Make pret-lam support superscripts.

[matthijs/master-project/report.git] / pret-lam.lua

diff --git a/pret-lam.lua b/pret-lam.lua
index a87137c0dc0b2428a1484552fe7d0357781f2fcf..77fcc5dcda8fdab9b456ac95fcb61e65b4d2f5de 100644 (file)
--- a/pret-lam.lua
+++ b/pret-lam.lua
@@ -17,51 +17,101 @@ buffers.visualizers.lam.colors = {
     "prettyfour"
 }
 
-buffers.visualizers.lam.tokens = {
+buffers.visualizers.lam.symbols = {
     [' '] = {repr = '\\obs '},
     ['_'] = {repr = '\\_'},
     ['->'] = {repr = '\\rightarrow'},
     -- The default * sits very high above the baseline, \ast (u+2217) looks
     -- better.
     ['*'] = {repr = '\\ast'},
-    ['case'] = {style = 'bold'},
-    ['of'] = {style = 'bold'},
-    ['let'] = {style = 'bold'},
-    ['in'] = {style = 'bold'},
 }
 
+buffers.visualizers.lam.keywords = {
+    ['case'] = {},
+    ['of'] = {},
+    ['let'] = {},
+    ['in'] = {},
+}
+
+
+function buffers.visualizers.lam.take_symbol(str)
+    -- See if str starts with a symbol, and return the remaining
+    -- string and that symbol. If no symbol from the table is matched,
+    -- just returns the first character.
+    -- We can do a lookup directly, since symbols can be different in
+    -- length, so we just loop over all symbols, trying them in turn.
+    for symbol,props in pairs(buffers.visualizers.lam.symbols) do
+        -- Try to remove symbol from the start of str 
+        symbol, newstr = utf.match(str, "^(" .. symbol .. ")(.*)")
+        if symbol then
+            -- Return this tokens repr, or just the token if it has no
+            -- repr.
+            res = props.repr or symbol
+            -- Enclose the token in {\style .. }
+            if props.style then
+                res = "{\\" .. props.style ..  " " .. res ..  "}"
+            end
+            return res, newstr
+        end
+    end
+    -- No symbol found, just return the first character
+    return utf.match(str, "^(.)(.*)")
+end
+
+-- Take a single word from str, if posible. Returns the rest of the string and
+-- the word taken.
+function buffers.visualizers.lam.take_word(str)
+        res, newstr = utf.match(str, "^(%a[%a%d_]+)(.*)")
+        return res, newstr or str
+end
+
+-- Tries to match each of the patterns and returns the captures of the first
+-- matching pattern (up to 5 captures are supported). Returns nil when nothing
+-- matches.
+function buffers.visualizers.lam.match_mul(str, patterns)
+    for i, pat in ipairs(patterns) do
+        a, b, c, d, e = utf.match(str, pat)
+        if a then
+            return a, b, c, d, e
+        end
+    end
+    return nil
+end
+
 function buffers.visualizers.lam.flush_line(str,nested)
     local result, state = { }, 0
     local finish, change = buffers.finish_state, buffers.change_state
+    local take_symbol = buffers.visualizers.lam.take_symbol
+    local take_word = buffers.visualizers.lam.take_word
+    local match_mul = buffers.visualizers.lam.match_mul
     -- Set the colorscheme, which is used by finish_state and change_state
     buffers.currentcolors = buffers.visualizers.lam.colors
 
     while str ~= "" do
-        found = false
-        for tok,props in pairs(buffers.visualizers.lam.tokens) do
-            -- Try to remove tok from the start of str 
-            str, count = utf.gsub(str, "^" .. tok, "")
-            if count ~= 0 then
-                -- Replace the token with the repr, or just itself
-                result[#result+1] = props.repr or tok
-                -- Enclose the token in {\style .. }
-                if props.style then
-                    result[#result] = "{\\" .. props.style ..
-                                        " " .. result[#result] ..
-                                        "}"
+        local found = false
+        local word, symbol
+        -- See if the next token is a word
+        word, str = take_word(str)
+        if word then
+            if buffers.visualizers.lam.keywords[res] then
+                -- Make all keywords bold
+                word = "{\\bold " .. word ..  "}"
+            else
+                -- Allow subscripts using _ or just appending a number.
+                base, sub = match_mul(res, {"^(.*)_([%a%d,]+)$", "^(.*[^%d])(%d*)$"})
+                if sub then
+                    word = base .. "\\low{" .. sub .. "}"
                 end
-                -- We found a token, now start over
-                found = true
-                break
             end
+        else
+            -- The next token is not a word, it must be a symbol
+            symbol, str = take_symbol(str)
         end
-        if not found then
-            -- Take a full word, or a single character. Taking a full word is
-            -- to prevent matching "in" in "bind" , for example.
-            result[#result+1] = utf.match(str, "^%a+") or utf.sub(str, 1, 1)
-            str = utf.sub(str, utf.len(result[#result]) + 1)
-        end
+
+        -- Append the resulting token
+        result[#result+1] = word or symbol
     end
+
     state = finish(state, result)
     buffers.flush_result(result,nested)
 end