From 07df4b80eb5719e1ad9f7c8109d7fb51949601a8 Mon Sep 17 00:00:00 2001
From: Erez Shinan <erezshin+git@gmail.com>
Date: Mon, 20 Feb 2017 00:11:09 +0200
Subject: [PATCH] My Earley parser is now working. Not yet plugged in

---
 lark/parsers/earley2.py | 80 ++++++++++++++++++++++++++---------------
 1 file changed, 52 insertions(+), 28 deletions(-)

diff --git a/lark/parsers/earley2.py b/lark/parsers/earley2.py
index 80144f9..c41dfa5 100644
--- a/lark/parsers/earley2.py
+++ b/lark/parsers/earley2.py
@@ -1,11 +1,13 @@
-from ..utils import classify, classify_bool, bfs, fzset
-from ..common import GrammarError, is_terminal
-from lalr_analysis import Rule, RulePtr, GrammarAnalyzer
+from ..common import ParseError, UnexpectedToken, is_terminal
+from lalr_analysis import GrammarAnalyzer
+
+from ..tree import Tree
 
 class Item:
-    def __init__(self, rule_ptr, start):
+    def __init__(self, rule_ptr, start, data):
         self.rule_ptr = rule_ptr
         self.start = start
+        self.data = data
 
     @property
     def expect(self):
@@ -15,8 +17,12 @@ class Item:
     def is_complete(self):
         return self.rule_ptr.is_satisfied
 
-    def advance(self):
-        return Item(self.rule_ptr.advance(self.expect), self.start)
+    @property
+    def name(self):
+        return self.rule_ptr.rule.origin
+
+    def advance(self, data):
+        return Item(self.rule_ptr.advance(self.expect), self.start, self.data + [data])
 
     def __eq__(self, other):
         return self.rule_ptr == other.rule_ptr and self.start == other.start
@@ -26,6 +32,7 @@ class Item:
     def __repr__(self):
         return '%s (%s)' % (self.rule_ptr, self.start)
 
+
 class Parser:
     def __init__(self, rules, start):
         self.analyzer = GrammarAnalyzer(rules, start)
@@ -37,19 +44,19 @@ class Parser:
 
         def predict(symbol, i):
             assert not is_terminal(symbol), symbol
-            return {Item(rp, i) for rp in self.analyzer.expand_rule(symbol)}
+            return {Item(rp, i, []) for rp in self.analyzer.expand_rule(symbol)}
 
         def scan(item, inp):
             if item.expect == inp:   # TODO Do a smarter match, i.e. regexp
-                return {item.advance()}
+                return {item.advance(inp)}
             else:
                 return set()
 
         def complete(item, table):
-            print "Complete:", item
-            name = item.rule_ptr.rule.origin
-            return {old_item.advance() for old_item in table[item.start]
-                                       if old_item.expect == name}
+            name = item.name
+            item.data = Tree(name, item.data)
+            return {old_item.advance(item.data) for old_item in table[item.start]
+                    if not old_item.is_complete and old_item.expect == name}
 
         def process_column(i, char):
             cur_set = table[-1]
@@ -71,6 +78,10 @@ class Parser:
                 to_process = new_items - cur_set
                 cur_set |= to_process
 
+            if not next_set and char != '$end':
+                expect = filter(is_terminal, [i.expect for i in cur_set if not i.is_complete])
+                raise UnexpectedToken(char, expect, stream, i)
+
         # Main loop starts
 
         table = [predict(self.start, 0)]
@@ -78,7 +89,16 @@ class Parser:
         for i, char in enumerate(stream):
             process_column(i, char)
 
-        process_column(len(stream), None)
+        process_column(len(stream), '$end')
+
+        # Parse ended. Now build a parse tree
+        solutions = [n.data for n in table[len(stream)]
+                     if n.is_complete and n.name==self.start and n.start==0]
+
+        if not solutions:
+            raise ParseError('Incomplete parse: Could not find a solution to input')
+
+        return solutions
 
 
 
@@ -86,22 +106,26 @@ class Parser:
 
 # rules = [
 #     ('a', ['a', 'A']),
+#     ('a', ['a', 'A', 'a']),
+#     ('a', ['a', 'A', 'A', 'a']),
 #     ('a', ['A']),
 # ]
 
 # p = Parser(rules, 'a')
-# p.parse('AAA')
-
-rules = [
-    ('sum', ['sum', "A", 'product']),
-    ('sum', ['product']),
-    ('product', ['product', "M", 'factor']),
-    ('product', ['factor']),
-    ('factor', ['L', 'sum', 'R']),
-    ('factor', ['number']),
-    ('number', ['N', 'number']),
-    ('number', ['N']),
-]
-
-p = Parser(rules, 'sum')
-p.parse('NALNMNANR')
+# for x in p.parse('AAAA'):
+#     print '->'
+#     print x.pretty()
+
+# rules = [
+#     ('sum', ['sum', "A", 'product']),
+#     ('sum', ['product']),
+#     ('product', ['product', "M", 'factor']),
+#     ('product', ['factor']),
+#     ('factor', ['L', 'sum', 'R']),
+#     ('factor', ['number']),
+#     ('number', ['N', 'number']),
+#     ('number', ['N']),
+# ]
+
+# p = Parser(rules, 'sum')
+# print p.parse('NALNMNANR')