Use the transformation pretty printer.

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

diff --git a/pret-lam.lua b/pret-lam.lua
index 77fcc5dcda8fdab9b456ac95fcb61e65b4d2f5de..f14b962ba2cd162a8f9743e73aeef8e30b695776 100644 (file)
--- a/pret-lam.lua
+++ b/pret-lam.lua
@@ -17,6 +17,7 @@ buffers.visualizers.lam.colors = {
     "prettyfour"
 }
 
+-- Symbols that should have a different representation
 buffers.visualizers.lam.symbols = {
     [' '] = {repr = '\\obs '},
     ['_'] = {repr = '\\_'},
@@ -24,8 +25,11 @@ buffers.visualizers.lam.symbols = {
     -- The default * sits very high above the baseline, \ast (u+2217) looks
     -- better.
     ['*'] = {repr = '\\ast'},
+    ['~'] = {repr = '\\sim'},
 }
 
+
+-- Keywords that should be bold
 buffers.visualizers.lam.keywords = {
     ['case'] = {},
     ['of'] = {},
@@ -33,13 +37,11 @@ buffers.visualizers.lam.keywords = {
     ['in'] = {},
 }
 
-
+-- 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.
 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 .. ")(.*)")
@@ -61,7 +63,9 @@ 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_]+)(.*)")
+        -- A word must always start with a-z (in particular, λ is not a valid
+        -- start of a word).
+        res, newstr = utf.match(str, "^([a-zA-Z][%a%d_]+)(.*)")
         return res, newstr or str
 end
 
@@ -78,15 +82,42 @@ function buffers.visualizers.lam.match_mul(str, patterns)
     return nil
 end
 
+-- Find any subscripts in the given word and typeset them
+function buffers.visualizers.lam.do_subscripts(word)
+    local match_mul = buffers.visualizers.lam.match_mul
+    base, sub = match_mul(res, submatches)
+    if sub then
+        word = base .. "\\low{" .. sub .. "}"
+        -- After a word has been used as a base, allow subscripts
+        -- without _, even for non-numbers.
+        if not bases[base] then
+            -- Register that we've added this base
+            bases[base] = true
+            -- Add a pattern for this base
+            submatches[#submatches+1] = "^(" .. base .. ")([%a%d,]+)$"
+        end
+    end
+    return word
+end
+
+function buffers.visualizers.lam.begin_of_buffer(type, name)
+    -- Initially allow subscripts using _ or just appending a number (later,
+    -- we will add extra patterns here.
+    submatches = {"^(.*)_([%a%d,]+)$", "^(.*[^%d])(%d+)$"}
+    -- This stores all the bases we've encountered so far (to prevent
+    -- duplicates). For each of them there will be a pattern in submatches
+    -- above.
+    bases = {}
+end
+
 function buffers.visualizers.lam.flush_line(str,nested)
     local result, state = { }, 0
     local finish, change = buffers.finish_state, buffers.change_state
     local take_symbol = buffers.visualizers.lam.take_symbol
     local take_word = buffers.visualizers.lam.take_word
-    local match_mul = buffers.visualizers.lam.match_mul
+    local do_subscripts = buffers.visualizers.lam.do_subscripts
     -- Set the colorscheme, which is used by finish_state and change_state
     buffers.currentcolors = buffers.visualizers.lam.colors
-
     while str ~= "" do
         local found = false
         local word, symbol
@@ -97,11 +128,8 @@ function buffers.visualizers.lam.flush_line(str,nested)
                 -- Make all keywords bold
                 word = "{\\bold " .. word ..  "}"
             else
-                -- Allow subscripts using _ or just appending a number.
-                base, sub = match_mul(res, {"^(.*)_([%a%d,]+)$", "^(.*[^%d])(%d*)$"})
-                if sub then
-                    word = base .. "\\low{" .. sub .. "}"
-                end
+                -- Process any subscripts in the word
+                word = do_subscripts(word)
             end
         else
             -- The next token is not a word, it must be a symbol