diff options
Diffstat (limited to 'tex/context/modules/mkiv/s-xml-analyzers.lua')
| -rw-r--r-- | tex/context/modules/mkiv/s-xml-analyzers.lua | 189 |
1 files changed, 159 insertions, 30 deletions
diff --git a/tex/context/modules/mkiv/s-xml-analyzers.lua b/tex/context/modules/mkiv/s-xml-analyzers.lua index c356d4c37..6e7f7f2ba 100644 --- a/tex/context/modules/mkiv/s-xml-analyzers.lua +++ b/tex/context/modules/mkiv/s-xml-analyzers.lua @@ -11,6 +11,8 @@ moduledata.xml.analyzers = moduledata.xml.analyzers or { } local next, type = next, type local utfvalues = string.utfvalues +local formatters = string.formatters +local setmetatableindex = table.setmetatableindex local context = context local NC, NR, HL, FL, LL, SL, TB = context.NC, context.NR, context.HL, context.TB, context.FL, context.LL, context.SL local sortedhash, sortedkeys, concat, sequenced = table.sortedhash, table.sortedkeys, table.concat, table.sequenced @@ -43,28 +45,29 @@ local function analyze(filename) attr = { } ents = { } - table.setmetatableindex(tags,function(t,k) + local function att(t,k) + local v = setmetatableindex("number") + t[k] = v + return v + end + + local function add(t,k) local v = { n = 0, - attributes = { }, - children = { }, + attributes = setmetatableindex(att), + children = setmetatableindex(add), } t[k] = v return v - end) + end - table.setmetatableindex(char,function(t,k) - t[k] = 0 - return 0 - end) + setmetatableindex(tags,add) - table.setmetatableindex(attr,function(t,k) - char[k] = char[k] or 0 - t[k] = 0 - return 0 - end) + setmetatableindex(ents,"number") + setmetatableindex(char,"number") - table.setmetatableindex(ents,function(t,k) + setmetatableindex(attr,function(t,k) + char[k] = char[k] or 0 t[k] = 0 return 0 end) @@ -85,24 +88,25 @@ local function analyze(filename) local tg = e.tg local tag = tags[tg] tag.n = tag.n + 1 + local children = parent and tags[parent].children[tg] + local childatt = children and children.attributes + if children then + children.n = children.n + 1 + end if at then local attributes = tag.attributes for k, v in next, at do local a = attributes[k] - if a then - a[v] = (a[v] or 0) + 1 - else - attributes[k] = { [v] = 1 } + a[v] = a[v] + 1 + if childatt then + local a = childatt[k] + a[v] = a[v] + 1 end for s in utfvalues(v) do attr[s] = attr[s] + 1 end end end - if parent then - local children = tags[parent].children - children[tg] = (children[tg] or 0) + 1 - end if dt then for i=1,#dt do local d = dt[i] @@ -119,7 +123,11 @@ local function analyze(filename) end for i=1,#filename do - local root = xml.load(filename[i]) + local name = filename[i] + local root = xml.load(name) + -- + logs.report("xml analyze","loaded: %s",name) + -- collect(root) -- local names = root.statistics.entities.names @@ -128,10 +136,10 @@ local function analyze(filename) end end - table.setmetatableindex(tags,nil) - table.setmetatableindex(char,nil) - table.setmetatableindex(attr,nil) - table.setmetatableindex(ents,nil) + setmetatableindex(tags,nil) + setmetatableindex(char,nil) + setmetatableindex(attr,nil) + setmetatableindex(ents,nil) end @@ -153,16 +161,20 @@ function moduledata.xml.analyzers.structure(filename) NC() context.bold("element") NC() context.darkred(name) NC() NR() NC() context.bold("frequency") NC() context(data.n) NC() NR() if next(children) then - NC() context.bold("children") NC() context.puretext(sequenced(children)) NC() NR() + local t = { } + for k, v in next, children do + t[k] = v.n + end + NC() context.bold("children") NC() context.puretext(sequenced(t)) NC() NR() end if next(attributes) then NC() context.bold("attributes") NC() context.puretext.darkgreen(concat(sortedkeys(attributes)," ")) NC() NR() for attribute, values in sortedhash(attributes) do local n = table.count(values) if attribute == "id" or attribute == "xml:id" or n > maxnofattributes then - NC() context(attribute) NC() context("%s different values",n) NC() NR() + NC() context("@%s",attribute) NC() context("%s different values",n) NC() NR() else - NC() context(attribute) NC() context.puretext(sequenced(values)) NC() NR() + NC() context("@%s",attribute) NC() context.puretext(sequenced(values)) NC() NR() end end end @@ -195,4 +207,121 @@ function moduledata.xml.analyzers.entities(filename) context.stoptabulate() end +local f_parent_s = formatters["xml:%s"] +local f_parent_n = formatters["\\startxmlsetups xml:%s\n \\xmlflush{#1}\n\\stopxmlsetups"] +local f_parent_a = formatters["\\startxmlsetups xml:%s\n %% @ % t\n \\xmlflush{#1}\n\\stopxmlsetups"] +local f_child_s = formatters["xml:%s:%s"] +local f_child_n = formatters["\\startxmlsetups xml:%s:%s\n \\xmlflush{#1}\n\\stopxmlsetups"] +local f_child_a = formatters["\\startxmlsetups xml:%s:%s\n %% @ % t\n \\xmlflush{#1}\n\\stopxmlsetups"] + +local f_template = formatters [ [[ +%% file: %s + +%% Beware, these are all (first level) setups. If you have a complex document +%% it often makes sense to use \\xmlfilter or similar local filter options. + +%% presets + +\startxmlsetup xml:presets:all + \xmlsetsetups {#1} { + %s + } +\stopxmlsetups + +%% setups + +%s +]] ] + +function moduledata.xml.analyzers.allsetups(filename,usedname) + analyze(filename) + local result = { } + local setups = { } + for name, data in table.sortedhash(tags) do + local children = data.children + local attributes = data.attributes + if next(attributes) then + result[#result+1] = f_parent_a(name,sortedkeys(attributes)) + else + result[#result+1] = f_parent_n(name) + end + setups[#setups+1] = f_parent_s(name) + if next(children) then + for k, v in sortedhash(children) do + local attributes = v.attributes + if next(attributes) then + result[#result+1] = f_child_a(name,k,sortedkeys(attributes)) + else + result[#result+1] = f_child_n(name,k) + end + setups[#setups+1] = f_child_s(name,k) + end + end + end + table.sort(setups) + -- + if type(filename) == "table" then + filename = concat(filename," | ") + end + -- + usedname = usedname or "xml-analyze-template.tex" + -- + io.savedata(usedname,f_template(filename,concat(setups,"|\n "),concat(result,"\n\n"))) + logs.report("xml analyze","presets saved in: %s",usedname) +end + +-- example: + +-- local t = { } +-- local x = xml.load("music-collection.xml") +-- for c in xml.collected(x,"//*") do +-- if not c.special and not t[c.tg] then +-- t[c.tg] = true +-- end +-- end +-- inspect(table.sortedkeys(t)) + +-- xml.finalizers.taglist = function(collected) +-- local t = { } +-- for i=1,#collected do +-- local c = collected[i] +-- if not c.special then +-- local tg = c.tg +-- if tg and not t[tg] then +-- t[tg] = true +-- end +-- end +-- end +-- return t +-- end +-- local x = xml.load("music-collection.xml") +-- inspect(table.sortedkeys(xml.applylpath(x,"//*/taglist()"))) + +-- xml.finalizers.taglist = function(collected,parenttoo) +-- local t = { } +-- for i=1,#collected do +-- local c = collected[i] +-- if not c.special then +-- local tg = c.tg +-- if tg and not t[tg] then +-- t[tg] = true +-- end +-- if parenttoo then +-- local p = c.__p__ +-- if p and not p.special then +-- local tg = p.tg .. ":" .. tg +-- if tg and not t[tg] then +-- t[tg] = true +-- end +-- end +-- end +-- end +-- end +-- return t +-- end + +-- local x = xml.load("music-collection.xml") +-- inspect(table.sortedkeys(xml.applylpath(x,"//*/taglist()"))) +-- local x = xml.load("music-collection.xml") +-- inspect(table.sortedkeys(xml.applylpath(x,"//*/taglist(true)"))) |