add some function incl. cjk+latn passes

example/ex1.ug

@@ -1,8 +1,12 @@
-{@lang|en}%
+{@def|@lang|en}%
 %{@def|@font|{@quote|FreeSans|AR PL UKai TW}}%
 %{@def|@fontsize|12}%
-%{@def|@linewidth|200} %in px
+%{@def|@linewidth|400} %in px
-%{@def|@spacing|{@hglue|{@ex|1}|2}}%
+%{@def|@hori_align|left} % left| right|center
-%{@def|@cjk_spacing|{@hglue|{@ex|0}|0.001}}%
+%{@def|@spacing|{@hglue|{@ex|0.4}|0.2}}%
+%{@def|@cjk_lat_spacing|{@hglue|{@ex|0.2}|0.2}}%
+%{@def|@cjk_spacing|{@hglue|{@ex|0}|0.2}}%
 %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
+the implementation of typesetting language.{@set|@fontsize|14} However, it is good to give TEXBook a try. 測試一段漢字，究竟會不會斷行呢？「『《臺灣連翹》是吳濁流的著作，涉及臺灣戰後歷史與政治。』」{@set|@font|{@quote|jf open 粉圓 2.0|AR PL UMing TW}}
+
+願主耶穌ê恩典kap眾聖徒saⁿ-kap tī-leh。À-bēng。（啟示錄bóe-á-cha̍t）原起頭上帝創造天kap地，是創世記頭chi̍t cha̍t。

src/arrange.jl

@@ -54,7 +54,13 @@ function total_cost(items, n, linewidth, last_of_queue=false)
         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)
+            total_c = total_cost(items, j, linewidth, false) 
+            cost_tail = cost(items, j+1, n, linewidth, last_of_queue)
+            if cost_tail < +Inf && last_of_queue == true
+                tmp_cost = total_c
+            else
+                tmp_cost = total_c + cost_tail
+            end
             if tmp_cost < mininal_cost
                 mininal_cost = tmp_cost
                 prev_breakpoint = j
@@ -98,7 +104,11 @@ function arrange(vbox, env)
     for (x, i) in enumerate(1:1:length(eles))
         item = eles[x]
         if !(x in breakpoint_list_reversed)
-            if typeof(item) == u.Disc
+            if typeof(item) == u.Par
+                horizonal_align = env["hori_align"]
+                result_vbox_inner[end].eles = add_aligner(result_vbox_inner[end].eles, horizonal_align)
+                push!(result_vbox_inner, u.HBox([],nothing,nothing,nothing,nothing,nothing))
+            elseif typeof(item) == u.Disc
                 if item.orig != []
                     push!(result_vbox_inner[end].eles, item.orig)
                 end
@@ -107,10 +117,15 @@ function arrange(vbox, env)
             end
         # x is the last one
         elseif i ==  length(eles)
+            if typeof(item) !== u.Par
                 push!(result_vbox_inner[end].eles, item)
+            end
         # x in breakpoint_list_reversed
         else
-            if typeof(item) == u.Disc
+            if typeof(item) == u.Par
+                push!(result_vbox_inner, u.HBox([],nothing,nothing,nothing,nothing,nothing))
+
+            elseif typeof(item) == u.Disc
                 if item.before != []
                     push!(result_vbox_inner[end].eles, item.before)
                 end
@@ -127,6 +142,17 @@ function arrange(vbox, env)
     return result_vbox_inner
 end
 
+function add_aligner(hbox_eles, horizonal_align)
+    aligner_glue = u.HGlue(0.0,10000.0)
+    if horizonal_align == "right"
+        hbox_eles = reverse(push!(reverse(hbox_eles), aligner_glue))
+    elseif horizonal_align == "middle"
+        return hbox_eles
+    else # left/default
+        hbox_eles = push!(hbox_eles, aligner_glue)
+    end
+    return hbox_eles
+end
 
 function position(box_inner, env#=to be used later=#)
     pages = [] #a subarray is the content of a page
@@ -135,19 +161,57 @@ function position(box_inner, env#=to be used later=#)
     posX = orig_x #cursor x
     posY = orig_y #cursor y
     baselineskip = 30  # it can be derived from env
-    for hbox in box_inner
+    linewidth = env["linewidth"]
-        pages = position_chbox(hbox.eles, pages, posX, posY)
+    horizonal_align = env["hori_align"]
+
+    for (idx, hbox) in enumerate(box_inner)
+        hbox_eles = hbox.eles
+        aligner_glue = u.HGlue(0.0,10000.0)
+        if idx == length(box_inner)
+            hbox_eles = add_aligner(hbox_eles, horizonal_align)
+        end
+        residual_space = parse(Float64, linewidth) - width_sum_of_eles(hbox.eles)
+        sum_of_stretch = reduce((x,y) -> x+y, map(hglue_stretch, hbox.eles))
+        space_stretch_ratio =  residual_space / sum_of_stretch
+
+        pages = position_chbox(hbox_eles, pages, posX, posY, space_stretch_ratio)
         posX = orig_x
         posY -= baselineskip
     end
     return pages
 end
 
+function hglue_stretch(ele)
+    if typeof(ele) == u.HGlue
+        return ele.stretch # stretch
+    else
+        return 0
+    end
+
+end
+
+function width_sum_of_eles(eles)
+    width_sum = x -> begin
+        if hasproperty(x, :wd)
+            return x.wd
+        else
+            return 0.0
+        end
+    end
+    if length(eles) == 0
+        return 0
+    elseif length(eles) == 1
+        return width_sum(eles[1])
+    else
+        return reduce((x,y)-> x+y, map(width_sum, eles))
+    end
+end
+
 """positioning all the chboxes in a HBox"""
-function position_chbox(hbox_eles, pages, posX, posY)
+function position_chbox(hbox_eles, pages, posX, posY, space_stretch_ratio)
     for i in hbox_eles
         if typeof(i) == u.HGlue
-            deltaX = i.wd
+            deltaX = i.wd + (i.stretch * space_stretch_ratio)
             posX += deltaX
         else #ChBox
             deltaX = i.wd 

src/hyphenating.jl

@@ -7,7 +7,7 @@ c = Main.uahgi.Parsing.Passes.Classes
 
 function match_lang(item)
     @match item begin
-    c.SEQ([c.ELE([c.ID("lang")]), c.ELE([c.CHAR(v1)])]) => begin
+    c.SEQ([c.ELE([c.ID("def")]),c.ELE([c.ID("lang")]), c.ELE([c.CHAR(v1)])]) => begin
         return v1
         end
     _ => false

src/interp.jl

@@ -35,9 +35,14 @@ function interp(ast, env, res_box, put_char=true_)
     @match ast begin
         c.ID(id) => 
             begin
-                print("ID____", id)
+                #print("ID____", id)
                 return interp(env[id], env, res_box, true_)
             end
+        c.SEQ([c.ID("par")]) => begin
+            par_box = u.Par()
+            push!(res_box.eles[end].eles, par_box)
+            return interp("", env, res_box)
+        end
         c.SEQ([c.ELE([c.ID("def")]),
             c.ELE([c.ID(id)]),val]) => 
                         begin
@@ -76,11 +81,7 @@ function interp(ast, env, res_box, put_char=true_)
                             return (ele, env, res_box)
             
                         end
-        c.SEQ([c.ELE([c.ID("par")])]) =>begin
+
-                                            push!(res_box.eles, u.HBox([],
-                                             nothing, nothing, nothing, nothing, nothing))
-                                            return (ast, env, res_box)
-                                        end
         c.SEQ([c.ELE([c.ID("ex")]), val]) =>begin
                                         x = 'x'
                                         font_family = select_font(x, env)
@@ -137,7 +138,6 @@ function interp(ast, env, res_box, put_char=true_)
                         
                         end
                     end
-
         c.SEQ([c.ELE([c.ID("disc")]),
                 c.ELE(before),
                 c.ELE(after),
@@ -172,8 +172,10 @@ function interp(ast, env, res_box, put_char=true_)
                         end
         # empty item
         [] => return (ast, env, res_box)
+        "" => return (ast, env, res_box)
+
         _ => begin
-              println("unknown token", ast)
+              println("unknown token: ", ast)
               val_evaled = nothing
               return (val_evaled, env, res_box)
             end

src/passes.jl

@@ -24,16 +24,74 @@ two_nl_to_par_pass = Pass(two_nl_to_par_pattern,
     two_nl_to_par_pass_func)
 push!(processed_passes, two_nl_to_par_pass)
 
+# hyphen between 2 chars to disc
+function discretize_hyphen(two_nl)
+    inner = Classes.SEQ([Classes.ELE([Classes.ID("disc")]),
+        Classes.ELE([Classes.CHAR("-")]),
+        Classes.ELE([]),
+        Classes.ELE([Classes.CHAR("-")])])
+    return [two_nl[1]; inner; two_nl[3]]
+end
+hyphen_pattern = [Classes.CHAR(r"[^\-]"),
+    Classes.CHAR(r"[-]"), Classes.CHAR(r"[^\-]")]
+
+hyphen_disc_pass = Pass(hyphen_pattern,
+    discretize_hyphen)
+push!(processed_passes, hyphen_disc_pass)
+
+adjacent_cjk_pattern = [Classes.CHAR(r"[\p{Han}]"),
+    Classes.CHAR(r"[\p{Han}]")]
+
+# in latin+cjk and latin+cjk add glue.
+function insert_hglue_in_adjacent_cjk_lat(two_nl)
+    inner = Classes.ID("cjk_lat_spacing")
+    return [two_nl[1]; inner; two_nl[2]]
+end
+adjacent_cjk_lat_pattern = [Classes.CHAR(r"[\p{Han}]"),
+    Classes.CHAR(r"[^\p{Han}·，。！？：」』》】］〕』〗〉｝…—「『《〔［【『〖〈｛]")]
+adjacent_cjk_lat_pattern2 = [Classes.CHAR(r"[^\p{Han}·，。！？：」』》】］〕』〗〉｝…—「『《〔［【『〖〈｛]"),
+    Classes.CHAR(r"[\p{Han}]")]
+
+adjacent_cjk_lat_pass = Pass(adjacent_cjk_lat_pattern,
+    insert_hglue_in_adjacent_cjk_lat)
+push!(processed_passes, adjacent_cjk_lat_pass)
+
+adjacent_cjk_lat2_pass = Pass(adjacent_cjk_lat_pattern2,
+    insert_hglue_in_adjacent_cjk_lat)
+push!(processed_passes, adjacent_cjk_lat2_pass)
+
 # in 2 hanzi add glue.
-function insert_hglue_in_adjacent_chinese(two_nl)
+function insert_hglue_in_adjacent_cjk(two_nl)
     inner = Classes.ID("cjk_spacing")
     return [two_nl[1]; inner; two_nl[2]]
 end
-adjacent_chinese_pattern = [Classes.CHAR(r"[\p{Han}，。！？：「」『』…]"),
+adjacent_cjk_pattern = [Classes.CHAR(r"[\p{Han}]"),
-    Classes.CHAR(r"[\p{Han}，。！？：「」『』…]")]
+    Classes.CHAR(r"[\p{Han}]")]
 
-adjacent_glue_pass = Pass(adjacent_chinese_pattern,
+adjacent_cjk_pass = Pass(adjacent_cjk_pattern,
-insert_hglue_in_adjacent_chinese)
+insert_hglue_in_adjacent_cjk)
+push!(processed_passes, adjacent_cjk_pass)
+
+# line breaking rule in CJK 避頭尾/禁則処理
+adjacent_cjk_punc_pattern = [Classes.CHAR(r"[·，。！？：」』》】］〕』〗〉｝]"),
+    Classes.CHAR(r"[^·，。！？：」』》】］〕』〗〉｝]")]
+
+adjacent_cjk_punc_pattern2 = [Classes.CHAR(r"[^「『《〔［【『〖〈｛]"),
+    Classes.CHAR(r"[「『《〔［【『〖〈｛]")]
+adjacent_cjk_punc_pattern3 = [Classes.CHAR(r"[…—]"),
+    Classes.CHAR(r"[^…—]")]
+
+adjacent_glue_pass = Pass(adjacent_cjk_pattern,
+insert_hglue_in_adjacent_cjk)
 push!(processed_passes, adjacent_glue_pass)
 
+adjacent_glue_pass2 = Pass(adjacent_cjk_punc_pattern2,
+insert_hglue_in_adjacent_cjk)
+push!(processed_passes, adjacent_glue_pass2)
+
+adjacent_glue_pass3 = Pass(adjacent_cjk_punc_pattern3,
+insert_hglue_in_adjacent_cjk)
+push!(processed_passes, adjacent_glue_pass3)
+
+
 end

src/uahgi.jl

@@ -45,13 +45,19 @@ end
 """
 Horizonal Glue (HGlue)
 - wd : width
-- str : stretch
+- stretch : stretch
 """
 mutable struct HGlue<:Box
     wd
-    str
+    stretch
 end
 
+"""
+Par
+- for paragraph marker.
+"""
+mutable struct Par<:Box
+end
 
 """
 Vertical Box