cpdf in.pdf -output-json -o out.json
[-output-json-parse-content-streams]
[-output-json-no-stream-data]
[-output-json-decompress-streams]
[-output-json-clean-strings]
[-utf8]
cpdf -j in.json -o out.pdf
In addition to reading and writing PDF files in the original Adobe format, cpdf can read and write them in its own CPDFJSON format, for somewhat easier extraction of information, modification of PDF files, and so on.
We convert a PDF file to JSON format like this:
cpdf -output-json in.pdf -o out.json
The resultant JSON file is an array of arrays containing an object number followed by an object, one for each object in the file and two special ones:
Object -1: CPDF’s own data with the PDF version number, CPDF JSON format number, and flags used when writing (which may be required when reading):
/CPDFJSONformatversion (CPDFJSON integer (see below), currently 3)
/CPDFJSONcontentparsed (boolean, true if content streams have been parsed)
/CPDFJSONstreamdataincluded (boolean, true if stream data included. Cannot round-trip if false).
/CPDFJSONmajorpdfversion (CPDFJSON integer)
/CPDFJSONminorpdfversion (CPDFJSON integer)
Object 0: The PDF’s trailer dictionary
Objects 1..n: The PDF’s objects.
Objects are formatted thus:
PDF arrays, dictionaries, booleans, and strings are the same as in JSON.
Integers are written as {"I": 0}
Floats are written as {"F": 0.0}
Names are written as {"N": "/Pages"}
Indirect references are integers
Streams are {"S": [dict, data]}
Strings are converted to JSON string format in a way which, when reversed, results in the original string. For best results when editing files, use the -utf8 option. The string representation is again reversible, but easier to edit. Unicode strings are written as {"U": "the text"}.
Here is an example of the output for a small PDF:
[ [ -1, { "/CPDFJSONformatversion": { "I": 2 }, "/CPDFJSONcontentparsed": false, "/CPDFJSONstreamdataincluded": true, "/CPDFJSONmajorpdfversion": { "I": 1 }, "/CPDFJSONminorpdfversion": { "I": 1 } } ], [ 0, { "/Size": { "I": 4 }, "/Root": 4, "/ID" : [ <elided>, <elided>] } ], [ 1, { "/Type": { "N": "/Pages" }, "/Kids": [ 3 ], "/Count": { "I": 1 } } ], [ 2, {"S": [{ "/Length": { "I": 49 } }, "1 0 0 1 50 770 cm BT/F0 36 Tf(Hello, World!)Tj ET"] } ], [ 3, { "/Type": { "N": "/Page" }, "/Parent": 1, "/Resources": { "/Font": { "/F0": { "/Type": { "N": "/Font" }, "/Subtype": { "N": "/Type1" }, "/BaseFont": { "N": "/Times-Italic" } } } }, "/MediaBox": [{ "I": 0 }, { "I": 0 }, { "F": 595.2755905510001 }, { "F": 841.88976378 }], "/Rotate": { "I": 0 }, "/Contents": [ 2 ] } ], [ 4, { "/Type": { "N": "/Catalog" }, "/Pages": 1 } ] ]
The option -output-json-parse-content-streams will also convert content streams to JSON, so our example content stream will be expanded:
2, { "S": [ {}, [ [ { "F": 1.0 }, { "F": 0.0 }, { "F": 0.0 }, { "F": 1.0 }, { "F": 50.0 }, { "F": 770.0 }, "cm" ], [ "BT" ], [ "/F0", { "F": 36.0 }, "Tf" ], [ "Hello, World!", "Tj" ], [ "ET" ] ] ] } ], [
The option -output-json-no-stream-data simply elides the stream data instead, leading to much smaller JSON files. But these may not be round-tripped back into PDF, of course.
The option -output-json-decompress-streams keeps the streams intact, and decompresses them.
The option -output-json-clean-strings converts any UTF16BE strings with no high bytes to PDFDocEncoding prior to output, so that editing them is easier. Note: this is deprecated as of version 2.6 in favour of -utf8.
We can load a JSON PDF file with the -j option in place of a PDF file anywhere in a normal cpdf command. A range may be applied, just like any other file.
cpdf -j in.json -o out.pdf
It is not required that /Length entries in CPDFJSON stream dictionaries be correctly updated when the JSON file is edited: cpdf will fix them when loading.