more

cpdfbookmarks.ml

@@ -192,17 +192,7 @@ let output_json_marks ch calculate_page_number pdf fastrefnums marks =
   let json = `List (map json_of_mark marks) in
     J.pretty_to_channel ch json
 
-(* List the bookmarks in the given range to the given output *)
+let process_string encoding s =
-let list_bookmarks ~json encoding range pdf output =
-  let process_stripped escaped =
-    let b = Buffer.create 200 in
-      iter
-        (fun x ->
-           if x <= 127 then Buffer.add_char b (char_of_int x))
-        escaped;
-      Buffer.contents b
-  in
-  let process_string s =
   let rec replace c x y = function
   | [] -> []
   | h::t when h = c -> x::y::replace c x y t
@@ -216,12 +206,22 @@ let list_bookmarks ~json encoding range pdf output =
         and n = int_of_char 'n'
         and q = int_of_char '\"' in
           replace q bs q (replace nl bs n (replace bs bs bs codepoints))
+      in
+        let process_stripped escaped =
+          let b = Buffer.create 200 in
+            iter
+              (fun x ->
+                 if x <= 127 then Buffer.add_char b (char_of_int x))
+              escaped;
+            Buffer.contents b
         in
         match encoding with
         | Cpdfmetadata.UTF8 -> Pdftext.utf8_of_codepoints escaped
         | Cpdfmetadata.Stripped -> process_stripped escaped
         | Cpdfmetadata.Raw -> s
-    in
+
+(* List the bookmarks in the given range to the given output *)
+let list_bookmarks ~json encoding range pdf output =
   let bookmarks = Pdfmarks.read_bookmarks pdf in
   let refnums = Pdf.page_reference_numbers pdf in
   let rangetable = hashset_of_list range in
@@ -256,7 +256,7 @@ let list_bookmarks ~json encoding range pdf output =
                output.Pdfio.output_string
                  (Printf.sprintf "%i \"%s\" %i%s %s\n"
                    mark.Pdfmarks.level
-                       (process_string mark.Pdfmarks.text)
+                   (process_string encoding mark.Pdfmarks.text)
                    (calculate_page_number mark)
                    (if mark.Pdfmarks.isopen then " open" else "")
                    (output_string_of_target pdf fastrefnums mark.Pdfmarks.target)))

cpdfmanual.tex

@@ -1562,6 +1562,9 @@ There are two options which turn off parts of the squeezer. They are \texttt{-sq
   \begin{framed}
   \small\noindent\verb!cpdf -list-bookmarks [-utf8 | -raw] in.pdf!
 
+  \vspace{1.5mm}
+  \small\noindent\verb!cpdf -list-bookmarks-json in.pdf!
+
   \vspace{1.5mm}
   \small\noindent\verb!cpdf -remove-bookmarks in.pdf -o out.pdf!
 
@@ -1608,6 +1611,24 @@ the ASCII range. To prevent this, and return unicode UTF8 output, add the
 \texttt{-utf8} option to the command. To prevent any processing, use the
 \texttt{-raw} option. See Section \ref{textencodings} for more information. A newline in a bookmark is represented as \texttt{"\textbackslash n"}.
 
+By using \texttt{-list-bookmarks-json} instead, the bookmarks are formatted as a JSON array, in order, of dictionaries formatted thus:
+
+\begin{verbatim}
+{ "level": 0,
+  "text": "1 Basic Usage",
+  "page": 17,
+  "open": false,
+  "target":
+    [ { "I": 17 },
+      { "N": "/XYZ" },
+      { "F": 85.039 },
+      { "F": 609.307 },
+      null ]
+}
+\end{verbatim}
+
+See chapter 15 for more details of cpdf's JSON formatting. Bookmark text in JSON bookmarks, however, is in UTF8 for ease of use.
+
 \subsection{Destinations}
 
 The destination is an extended description of where the bookmark should point to (i.e it can be more detailed than just giving the page). For example, it may point to a section heading halfway down a page. Here are the possibilities: