initial released

README.md

@@ -4,3 +4,14 @@ another experimential typesetting tool
 
 ## Known issues
 - Some .ttc file(eg. Noto Sans/Serif CJK) can't be loaded. It seems to be a libharu bug. For those want to use Noto Sans/Serif CJK, please download the .ttf file on Google Fonts. [Example (Trad. Chinese Version)](https://fonts.google.com/noto/specimen/Noto+Sans+TC)
+
+## How to test
+To test with the example, please using the command:
+
+```
+cd /path/to/uahgi
+julia --project="." src/uahgi.jl example/ex1.ug
+```
+
+## Origin of the name
+The name is after the Taiwanese/Hokkien word for the movable type, which is ua̍h-gī (活字), in Taichung-Basin accent.

example/ex1.ug

@@ -1,7 +1,8 @@
-{@lang|en}
+{@lang|en}%
-{@def|@font|{@quote|FreeSans|AR PL UKai TW}}
+%{@def|@font|{@quote|FreeSans|AR PL UKai TW}}%
-{@def|@fontsize|12}
+%{@def|@fontsize|12}%
-{@def|@linewidth|200} %in px%
+%{@def|@linewidth|200} %in px
-{@def|@spacing|{@hglue|{@ex|1}|2}}
+%{@def|@spacing|{@hglue|{@ex|1}|2}}%
-{@def|@cjk_spacing|{@hglue|{@ex|0}|0.001}}
+%{@def|@cjk_spacing|{@hglue|{@ex|0}|0.001}}%
-11111 222 3333333 444444 555555 666666 77777 88888 99999 000000 11111 22222 33333 55555
+%the processing of typesetting technology is so sophisticated that many people are afraid of
+the implementation of typesetting language. However, it is good to give TEXBook a try.測試一段漢字，究竟漢字會不會斷行呢

src/arrange.jl

@@ -3,35 +3,160 @@ using Match
 
 u = Main.uahgi
 
-# the cost of make ith to jth char a line
+total_cost_table = Dict()
-function cost(items, i, j, linewidth)
+
+# the cost of make ith to jth item a line
+function cost(items, i, j, linewidth, last_of_queue=false)
     slice = items[i:j]
-    cost_list = map(x -> item_cost(x), slice)
+    last = items[j]
-    sum_of_cost_list = reduce((x, y)-> x+y, cost_list)
+    if typeof(last) == u.ChBox && last_of_queue === false
+        return Inf
+    end
+    sum_of_cost_list = 0
+
+    # add disc (after) if it exists
+    
+    if i > 2 && typeof(items[i-1]) == u.Disc
+        previous = items[i-1]
+        sum_of_cost_list += item_width(previous.after)
+    end
+
+    if typeof(last) != u.Disc
+        cost_list = map(x -> item_width(x), slice)
+        sum_of_cost_list += reduce((x, y)-> x+y, cost_list) - item_width(slice[end])
+    else
+        cost_list = map(x -> item_width(x), slice[1:end-1])
+        if cost_list != []
+            sum_of_cost_list += reduce((x, y)-> x+y, cost_list)
+        end
+        last_width = item_width(last.before)
+        sum_of_cost_list += last_width
+    end
     if sum_of_cost_list > linewidth
         return Inf
     else
-        return linewidth - sum_of_cost_list
+        return (linewidth - sum_of_cost_list) ^ 3
     end
 end
 
-function item_cost(i)
+"""the total cost from 1 to n"""
+function total_cost(items, n, linewidth, last_of_queue=false)
+    if haskey(total_cost_table, n)
+        return total_cost_table[n][1]
+    end
+    # first to nth item cost
+    fst_to_nth_cost = cost(items, 1, n, linewidth, last_of_queue)
+
+    if fst_to_nth_cost < Inf
+        total_cost_table[n] = [fst_to_nth_cost, 0]
+        return fst_to_nth_cost
+    else
+        mininal_cost = +Inf
+        prev_breakpoint = nothing
+        for j in 1:1:(n-1)
+            tmp_cost = total_cost(items, j, linewidth, false) + cost(items, j+1, n, linewidth, last_of_queue)
+            if tmp_cost < mininal_cost
+                mininal_cost = tmp_cost
+                prev_breakpoint = j
+            end
+        end
+
+        total_cost_table[n] = [mininal_cost, prev_breakpoint]
+        return mininal_cost
+    end
+end
+
+"""width of a item"""
+function item_width(i)
     if typeof(i) == u.HGlue
         return i.wd
     elseif typeof(i) == u.ChBox
         return i.wd
+    elseif typeof(i) == u.Disc
+        return item_width(i.orig)
     else
         return 0
     end
 end
 
 function arrange(vbox, env)
-    result_vbox_inner = []
+
+    result_vbox_inner = [u.HBox([],nothing,nothing,nothing,nothing,nothing)]
     linewidth = parse(Int, env["linewidth"])
-    print(cost(vbox.eles[1].eles, 1, 20, linewidth))
+    eles = vbox.eles[1].eles
+    total_cost(eles, length(eles), linewidth, true)
 
-    return vbox
 
+    current_point = length(eles)
+    breakpoint_list = [current_point]
+    while total_cost_table[current_point][2] !== 0.0
+        current_point = total_cost_table[current_point][2]
+        push!(breakpoint_list, current_point)
     end
 
+    breakpoint_list_reversed = reverse(breakpoint_list)
+    for (x, i) in enumerate(1:1:length(eles))
+        item = eles[x]
+        if !(x in breakpoint_list_reversed)
+            if typeof(item) == u.Disc
+                if item.orig != []
+                    push!(result_vbox_inner[end].eles, item.orig)
+                end
+            else
+                push!(result_vbox_inner[end].eles, item)
+            end
+        # x is the last one
+        elseif i ==  length(eles)
+            push!(result_vbox_inner[end].eles, item)
+        # x in breakpoint_list_reversed
+        else
+            if typeof(item) == u.Disc
+                if item.before != []
+                    push!(result_vbox_inner[end].eles, item.before)
+                end
+                push!(result_vbox_inner, u.HBox([],nothing,nothing,nothing,nothing,nothing))
+                if item.after != []
+                    push!(result_vbox_inner[end].eles, item.after)
+                end
+            else
+                push!(result_vbox_inner, u.HBox([],nothing,nothing,nothing,nothing,nothing))
+            end
+        end
+    end
+
+    return result_vbox_inner
+end
+
+
+function position(box_inner, env#=to be used later=#)
+    pages = [] #a subarray is the content of a page
+    orig_y = 700 # it can be derived from env
+    orig_x = 100 # it can be derived from env
+    posX = orig_x #cursor x
+    posY = orig_y #cursor y
+    baselineskip = 30  # it can be derived from env
+    for hbox in box_inner
+        pages = position_chbox(hbox.eles, pages, posX, posY)
+        posX = orig_x
+        posY -= baselineskip
+    end
+    return pages
+end
+
+"""positioning all the chboxes in a HBox"""
+function position_chbox(hbox_eles, pages, posX, posY)
+    for i in hbox_eles
+        if typeof(i) == u.HGlue
+            deltaX = i.wd
+            posX += deltaX
+        else #ChBox
+            deltaX = i.wd 
+            i.x = posX
+            i.y = posY
+            push!(pages, i)
+            posX += deltaX
+        end
+    end
+    return pages
+end
 end

src/hyphenating.jl

@@ -89,9 +89,9 @@ function match_char(ast_item, patterns)
                 for i in splitted
                     push!(final, c.CHAR(i))
                     push!(final, c.SEQ([c.ELE([c.ID("disc")]),
+                                        c.ELE([c.CHAR("-")]),
                                         c.ELE([]),
-                                        c.ELE([]),
+                                        c.ELE([])]))
-                                        c.ELE([c.CHAR("-")])]))
                 end
 
                 final = final[1:end-1]

src/interp.jl

@@ -33,6 +33,11 @@ interp: the intepreter of the uahgi.
 function interp(ast, env, res_box, put_char=true_)
     #println("INTERP", ast)
     @match ast begin
+        c.ID(id) => 
+            begin
+                print("ID____", id)
+                return interp(env[id], env, res_box, true_)
+            end
         c.SEQ([c.ELE([c.ID("def")]),
             c.ELE([c.ID(id)]),val]) => 
                         begin
@@ -87,7 +92,7 @@ function interp(ast, env, res_box, put_char=true_)
                                             font_size)
                                         ex_in_px = charmetrics.wd
                                         (val, _, _) = interp(val, env, res_box, false_)
-                                        val_ex_in_px = ex_in_px * parse(Int,val)
+                                        val_ex_in_px = ex_in_px * parse(Float64,val)
                                         return (val_ex_in_px, env, res_box)
                                     end
         c.SEQ([c.ELE([c.ID("hglue")]),
@@ -96,7 +101,7 @@ function interp(ast, env, res_box, put_char=true_)
                                         (width_evaled, _, _) = interp(width, env, res_box, false_)
                                         (stretch_evaled, _, _) = interp(stretch, env, res_box, false_)
                                         push!(res_box.eles[end].eles,
-                                            u.HGlue(width_evaled, parse(Int, stretch_evaled))
+                                            u.HGlue(width_evaled, parse(Float64, stretch_evaled))
                                         )
                                     end
 
@@ -145,7 +150,6 @@ function interp(ast, env, res_box, put_char=true_)
                 (after_evaled, _, _) = interp(after_item, env, res_box, gen_chbox_)
                 (orig_evaled, _, _) = interp(orig_item, env, res_box, gen_chbox_)
                 ret = u.Disc(before_evaled, after_evaled, orig_evaled)
-                println("RETURN", ret)
                 push!(res_box.eles[end].eles, ret)
                 return (ret, env, res_box)
             
@@ -169,8 +173,8 @@ function interp(ast, env, res_box, put_char=true_)
         # empty item
         [] => return (ast, env, res_box)
         _ => begin
-              println("不知道")
+              println("unknown token", ast)
-              val_evaled = "不知道"
+              val_evaled = nothing
               return (val_evaled, env, res_box)
             end
     end

src/parsing.jl

@@ -45,9 +45,7 @@ part =  seq | comment | space | newline | id | char
 all = (Repeat(part) + Eos()) |> Passes.Classes.PROG
 
 function parse(input)
-    print(input)
     ast = parse_one(input, all)[1]
-    print("\n" * string(ast) * "\n")
     
     
     #print(parse_one(, Pattern(r".b.")))
@@ -63,7 +61,6 @@ function parse(input)
     end
 
     new_ast = Passes.Classes.PROG(ast_val)
-    print(ast_to_string(new_ast))
     return new_ast
 end
 

src/pdfoperating.jl

@@ -101,11 +101,11 @@ function check_char_size(char, font_path, font_size, font_index=0)
 
     metrics = glyphrec.metrics
     # horizonal mode
-    width = metricToPx(metrics.horiAdvance, font_size)
+    width = metricToPx(metrics.horiAdvance, font_size) * 0.75
     #from baseline up to top of the glyph
-    height = metricToPx(metrics.horiBearingY, font_size)
+    height = metricToPx(metrics.horiBearingY, font_size) * 0.75
     #from baseline down to the bottom of the glyph
-    depth = metricToPx(metrics.height - metrics.horiBearingY, font_size)
+    depth = metricToPx(metrics.height - metrics.horiBearingY, font_size) * 0.75
 
 
     return CharMetrics(height, depth, width)
@@ -131,5 +131,15 @@ end
 #    '安',
 #    default_font_path,
 #    20)
+"""generate_pdf"""
+function generate_pdf(box_list, file_path)
+    pdf = createPDF()
+    page = addPage(pdf)
+    for ch in box_list
+        font = load_font(pdf, ch.font_path)
+        put_text(page, ch.char, font, ch.size, ch.x, ch.y)
+    end
+    save_pdf(pdf, file_path)
+end
 
 end

src/uahgi.jl

@@ -2,10 +2,12 @@ module uahgi
 include("parsing.jl")
 include("interp.jl")
 include("arrange.jl")
+include("pdfoperating.jl")
 
 using .Interp
 using .Parsing
 using .Arrange
+using .PDFOperating
 using ArgParse
 
 export ChBox, HGlue
@@ -136,12 +138,9 @@ function main()
     output_box_result = Interp.interp_main(ast, default_env, output_box_orig)
     env = output_box_result[2]
     output_box =  output_box_result[3]
-    print(output_box)
-    #TODO
     arranged = Arrange.arrange(output_box, env)
-
+    unit_positioned = Arrange.position(arranged, env)
-
+    generated_pdf = PDFOperating.generate_pdf(unit_positioned, file_path[1:end-3] * ".pdf")
-
 end
 
 main()