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

Scaffolding for extended -obj and friends

This commit is contained in:
John Whitington 2024-06-24 15:29:32 +01:00
parent 0a774c9a88
commit 0244a8d09b
2 changed files with 25 additions and 17 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
Changes
View File

@ -8,7 +8,13 @@ o Verify compliance to PDF/UA via the Matterhorn protocol
o Extract, edit and reapply document structure tree
o Split structure tree when splitting PDF to save size
o Combine structure trees when stamping PDFs
o Report natural language on -info
o Set the natural language of a file
Extended features:
o Allow -obj/-extract-stream to lookup nested PDF information
o Merge structure trees better when merging files
o Report top-level natural language on -info
o Report mark information dictionary contents on -info
2.7 (February 2024)

34
cpdfcommand.ml
View File

@ -24,7 +24,6 @@ let initial_file_size = ref 0
let empty = Pdf.empty ()
(* Wrap up the file reading functions to exit with code 1 when an encryption
problem occurs. This happens when object streams are in an encrypted document
and so it can't be read without the right password... The existing error
@ -217,8 +216,8 @@ type op =
| Chop of int * int
| ChopHV of bool * float
| ProcessImages
| ExtractStream of int
| PrintObj of int
| ExtractStream of string
| PrintObj of string
| Verify of string
| MarkAs of string
| RemoveMark of string
@ -1841,15 +1840,15 @@ let setjbig2_lossy_threshold f =
let setprocessimagesinfo () =
set Cpdfimage.debug_image_processing
let setextractstream i =
args.op <- Some (ExtractStream i)
let setextractstream s =
args.op <- Some (ExtractStream s)
let setextractstreamdecomp i =
args.op <- Some (ExtractStream i);
let setextractstreamdecomp s =
args.op <- Some (ExtractStream s);
args.extract_stream_decompress <- true
let setprintobj i =
args.op <- Some (PrintObj i)
let setprintobj s =
args.op <- Some (PrintObj s)
(* Parse a control file, make an argv, and then make Arg parse it. *)
let rec make_control_argv_and_parse filename =
@ -2815,9 +2814,9 @@ and specs =
("-rise", Arg.Float (fun f -> Cpdfdrawcontrol.addop (Cpdfdraw.Rise f)), " Set text rise");
("-nl", Arg.Unit (fun () -> Cpdfdrawcontrol.addop Cpdfdraw.Newline), " New line");
("-newpage", Arg.Unit Cpdfdrawcontrol.addnewpage, " Move to a fresh page");
("-extract-stream", Arg.Int setextractstream, " Extract a stream");
("-extract-stream-decompress", Arg.Int setextractstreamdecomp, "Extract a stream, decompressing");
("-obj", Arg.Int setprintobj, "Print object");
("-extract-stream", Arg.String setextractstream, " Extract a stream");
("-extract-stream-decompress", Arg.String setextractstreamdecomp, "Extract a stream, decompressing");
("-obj", Arg.String setprintobj, "Print object");
("-json", Arg.Unit (fun () -> args.format_json <- true), "Format output as JSON");
("-verify", Arg.String (fun s -> setop (Verify s) ()), "Verify conformance to a standard");
("-mark-as", Arg.String (fun s -> setop (MarkAs s) ()), "Mark as conforming to a standard");
@ -3413,6 +3412,9 @@ let build_enc () =
Pdfwrite.user_password = args.user;
Pdfwrite.permissions = banlist_of_args ()}
let objnum_of_objspec s =
int_of_string s
let extract_stream pdf decomp objnum =
let obj = Pdf.lookup_obj pdf objnum in
Pdf.getstream obj;
@ -4454,12 +4456,12 @@ let go () =
~dpi_threshold:args.dpi_threshold ~factor:args.resample_factor ~interpolate:args.resample_interpolate
~path_to_jbig2enc:args.path_to_jbig2enc ~path_to_convert:args.path_to_im range pdf;
write_pdf false pdf
| Some (ExtractStream i) ->
| Some (ExtractStream s) ->
let pdf = get_single_pdf args.op true in
extract_stream pdf args.extract_stream_decompress i
| Some (PrintObj i) ->
extract_stream pdf args.extract_stream_decompress (objnum_of_objspec s)
| Some (PrintObj s) ->
let pdf = get_single_pdf args.op true in
print_obj pdf i
print_obj pdf (objnum_of_objspec s)
| Some (Verify standard) ->
begin match standard with
| "PDF/UA-1(matterhorn)" ->