Add interactive parser ability, tidy up the Python test and skeleton slightly

diff --git a/lr1.py b/lr1.py
index 7fba214..c17def1 100644 (file)
--- a/lr1.py
+++ b/lr1.py
@@ -338,7 +338,10 @@ class LR1:
 
     while len(_lr1dfa.states) < len(state_to_items):
       items = state_to_items[len(_lr1dfa.states)]
-      state_desc = ([], [], [], [])
+      terminal_breaks = []
+      actions = []
+      nonterminal_breaks = []
+      gotos = []
 
       def add_state(next_items, next_item_to_index):
         self.lookahead_item_set_closure(next_items, next_item_to_index)
@@ -407,8 +410,8 @@ class LR1:
           action = add_state(next_items, next_item_to_index) * 2 # shift
         else:
           action = -1
-        state_desc[0].append(terminal1)
-        state_desc[1].append(action)
+        terminal_breaks.append(terminal1)
+        actions.append(action)
         terminal = terminal1
 
       nonterminal = 0
@@ -421,10 +424,28 @@ class LR1:
           goto = add_state(next_items, next_item_to_index)
         else:
           goto = -1
-        state_desc[2].append(nonterminal1)
-        state_desc[3].append(goto)
+        nonterminal_breaks.append(nonterminal1)
+        gotos.append(goto)
         nonterminal = nonterminal1
 
+      # test for a default reduce action, useful for interactive parsers
+      valid_actions = [i for i in actions if i != -1]
+      default_reduce = -1
+      if len(valid_actions):
+        valid_action = valid_actions[0]
+        if (
+          (valid_action & 1) and
+          all([i == valid_action for i in valid_actions[1:]])
+        ):
+          default_reduce = valid_action >> 1
+
+      state_desc = (
+        terminal_breaks,
+        actions,
+        nonterminal_breaks,
+        gotos,
+        default_reduce
+      )
       _lr1dfa.states.append(state_desc)
     return _lr1dfa
 
@@ -456,7 +477,10 @@ class LR1:
       i = queue[qhead]
       in_queue[i] = False
       items = state_to_items[i]
-      state_desc = ([], [], [], [])
+      terminal_breaks = []
+      actions = []
+      nonterminal_breaks = []
+      gotos = []
 
       def add_state(next_items, next_item_to_index):
         self.lookahead_item_set_closure(next_items, next_item_to_index)
@@ -536,8 +560,8 @@ class LR1:
           action = add_state(next_items, next_item_to_index) * 2 # shift
         else:
           action = -1
-        state_desc[0].append(terminal1)
-        state_desc[1].append(action)
+        terminal_breaks.append(terminal1)
+        actions.append(action)
         terminal = terminal1
 
       nonterminal = 0
@@ -550,10 +574,28 @@ class LR1:
           goto = add_state(next_items, next_item_to_index)
         else:
           goto = -1
-        state_desc[2].append(nonterminal1)
-        state_desc[3].append(goto)
+        nonterminal_breaks.append(nonterminal1)
+        gotos.append(goto)
         nonterminal = nonterminal1
 
+      # test for a default reduce action, useful for interactive parsers
+      valid_actions = [i for i in actions if i != -1]
+      default_reduce = -1
+      if len(valid_actions):
+        valid_action = valid_actions[0]
+        if (
+          (valid_action & 1) and
+          all([i == valid_action for i in valid_actions[1:]])
+        ):
+          default_reduce = valid_action >> 1
+
+      state_desc = (
+        terminal_breaks,
+        actions,
+        nonterminal_breaks,
+        gotos,
+        default_reduce
+      )
       if i < len(_lr1dfa.states):
         _lr1dfa.states[i] = state_desc
       else:

diff --git a/lr1dfa.py b/lr1dfa.py
index a5d4086..03cb711 100644 (file)
--- a/lr1dfa.py
+++ b/lr1dfa.py
@@ -17,7 +17,7 @@ class LR1DFA:
     eof_terminal = n_characters
   ):
     # states: list of state_desc
-    # state_desc: (terminal breaks, actions, nonterminal breaks, gotos)
+    # state_desc: (terminal br, actions, nonterminal br, gotos, default reduce)
     # action: shift = new state * 2, reduce = production * 2 + 1, error = -1
     # goto: reduce = production, error = -1 (but error can't really happen)
     # productions: list of production
@@ -80,7 +80,7 @@ class LR1DFA:
       numpy.int16
     )
     for i in range(len(self.states)):
-      terminal_breaks, actions, nonterminal_breaks, gotos = self.states[i]
+      terminal_breaks, actions, nonterminal_breaks, gotos, _ = self.states[i]
 
       terminal0 = 0
       for j in range(len(terminal_breaks)):

diff --git a/skel/skel_py.py b/skel/skel_py.py
index 028f25f..c80d2e3 100644 (file)
--- a/skel/skel_py.py
+++ b/skel/skel_py.py
@@ -1,5 +1,4 @@
 import bisect
-import lex_yy
 
 # GENERATE SECTION1
 
@@ -24,39 +23,42 @@ def yyparse():
   state = 0
   yystack = []
   yylval = None
-  yytoken = lex_yy.yylex()
-  print('yytoken', yytoken, 'yylval', yylval)
+  yytoken = -1
   while True:
-    print('state', state, 'yystack', yystack)
-    action = yy_lr1dfa_states[state][1][
-      bisect.bisect_right(yy_lr1dfa_states[state][0], yytoken)
-    ]
-    if action == -1:
-      raise Exception('syntax error')
-    if (action & 1) == 0:
-      yystack.append((state, yylval))
-      state = action >> 1
-      print('shift', state)
-      yylval = None
-      yytoken = lex_yy.yylex()
-      print('yytoken', yytoken, 'yylval', yylval)
-    else:
-      yystack.append((state, None))
-      reduce = action >> 1
-      print('reduce', reduce)
-      len_symbols, ref_data = yy_lr1dfa_productions[reduce]
-      base = len(yystack) - len_symbols - 1
-      yyval = yystack[base][1]
-      ref_data()
-      del yystack[base + 1:]
-      if reduce == 0:
-        assert base == 0
-        return
-      state = yystack[-1][0]
-      yystack[-1] = (state, yyval)
-      state = yy_lr1dfa_states[state][3][
-        bisect.bisect_right(yy_lr1dfa_states[state][2], reduce)
+    #print('state', state, 'yystack', yystack)
+    reduce = yy_lr1dfa_states[state][4]
+    if reduce == -1:
+      if yytoken == -1:
+        yylval = None
+        yylloc = None
+        yytoken = yylex()
+        #print('yytoken', yytoken, 'yylval', yylval, 'yylloc', yylloc)
+      action = yy_lr1dfa_states[state][1][
+        bisect.bisect_right(yy_lr1dfa_states[state][0], yytoken)
       ]
-      assert state != -1
+      if action == -1:
+        raise Exception('syntax error')
+      if (action & 1) == 0:
+        yystack.append((state, yylval, yylloc))
+        state = action >> 1
+        #print('shift', state)
+        yytoken = -1
+        continue
+      reduce = action >> 1
+    #print('reduce', reduce)
+    len_symbols, ref_data = yy_lr1dfa_productions[reduce]
+    base = len(yystack) - len_symbols
+    yystack.append((state, None, None))
+    state, yyval, yyloc = yystack[base]
+    ref_data()
+    del yystack[base:]
+    if reduce == 0:
+      assert base == 0
+      break
+    yystack.append((state, yyval, yyloc))
+    state = yy_lr1dfa_states[state][3][
+      bisect.bisect_right(yy_lr1dfa_states[state][2], reduce)
+    ]
+    assert state != -1
 
 # GENERATE SECTION3

diff --git a/tests/cal.py b/tests/cal.py
index d2978ba..d43ab5c 100755 (executable)
--- a/tests/cal.py
+++ b/tests/cal.py
@@ -1,6 +1,8 @@
 #!/usr/bin/env python3
 
+import sys
 import y_tab
 
-print('Enter the expression: ', end = '')
+sys.stdout.write('Enter the expression: ')
+sys.stdout.flush()
 y_tab.yyparse()

diff --git a/tests/cal_py.l b/tests/cal_py.l
index fcee582..272fba6 100644 (file)
--- a/tests/cal_py.l
+++ b/tests/cal_py.l
@@ -4,8 +4,6 @@ import y_tab
 
 DIGIT [0-9]+\.?|[0-9]*\.[0-9]+
 
-%option noecs nometa-ecs noyywrap reject yymore
-
 %%
 
 [ ]

diff --git a/tests/cal_py.y b/tests/cal_py.y
index f2b532a..4fea55a 100644 (file)
--- a/tests/cal_py.y
+++ b/tests/cal_py.y
@@ -1,4 +1,5 @@
 %{
+from lex_yy import yylex
 import sys
 %}
 %token NUM
@@ -10,8 +11,7 @@ import sys
 %%
 
 S : S E '\n' {
-  print('Answer:', $2)
-  print('Enter:')
+  sys.stdout.write('Answer: {0:g}\nEnter:\n'.format($2))
 }
   | S '\n'
   |
@@ -30,9 +30,7 @@ E : E '+' E { $$ = $1 + $3 }
   | NUM
   ;
 %%
-# this is section 3
 
 def yyerror(s):
-  print(s)
+  sys.stdout.write('{0:s}\n'.format(s))
   sys.exit(1)
-