftype.NewLine::Cpdftype.NewLine::Cpdftype.Tag ("P", 0)::tag_paragraphs t
|
2024-10-02 15:45:40 +02:00
|
|
|
| x::t -> x::tag_paragraphs t
|
|
|
|
| [] -> [Cpdftype.EndTag]
|
|
|
|
|
|
|
|
let tag_paragraphs l =
|
2024-10-02 17:39:16 +02:00
|
|
|
Cpdftype.Tag ("P", 0)::tag_paragraphs l
|
2024-10-02 15:45:40 +02:00
|
|
|
|
2024-10-02 14:27:57 +02:00
|
|
|
let typeset ~process_struct_tree ?subformat ?title ~papersize ~font ~fontsize text =
|
|
|
|
let process_struct_tree =
|
2024-10-03 15:47:57 +02:00
|
|
|
match process_struct_tree, subformat with
|
|
|
|
| _, (Some Cpdfua.PDFUA1 | Some Cpdfua.PDFUA2) | true, _ -> true
|
|
|
|
| _ -> false
|
2024-10-02 14:27:57 +02:00
|
|
|
in
|
2024-10-02 15:20:51 +02:00
|
|
|
let pdf, title =
|
|
|
|
match subformat with
|
|
|
|
| None -> Pdf.empty (), begin match title with Some x -> x | None -> "" end
|
|
|
|
| Some Cpdfua.PDFUA1 ->
|
|
|
|
begin match title with
|
|
|
|
| None -> error "no -title given"
|
|
|
|
| Some title -> Cpdfua.create_pdfua1 title papersize 1, title
|
|
|
|
end
|
|
|
|
| Some Cpdfua.PDFUA2 ->
|
|
|
|
begin match title with
|
|
|
|
| None -> error "no -title given"
|
|
|
|
| Some title -> Cpdfua.create_pdfua2 title papersize 1, title
|
|
|
|
end
|
|
|
|
in
|
2023-06-13 16:21:23 +02:00
|
|
|
let codepoints = setify (Pdftext.codepoints_of_utf8 (Pdfio.string_of_bytes text)) in
|
2023-06-13 15:35:01 +02:00
|
|
|
let fontpack =
|
2022-10-19 14:48:13 +02:00
|
|
|
match font with
|
2023-06-13 15:35:01 +02:00
|
|
|
| Cpdfembed.PreMadeFontPack t -> t
|
2022-10-19 14:48:13 +02:00
|
|
|
| Cpdfembed.EmbedInfo {fontfile; fontname; encoding} ->
|
2023-06-13 15:35:01 +02:00
|
|
|
Cpdfembed.embed_truetype pdf ~fontfile ~fontname ~codepoints ~encoding
|
|
|
|
| Cpdfembed.ExistingNamedFont ->
|
|
|
|
raise (Pdf.PDFError "Can't use existing named font for text-to-PDF")
|
2022-10-19 14:48:13 +02:00
|
|
|
in
|
2022-10-20 15:28:14 +02:00
|
|
|
let instrs = of_utf8_with_newlines fontpack fontsize (Pdfio.string_of_bytes text) in
|
2024-10-02 16:26:30 +02:00
|
|
|
(*flprint (Cpdftype.to_string instrs);
|
|
|
|
flprint "------------------------------";*)
|
2024-10-04 15:42:31 +02:00
|
|
|
let tagged = if process_struct_tree then tag_paragraphs instrs else instrs in
|
2024-10-02 16:26:30 +02:00
|
|
|
(*flprint (Cpdftype.to_string tagged);*)
|
2023-06-13 21:23:59 +02:00
|
|
|
let margin = Pdfunits.points (Pdfpaper.width papersize) (Pdfpaper.unit papersize) /. 15. in
|
2024-05-06 09:37:58 +02:00
|
|
|
let instrs =
|
2024-10-02 15:45:40 +02:00
|
|
|
if tagged = [] then [] else
|
|
|
|
let firstfont = hd (keep (function Cpdftype.Font _ -> true | _ -> false) tagged) in
|
2024-10-02 16:26:30 +02:00
|
|
|
[firstfont; Cpdftype.BeginDocument] @ tagged
|
2024-05-06 09:37:58 +02:00
|
|
|
in
|
2024-10-03 17:18:09 +02:00
|
|
|
let pages, tags = Cpdftype.typeset ~process_struct_tree margin margin margin margin papersize pdf instrs in
|
2024-10-04 13:24:50 +02:00
|
|
|
(*iter (fun x -> Printf.printf "PAGE\n"; iter (fun (_, i) -> Printf.printf "Paragraph number %i\n" i) x) tags;*)
|
2024-10-03 17:18:09 +02:00
|
|
|
(* We make (tag number, page number, mcid) triples *)
|
|
|
|
let tagtriples =
|
|
|
|
flatten
|
|
|
|
(map2
|
|
|
|
(fun pn tags ->
|
|
|
|
map2 (fun (_, tagnum) mcid -> (tagnum, pn, mcid)) tags (indx0 tags))
|
|
|
|
(indx0 tags)
|
|
|
|
tags)
|
|
|
|
in
|
2024-10-04 13:24:50 +02:00
|
|
|
(* Printf.printf "(paragraph number, page number, mcid) triples:\n";
|
|
|
|
iter (fun (tagnum, pn, mcid) -> Printf.printf "%i, %i, %i\n" tagnum pn mcid) tagtriples;*)
|
2024-10-04 13:15:49 +02:00
|
|
|
(* Now work out the nodes and which MCIDs in which pages they point to. Each paragraph may point to 1 or more nodes. *)
|
|
|
|
let rec find_nodes (a : ((int * int * int) list) list) = function
|
|
|
|
| (para, page, mcid)::nodes ->
|
|
|
|
begin match a with
|
|
|
|
| ((para', page', mcid')::t)::rest when para = para' ->
|
|
|
|
find_nodes (((para, page, mcid)::(para', page', mcid')::t)::rest) nodes
|
|
|
|
| (h::t)::rest ->
|
|
|
|
find_nodes (([(para, page, mcid)])::(h::t)::rest) nodes
|
|
|
|
| []::rest ->
|
|
|
|
find_nodes (([(para, page, mcid)])::rest) nodes
|
|
|
|
| [] -> assert false
|
|
|
|
end
|
|
|
|
| [] -> rev (map rev a)
|
2024-10-03 17:18:09 +02:00
|
|
|
in
|
2024-10-04 13:15:49 +02:00
|
|
|
let nodes = find_nodes [[]] tagtriples in
|
2024-10-04 15:42:31 +02:00
|
|
|
(*Printf.printf "Paragraphs and their page and MCIDs\n";
|
2024-10-03 17:18:09 +02:00
|
|
|
iter
|
2024-10-04 13:15:49 +02:00
|
|
|
(fun parts_of_para ->
|
|
|
|
Printf.printf "Paragraph:\n";
|
|
|
|
iter (fun (para, page, mcid) -> Printf.printf "Para %i, Page %i, MCID %i\n" para page mcid) parts_of_para)
|
2024-10-04 15:42:31 +02:00
|
|
|
nodes;*)
|
2024-10-03 17:18:09 +02:00
|
|
|
let pages =
|
|
|
|
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)
|
|
|
|
(indx0 pages)
|
|
|
|
pages
|
|
|
|
in
|
|
|
|
let pdf, pageroot = Pdfpage.add_pagetree pages pdf in
|
2024-10-04 14:41:47 +02:00
|
|
|
let pdf = Pdfpage.add_root pageroot [] pdf in
|
|
|
|
let refnums = let ns = Pdf.page_reference_numbers pdf in combine (indx0 ns) ns in
|
2024-10-04 16:26:47 +02:00
|
|
|
if process_struct_tree || subformat = Some Cpdfua.PDFUA1 || subformat = Some Cpdfua.PDFUA2 then
|
2024-10-04 14:41:47 +02:00
|
|
|
begin
|
2024-10-04 16:26:47 +02:00
|
|
|
let namespace = if subformat = Some Cpdfua.PDFUA2 then Pdf.addobj pdf (Pdf.Dictionary [("/NS", Pdf.String "http://iso.org/pdf2/ssn")]) else 0 in
|
|
|
|
let document = if subformat = Some Cpdfua.PDFUA2 then Pdf.addobj pdf Pdf.Null else 0 in
|
|
|
|
|
2024-10-04 14:41:47 +02:00
|
|
|
let str = Pdf.addobj pdf Pdf.Null in
|
|
|
|
let topks =
|
|
|
|
map
|
|
|
|
(fun parts_of_para ->
|
|
|
|
let ks =
|
|
|
|
map (fun (_, pagenumber, mcid) -> Pdf.Dictionary [("/Type", Pdf.Name "/MCR"); ("/Pg", Pdf.Indirect (unopt (lookup pagenumber refnums))); ("/MCID", Pdf.Integer mcid)]) parts_of_para
|
|
|
|
in
|
2024-10-04 16:26:47 +02:00
|
|
|
Pdf.Indirect (Pdf.addobj pdf (Pdf.Dictionary [("/K", Pdf.Array ks); ("/P", Pdf.Indirect (if subformat = Some Cpdfua.PDFUA2 then document else str)); ("/S", Pdf.Name "/P")])))
|
2024-10-04 14:41:47 +02:00
|
|
|
nodes
|
|
|
|
in
|
2024-10-04 16:26:47 +02:00
|
|
|
if subformat = Some Cpdfua.PDFUA2 then
|
|
|
|
Pdf.addobj_given_num pdf (document, Pdf.Dictionary [("/K", Pdf.Array topks); ("/P", Pdf.Indirect str); ("/S", Pdf.Name "/Document"); ("/NS", Pdf.Indirect namespace)]);
|
2024-10-04 15:32:49 +02:00
|
|
|
let parent_tree =
|
|
|
|
let pairs =
|
|
|
|
map
|
|
|
|
(fun pn ->
|
2024-10-04 16:26:47 +02:00
|
|
|
let this_page_triples = keep (fun (para, page, mcid) -> page = pn) tagtriples in
|
2024-10-04 15:32:49 +02:00
|
|
|
(string_of_int pn, Pdf.Array (map (function (para, _, _) -> (List.nth topks para)) this_page_triples)))
|
|
|
|
(indx0 pages)
|
|
|
|
in
|
|
|
|
Pdf.addobj pdf (Pdftree.build_name_tree true pdf pairs)
|
|
|
|
in
|
2024-10-04 16:26:47 +02:00
|
|
|
let stns =
|
|
|
|
if subformat = Some Cpdfua.PDFUA2 then [("/Namespaces", Pdf.Array [Pdf.Indirect namespace])] else []
|
|
|
|
in
|
|
|
|
let k =
|
|
|
|
if subformat = Some Cpdfua.PDFUA2 then Pdf.Indirect document else Pdf.Array topks
|
|
|
|
in
|
|
|
|
Pdf.addobj_given_num pdf (str, Pdf.Dictionary (stns @ [("/Type", Pdf.Name "/StructTreeRoot"); ("/K", k); ("/ParentTree", Pdf.Indirect parent_tree)]));
|
2024-10-23 14:44:31 +02:00
|
|
|
Pdf.replace_chain pdf ["/Root"; "/StructTreeRoot"] (Pdf.Indirect str)
|
2024-10-04 14:41:47 +02:00
|
|
|
end;
|
|
|
|
pdf
|