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

Finished -stamp-as-xobject

This commit is contained in:
John Whitington 2020-03-25 13:46:54 +00:00
parent e0308ccee9
commit 0a77fcfe3e
2 changed files with 21 additions and 23 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
cpdf.ml
View File

@ -2555,17 +2555,15 @@ let add_page_as_xobject pdf range page name =
in in
let xobject_dict = let xobject_dict =
["/Type", Pdf.Name "/XObject"; ["/Type", Pdf.Name "/XObject";
"/Subtype", Pdf.Name "/Form"; "/Subtype", Pdf.Name "/Form";
"/BBox", Pdf.Array [Pdf.Real 0.; Pdf.Real 0.; Pdf.Real 1000.; Pdf.Real 1000.]; "/BBox", page.Pdfpage.mediabox;
"/Resources", page.Pdfpage.resources; "/Resources", page.Pdfpage.resources;
"/Length", Pdf.Integer (bytes_size xobject_data); "/Length", Pdf.Integer (bytes_size xobject_data)]
]
in in
let xobject = let xobject =
Pdf.Stream {contents = (Pdf.Dictionary xobject_dict, Pdf.Got xobject_data)} Pdf.Stream {contents = (Pdf.Dictionary xobject_dict, Pdf.Got xobject_data)}
in in
let xobject_objnum = Pdf.addobj pdf xobject in let xobject_objnum = Pdf.addobj pdf xobject in
(* For each page in range, add the xobject to the list of xobjects in the resources of that page *)
let pages = Pdfpage.pages_of_pagetree pdf in let pages = Pdfpage.pages_of_pagetree pdf in
let new_pages = let new_pages =
List.map2 List.map2

34
cpdfwriteJSON.ml
View File

@ -25,25 +25,25 @@ let rec json_of_object pdf fcs no_stream_data = function
| _ -> ()) | _ -> ())
elts; elts;
J.Object (map (fun (k, v) -> (k, json_of_object pdf fcs no_stream_data v)) elts) J.Object (map (fun (k, v) -> (k, json_of_object pdf fcs no_stream_data v)) elts)
| P.Stream ({contents = (Pdf.Dictionary dict as d, stream)} as mut) as thestream -> | P.Stream ({contents = (P.Dictionary dict as d, stream)} as mut) as thestream ->
Pdf.getstream thestream; P.getstream thestream;
let str = let str =
begin match Pdf.lookup_direct pdf "/FunctionType" d with begin match P.lookup_direct pdf "/FunctionType" d with
| Some _ -> | Some _ ->
Pdfcodec.decode_pdfstream_until_unknown pdf thestream; Pdfcodec.decode_pdfstream_until_unknown pdf thestream;
begin match !mut with (_, Pdf.Got b) -> Pdfio.string_of_bytes b | _ -> "failure: decomp" end begin match !mut with (_, P.Got b) -> Pdfio.string_of_bytes b | _ -> "failure: decomp" end
| None -> | None ->
if no_stream_data then "<<stream data elided>>" else if no_stream_data then "<<stream data elided>>" else
match stream with Pdf.Got b -> Pdfio.string_of_bytes b | Pdf.ToGet _ -> "failure: toget" match stream with P.Got b -> Pdfio.string_of_bytes b | P.ToGet _ -> "failure: toget"
end end
in in
json_of_object pdf fcs no_stream_data (P.Array [P.Dictionary dict; P.String str]) json_of_object pdf fcs no_stream_data (P.Array [P.Dictionary dict; P.String str])
| P.Stream _ -> J.String "error: stream with not-a-dictioary" | P.Stream _ -> J.String "error: stream with not-a-dictioary"
| P.Indirect i -> | P.Indirect i ->
begin match Pdf.lookup_obj pdf i with begin match P.lookup_obj pdf i with
| P.Stream {contents = (Pdf.Dictionary dict as d, _)} -> | P.Stream {contents = (P.Dictionary dict as d, _)} ->
begin match Pdf.lookup_direct pdf "/Subtype" d with begin match P.lookup_direct pdf "/Subtype" d with
| Some (Pdf.Name "/Form") -> fcs i | Some (P.Name "/Form") -> fcs i
| _ -> () | _ -> ()
end end
| _ -> () | _ -> ()
@ -164,23 +164,23 @@ let json_of_op pdf no_stream_data = function
* resources, though? For now, don't worry about inherited resources: check in * resources, though? For now, don't worry about inherited resources: check in
* PDF standard. *) * PDF standard. *)
let parse_content_stream pdf resources bs = let parse_content_stream pdf resources bs =
let ops = Pdfops.parse_stream pdf resources [bs] in let ops = O.parse_stream pdf resources [bs] in
J.Array (map (json_of_op pdf false) ops) J.Array (map (json_of_op pdf false) ops)
let json_of_pdf parse_content no_stream_data pdf = let json_of_pdf parse_content no_stream_data pdf =
let trailerdict = (0, json_of_object pdf (fun x -> ()) no_stream_data pdf.Pdf.trailerdict) in let trailerdict = (0, json_of_object pdf (fun x -> ()) no_stream_data pdf.P.trailerdict) in
let content_streams = ref [] in let content_streams = ref [] in
let fcs n = content_streams := n::!content_streams in let fcs n = content_streams := n::!content_streams in
let pairs = let pairs =
let ps = ref [] in let ps = ref [] in
Pdf.objiter P.objiter
(fun i pdfobj -> (fun i pdfobj ->
ps := (i, json_of_object pdf fcs no_stream_data pdfobj)::!ps) ps := (i, json_of_object pdf fcs no_stream_data pdfobj)::!ps)
pdf; pdf;
trailerdict::!ps trailerdict::!ps
in in
if parse_content then if parse_content then
iter (fun n -> Pdfcodec.decode_pdfstream_until_unknown pdf (Pdf.lookup_obj pdf n)) !content_streams; iter (fun n -> Pdfcodec.decode_pdfstream_until_unknown pdf (P.lookup_obj pdf n)) !content_streams;
let pairs_parsed = let pairs_parsed =
if not parse_content then pairs else if not parse_content then pairs else
map map
@ -190,11 +190,11 @@ let json_of_pdf parse_content no_stream_data pdf =
| J.Array [dict; J.String _] -> | J.Array [dict; J.String _] ->
(* FIXME Proper resources here for reasons explained above *) (* FIXME Proper resources here for reasons explained above *)
let streamdata = let streamdata =
match Pdf.lookup_obj pdf objnum with match P.lookup_obj pdf objnum with
| Pdf.Stream {contents = (_, Pdf.Got b)} -> b | P.Stream {contents = (_, P.Got b)} -> b
| _ -> failwith "JSON: stream not decoded" | _ -> failwith "JSON: stream not decoded"
in in
(objnum, J.Array [dict; parse_content_stream pdf (Pdf.Dictionary []) streamdata]) (objnum, J.Array [dict; parse_content_stream pdf (P.Dictionary []) streamdata])
| _ -> failwith "json_of_pdf: stream parsing inconsistency" | _ -> failwith "json_of_pdf: stream parsing inconsistency"
end end
else else
@ -209,7 +209,7 @@ let json_of_pdf parse_content no_stream_data pdf =
let write fh parse_content no_stream_data pdf = let write fh parse_content no_stream_data pdf =
let b = Buffer.create 256 in let b = Buffer.create 256 in
let formatter = Format.formatter_of_buffer b in let formatter = Format.formatter_of_buffer b in
Tjjson.format formatter (json_of_pdf parse_content no_stream_data pdf); J.format formatter (json_of_pdf parse_content no_stream_data pdf);
Format.pp_print_flush formatter (); Format.pp_print_flush formatter ();
output_string fh (Buffer.contents b) output_string fh (Buffer.contents b)