if not modules then modules = { } end modules ['lpdf-xmp'] = { version = 1.001, comment = "companion to lpdf-ini.mkiv", author = "Hans Hagen, PRAGMA-ADE, Hasselt NL", copyright = "PRAGMA ADE / ConTeXt Development Team", license = "see context related readme files", comment = "with help from Peter Rolf", } local tostring, type = tostring, type local format, random, char, gsub, concat = string.format, math.random, string.char, string.gsub, table.concat local xmlfillin = xml.fillin local trace_xmp = false trackers.register("backend.xmp", function(v) trace_xmp = v end) local trace_info = false trackers.register("backend.info", function(v) trace_info = v end) local report_xmp = logs.reporter("backend","xmp") local report_info = logs.reporter("backend","info") local backends, lpdf = backends, lpdf local codeinjections = backends.pdf.codeinjections -- normally it is registered local pdfdictionary = lpdf.dictionary local pdfconstant = lpdf.constant local pdfreference = lpdf.reference local pdfflushstreamobject = lpdf.flushstreamobject -- I wonder why this begin end is empty / w (no time now to look into it) / begin can also be "?" local xpacket = [[ %s ]] local mapping = { -- user defined keys (pdfx:) ["ConTeXt.Jobname"] = { "context", "rdf:Description/pdfx:ConTeXt.Jobname" }, ["ConTeXt.Time"] = { "date", "rdf:Description/pdfx:ConTeXt.Time" }, ["ConTeXt.Url"] = { "context", "rdf:Description/pdfx:ConTeXt.Url" }, ["ConTeXt.Version"] = { "context", "rdf:Description/pdfx:ConTeXt.Version" }, ["ID"] = { "id", "rdf:Description/pdfx:ID" }, -- has date ["PTEX.Fullbanner"] = { "metadata","rdf:Description/pdfx:PTEX.Fullbanner" }, -- Adobe PDF schema ["Keywords"] = { "metadata","rdf:Description/pdf:Keywords" }, ["Producer"] = { "metadata","rdf:Description/pdf:Producer" }, -- ["Trapped"] = { "pdf", "rdf:Description/pdf:Trapped" }, -- '/False' in /Info, but 'False' in XMP -- Dublin Core schema ["Author"] = { "metadata","rdf:Description/dc:creator/rdf:Seq/rdf:li" }, ["Format"] = { "metadata","rdf:Description/dc:format" }, -- optional, but nice to have ["Subject"] = { "metadata","rdf:Description/dc:description/rdf:Alt/rdf:li" }, ["Title"] = { "metadata","rdf:Description/dc:title/rdf:Alt/rdf:li" }, -- XMP Basic schema ["CreateDate"] = { "date", "rdf:Description/xmp:CreateDate" }, ["CreationDate"] = { "date", "rdf:Description/xmp:CreationDate" }, -- dummy ["Creator"] = { "metadata","rdf:Description/xmp:CreatorTool" }, ["MetadataDate"] = { "date", "rdf:Description/xmp:MetadataDate" }, ["ModDate"] = { "date", "rdf:Description/xmp:ModDate" }, -- dummy ["ModifyDate"] = { "date", "rdf:Description/xmp:ModifyDate" }, -- XMP Media Management schema ["DocumentID"] = { "id", "rdf:Description/xmpMM:DocumentID" }, -- uuid ["InstanceID"] = { "id", "rdf:Description/xmpMM:InstanceID" }, -- uuid ["RenditionClass"] = { "pdf", "rdf:Description/xmpMM:RenditionClass" }, -- PDF/X-4 ["VersionID"] = { "pdf", "rdf:Description/xmpMM:VersionID" }, -- PDF/X-4 -- additional entries -- PDF/X ["GTS_PDFXVersion"] = { "pdf", "rdf:Description/pdfxid:GTS_PDFXVersion" }, -- optional entries -- all what is visible in the 'document properties --> additional metadata' window -- XMP Rights Management schema (optional) ["Marked"] = { "pdf", "rdf:Description/xmpRights:Marked" }, -- ["Owner"] = { "metadata", "rdf:Description/xmpRights:Owner/rdf:Bag/rdf:li" }, -- maybe useful (not visible) -- ["UsageTerms"] = { "metadata", "rdf:Description/xmpRights:UsageTerms" }, -- maybe useful (not visible) ["WebStatement"] = { "metadata", "rdf:Description/xmpRights:WebStatement" }, -- Photoshop PDF schema (optional) ["AuthorsPosition"] = { "metadata", "rdf:Description/photoshop:AuthorsPosition" }, ["Copyright"] = { "metadata", "rdf:Description/photoshop:Copyright" }, ["CaptionWriter"] = { "metadata", "rdf:Description/photoshop:CaptionWriter" }, } pdf.setsuppressoptionalinfo( 0 -- + 1 -- pdfnofullbanner + 2 -- pdfnofilename + 4 -- pdfnopagenumber + 8 -- pdfnoinfodict + 16 -- pdfnocreator + 32 -- pdfnocreationdate + 64 -- pdfnomoddate + 128 -- pdfnoproducer + 256 -- pdfnotrapped -- + 512 -- pdfnoid ) local included = table.setmetatableindex( { context = true, id = true, metadata = true, date = true, id = true, pdf = true, }, function(t,k) return true end) function lpdf.settrailerid(v) if v then local b = toboolean(v) or v == "" if b then v = "This file is processed by ConTeXt and LuaTeX." else v = tostring(v) end local h = md5.HEX(v) if b then report_info("using frozen trailer id") else report_info("using hashed trailer id %a (%a)",v,h) end pdf.settrailerid(format("[<%s> <%s>]",h,h)) end end function lpdf.setdates(v) local t = type(v) if t == "number" or t == "string" then t = converters.totime(v) if t then included.date = true included.id = "fake" report_info("forced date/time information %a will be used",lpdf.settime(t)) lpdf.settrailerid(false) return end end v = toboolean(v) included.date = v if v then included.id = true else report_info("no date/time but fake id information will be added") lpdf.settrailerid(true) included.id = "fake" -- maybe: lpdf.settime(231631200) -- 1975-05-05 % first entry of knuth about tex mentioned in DT end end function lpdf.id() -- overload of ini local banner = tex.jobname if included.date then return format("%s.%s",banner,lpdf.timestamp()) else return banner end end directives.register("backend.trailerid", lpdf.settrailerid) directives.register("backend.date", lpdf.setdates) local function permitdetail(what) local m = mapping[what] if m then return included[m[1]] and m[2] else return included[what] and true or false end end lpdf.permitdetail = permitdetail -- maybe some day we will load the xmp file at runtime local xmp, xmpfile, xmpname = nil, nil, "lpdf-pdx.xml" local function setxmpfile(name) if xmp then report_xmp("discarding loaded file %a",xmpfile) xmp = nil end xmpfile = name ~= "" and name end codeinjections.setxmpfile = setxmpfile interfaces.implement { name = "setxmpfile", arguments = "string", actions = setxmpfile } local function valid_xmp() if not xmp then -- local xmpfile = xmpfile or resolvers.findfile(xmpname) or "" if xmpfile and xmpfile ~= "" then xmpfile = resolvers.findfile(xmpfile) or "" end if not xmpfile or xmpfile == "" then xmpfile = resolvers.findfile(xmpname) or "" end if xmpfile ~= "" then report_xmp("using file %a",xmpfile) end local xmpdata = xmpfile ~= "" and io.loaddata(xmpfile) or "" xmp = xml.convert(xmpdata) end return xmp end function lpdf.addxmpinfo(tag,value,check) local pattern = permitdetail(tag) if type(pattern) == "string" then xmlfillin(xmp or valid_xmp(),pattern,value,check) end end -- redefined local pdfaddtoinfo = lpdf.addtoinfo local pdfaddxmpinfo = lpdf.addxmpinfo function lpdf.addtoinfo(tag,pdfvalue,strvalue) local pattern = permitdetail(tag) if pattern then pdfaddtoinfo(tag,pdfvalue) end if type(pattern) == "string" then local value = strvalue or gsub(tostring(pdfvalue),"^%((.*)%)$","%1") -- hack if trace_info then report_info("set %a to %a",tag,value) end xmlfillin(xmp or valid_xmp(),pattern,value,check) end end local pdfaddtoinfo = lpdf.addtoinfo -- used later -- for the do-it-yourselvers function lpdf.insertxmpinfo(pattern,whatever,prepend) xml.insert(xmp or valid_xmp(),pattern,whatever,prepend) end function lpdf.injectxmpinfo(pattern,whatever,prepend) xml.inject(xmp or valid_xmp(),pattern,whatever,prepend) end -- flushing local function randomstring(n) local t = { } for i=1,n do t[i] = char(96 + random(26)) end return concat(t) end randomstring(26) -- kind of initializes and kicks off random local function flushxmpinfo() commands.pushrandomseed() commands.setrandomseed(os.time()) local packetid = "no unique packet id here" -- 24 chars local documentid = "no unique document id here" local instanceid = "no unique instance id here" local producer = format("LuaTeX-%0.2f.%s",status.luatex_version/100,status.luatex_revision) local creator = "LuaTeX + ConTeXt MkIV" local time = lpdf.timestamp() local fullbanner = status.banner if included.id ~= "fake" then packetid = randomstring(24) documentid = "uuid:%s" .. os.uuid() instanceid = "uuid:%s" .. os.uuid() end pdfaddxmpinfo("DocumentID", documentid) pdfaddxmpinfo("InstanceID", instanceid) pdfaddxmpinfo("Producer", producer) pdfaddxmpinfo("CreatorTool", creator) pdfaddxmpinfo("CreateDate", time) pdfaddxmpinfo("ModifyDate", time) pdfaddxmpinfo("MetadataDate", time) pdfaddxmpinfo("PTEX.Fullbanner", fullbanner) pdfaddtoinfo("Producer", producer) pdfaddtoinfo("Creator", creator) pdfaddtoinfo("CreationDate", time) pdfaddtoinfo("ModDate", time) -- pdfaddtoinfo("PTEX.Fullbanner", fullbanner) -- no checking done on existence local blob = xml.tostring(xml.first(xmp or valid_xmp(),"/x:xmpmeta")) local md = pdfdictionary { Subtype = pdfconstant("XML"), Type = pdfconstant("Metadata"), } if trace_xmp then report_xmp("data flushed, see log file") logs.pushtarget("logfile") report_xmp("start xmp blob") logs.newline() logs.writer(blob) logs.newline() report_xmp("stop xmp blob") logs.poptarget() end blob = format(xpacket,packetid,blob) if not verbose and pdf.getcompresslevel() > 0 then blob = gsub(blob,">%s+<","><") end local r = pdfflushstreamobject(blob,md,false) -- uncompressed lpdf.addtocatalog("Metadata",pdfreference(r)) commands.poprandomseed() -- hack end -- his will be enabled when we can inhibit compression for a stream at the lua end lpdf.registerdocumentfinalizer(flushxmpinfo,1,"metadata") directives.register("backend.verbosexmp", function(v) verbose = v end)