-# Helper function
-def Assign(left, right):
- names = []
- if isinstance(left, ast.Name):
- # Single assignment on left
- return ast.Assign([ast.Name(left.id, ast.Store())], right)
- elif isinstance(left, ast.Tuple):
- # List of things - make sure they are Name nodes
- names = []
- for child in left.getChildren():
- if not isinstance(child, ast.Name):
- raise SyntaxError("that assignment not supported")
- names.append(child.name)
- ass_list = [ast.AssName(name, 'OP_ASSIGN') for name in names]
- return ast.Assign([ast.AssTuple(ass_list)], right)
- else:
- raise SyntaxError("Can't do that yet")
-
-
-# The grammar comments come from Python's Grammar/Grammar file
-
-## NB: compound_stmt in single_input is followed by extra NEWLINE!
-# file_input: (NEWLINE | stmt)* ENDMARKER
-def p_file_input_end(p):
- """file_input_end : file_input ENDMARKER"""
- print ("end", p[1])
- p[0] = p[1]
-
-def p_file_input(p):
- """file_input : file_input NEWLINE
- | file_input stmt
- | NEWLINE
- | stmt"""
- if isinstance(p[len(p)-1], str):
- if len(p) == 3:
- p[0] = p[1]
- else:
- p[0] = [] # p == 2 --> only a blank line
- else:
- if len(p) == 3:
- p[0] = p[1] + p[2]
- else:
- p[0] = p[1]
-
-
-# funcdef: [decorators] 'def' NAME parameters ':' suite
-# ignoring decorators
-def p_funcdef(p):
- "funcdef : DEF NAME parameters COLON suite"
- p[0] = ast.Function(None, p[2], list(p[3]), (), 0, None, p[5])
-
-# parameters: '(' [varargslist] ')'
-def p_parameters(p):
- """parameters : LPAR RPAR
- | LPAR varargslist RPAR"""
- if len(p) == 3:
- p[0] = []
- else:
- p[0] = p[2]
-
-
-# varargslist: (fpdef ['=' test] ',')* ('*' NAME [',' '**' NAME] | '**' NAME) |
-# highly simplified
-def p_varargslist(p):
- """varargslist : varargslist COMMA NAME
- | NAME"""
- if len(p) == 4:
- p[0] = p[1] + p[3]
- else:
- p[0] = [p[1]]
-
-# stmt: simple_stmt | compound_stmt
-def p_stmt_simple(p):
- """stmt : simple_stmt"""
- # simple_stmt is a list
- p[0] = p[1]
-
-def p_stmt_compound(p):
- """stmt : compound_stmt"""
- p[0] = [p[1]]
-
-# simple_stmt: small_stmt (';' small_stmt)* [';'] NEWLINE
-def p_simple_stmt(p):
- """simple_stmt : small_stmts NEWLINE
- | small_stmts SEMICOLON NEWLINE"""
- p[0] = p[1]
-
-def p_small_stmts(p):
- """small_stmts : small_stmts SEMICOLON small_stmt
- | small_stmt"""
- if len(p) == 4:
- p[0] = p[1] + [p[3]]
- else:
- p[0] = [p[1]]
-
-# small_stmt: expr_stmt | print_stmt | del_stmt | pass_stmt | flow_stmt |
-# import_stmt | global_stmt | exec_stmt | assert_stmt
-def p_small_stmt(p):
- """small_stmt : flow_stmt
- | break_stmt
- | expr_stmt"""
- p[0] = p[1]
-
-# expr_stmt: testlist (augassign (yield_expr|testlist) |
-# ('=' (yield_expr|testlist))*)
-# augassign: ('+=' | '-=' | '*=' | '/=' | '%=' | '&=' | '|=' | '^=' |
-# '<<=' | '>>=' | '**=' | '//=')
-def p_expr_stmt(p):
- """expr_stmt : testlist ASSIGN testlist
- | testlist """
- if len(p) == 2:
- # a list of expressions
- #p[0] = ast.Discard(p[1])
- p[0] = p[1]
- else:
- p[0] = Assign(p[1], p[3])
-
-def p_flow_stmt(p):
- "flow_stmt : return_stmt"
- p[0] = p[1]
-
-# return_stmt: 'return' [testlist]
-def p_return_stmt(p):
- "return_stmt : RETURN testlist"
- p[0] = ast.Return(p[2])
-
-
-def p_compound_stmt(p):
- """compound_stmt : if_stmt
- | while_stmt
- | for_stmt
- | funcdef
- """
- p[0] = p[1]
-
-def p_break_stmt(p):
- """break_stmt : BREAK
- """
- p[0] = ast.Break()
-
-def p_for_stmt(p):
- """for_stmt : FOR test FOREQ test TO test COLON suite
- """
- p[0] = ast.While(p[2], p[4], [])
- # auto-add-one (sigh) due to python range
- start = p[4]
- end = ast.BinOp(p[6], ast.Add(), ast.Constant(1))
- it = ast.Call(ast.Name("range"), [start, end], [])
- p[0] = ast.For(p[2], it, p[8], [])
-
-def p_while_stmt(p):
- """while_stmt : DO WHILE test COLON suite ELSE COLON suite
- | DO WHILE test COLON suite
- """
- if len(p) == 6:
- p[0] = ast.While(p[3], p[5], [])
- else:
- p[0] = ast.While(p[3], p[5], p[8])
-
-def p_if_stmt(p):
- """if_stmt : IF test COLON suite ELSE COLON suite
- | IF test COLON suite
- """
- if len(p) == 5:
- p[0] = ast.If(p[2], p[4], [])
- else:
- p[0] = ast.If(p[2], p[4], p[7])
-
-def p_suite(p):
- """suite : simple_stmt
- | NEWLINE INDENT stmts DEDENT"""
- if len(p) == 2:
- p[0] = p[1]
- else:
- p[0] = p[3]
-
-
-def p_stmts(p):
- """stmts : stmts stmt
- | stmt"""
- if len(p) == 3:
- p[0] = p[1] + p[2]
- else:
- p[0] = p[1]
-
-## No using Python's approach because Ply supports precedence
-
-# comparison: expr (comp_op expr)*
-# arith_expr: term (('+'|'-') term)*
-# term: factor (('*'|'/'|'%'|'//') factor)*
-# factor: ('+'|'-'|'~') factor | power
-# comp_op: '<'|'>'|'=='|'>='|'<='|'<>'|'!='|'in'|'not' 'in'|'is'|'is' 'not'
-
-def make_lt_compare(arg):
- (left, right) = arg
- return ast.Compare(left, [ast.Lt()], [right])
-def make_gt_compare(arg):
- (left, right) = arg
- return ast.Compare(left, [ast.Gt()], [right])
-def make_eq_compare(arg):
- (left, right) = arg
- return ast.Compare(left, [ast.Eq()], [right])
-
-
-binary_ops = {
- "+": ast.Add(),
- "-": ast.Sub(),
- "*": ast.Mult(),
- "/": ast.Div(),
- "<": make_lt_compare,
- ">": make_gt_compare,
- "==": make_eq_compare,
-}
-unary_ops = {
- "+": ast.Add,
- "-": ast.Sub,
- }
-precedence = (
- ("left", "EQ", "GT", "LT"),
- ("left", "PLUS", "MINUS"),
- ("left", "MULT", "DIV"),
- )
-
-def check_concat(node): # checks if the comparison is already a concat
- print (node)
- if not isinstance(node, ast.Call):
- return [node]
- if node[0].id != 'concat':
- return node
- return node[1]
-
-def p_comparison(p):
- """comparison : comparison PLUS comparison
- | comparison MINUS comparison
- | comparison MULT comparison
- | comparison DIV comparison
- | comparison LT comparison
- | comparison EQ comparison
- | comparison GT comparison
- | PLUS comparison
- | MINUS comparison
- | comparison APPEND comparison
- | power"""
- if len(p) == 4:
- print (list(p))
- if p[2] == '||':
- l = check_concat(p[1]) + check_concat(p[3])
- p[0] = ast.Call(ast.Name("concat"), l, [])
- elif p[2] in ['<', '>', '==']:
- p[0] = binary_ops[p[2]]((p[1],p[3]))
- else:
- p[0] = ast.BinOp(p[1], binary_ops[p[2]], p[3])
- elif len(p) == 3:
- p[0] = unary_ops[p[1]](p[2])
- else:
- p[0] = p[1]
-
-# power: atom trailer* ['**' factor]
-# trailers enables function calls (and subscripts).
-# I only allow one level of calls
-# so this is 'trailer'
-def p_power(p):
- """power : atom
- | atom trailer"""
- if len(p) == 2:
- p[0] = p[1]
- else:
- if p[2][0] == "CALL":
- p[0] = ast.Expr(ast.Call(p[1], p[2][1], []))
- #if p[1].id == 'print':
- # p[0] = ast.Printnl(ast.Tuple(p[2][1]), None, None)
- #else:
- # p[0] = ast.CallFunc(p[1], p[2][1], None, None)
- else:
- print (p[2][1])
- #raise AssertionError("not implemented %s" % p[2][0])
- subs = p[2][1]
- if len(subs) == 1:
- idx = subs[0]
- else:
- idx = ast.Slice(subs[0], subs[1], None)
- p[0] = ast.Subscript(p[1], idx)
-
-def p_atom_name(p):
- """atom : NAME"""
- p[0] = ast.Name(p[1], ctx=ast.Load())
-
-def p_atom_number(p):
- """atom : NUMBER
- | STRING"""
- p[0] = ast.Constant(p[1])
-
-#'[' [listmaker] ']' |
-
-def p_atom_listmaker(p):
- """atom : LBRACK listmaker RBRACK"""
- p[0] = p[2]
-
-def p_listmaker(p):
- """listmaker : test COMMA listmaker
- | test
- """
- if len(p) == 2:
- p[0] = ast.List([p[1]])
- else:
- p[0] = ast.List([p[1]] + p[3].nodes)
-
-def p_atom_tuple(p):
- """atom : LPAR testlist RPAR"""
- p[0] = p[2]
-
-# trailer: '(' [arglist] ')' | '[' subscriptlist ']' | '.' NAME
-def p_trailer(p):
- """trailer : trailer_arglist
- | trailer_subscript
- """
- p[0] = p[1]
-
-def p_trailer_arglist(p):
- "trailer_arglist : LPAR arglist RPAR"
- p[0] = ("CALL", p[2])
-
-def p_trailer_subscript(p):
- "trailer_subscript : LBRACK subscript RBRACK"
- p[0] = ("SUBS", p[2])
-
-#subscript: '.' '.' '.' | test | [test] ':' [test]
-
-def p_subscript(p):
- """subscript : test COLON test
- | test
- """
- if len(p) == 4:
- p[0] = [p[1], p[3]]
- else:
- p[0] = [p[1]]
-
-
-# testlist: test (',' test)* [',']
-# Contains shift/reduce error
-def p_testlist(p):
- """testlist : testlist_multi COMMA
- | testlist_multi """
- if len(p) == 2:
- p[0] = p[1]
- else:
- # May need to promote singleton to tuple
- if isinstance(p[1], list):
- p[0] = p[1]
- else:
- p[0] = [p[1]]
- # Convert into a tuple?
- if isinstance(p[0], list):
- p[0] = ast.Tuple(p[0])
-
-def p_testlist_multi(p):
- """testlist_multi : testlist_multi COMMA test
- | test"""
- if len(p) == 2:
- # singleton
- p[0] = p[1]
- else:
- if isinstance(p[1], list):
- p[0] = p[1] + [p[3]]
- else:
- # singleton -> tuple
- p[0] = [p[1], p[3]]
-
-
-# test: or_test ['if' or_test 'else' test] | lambdef
-# as I don't support 'and', 'or', and 'not' this works down to 'comparison'
-def p_test(p):
- "test : comparison"
- p[0] = p[1]
-
-
-
-# arglist: (argument ',')* (argument [',']| '*' test [',' '**' test] | '**' test)
-# XXX INCOMPLETE: this doesn't allow the trailing comma
-def p_arglist(p):
- """arglist : arglist COMMA argument
- | argument"""
- if len(p) == 4:
- p[0] = p[1] + [p[3]]
- else:
- p[0] = [p[1]]
-
-# argument: test [gen_for] | test '=' test # Really [keyword '='] test
-def p_argument(p):
- "argument : test"
- p[0] = p[1]
-
-def p_error(p):
- #print "Error!", repr(p)
- raise SyntaxError(p)
-
-
-class GardenSnakeParser(object):
- def __init__(self, lexer = None):
- if lexer is None:
- lexer = IndentLexer(debug=1)
- self.lexer = lexer
- self.parser = yacc.yacc(start="file_input_end",
- debug=False, write_tables=False)
-
- def parse(self, code):
- self.lexer.input(code)
- result = self.parser.parse(lexer = self.lexer, debug=False)
- return ast.Module(result)
-
-
-###### Code generation ######
-
-#from compiler import misc, syntax, pycodegen
-
-class GardenSnakeCompiler(object):
- def __init__(self):
- self.parser = GardenSnakeParser()
- def compile(self, code, mode="exec", filename="<string>"):
- tree = self.parser.parse(code)
- print ("snake")
- pprint(tree)
- return tree
- #misc.set_filename(filename, tree)
- return compile(tree, mode="exec", filename="<string>")
- #syntax.check(tree)
- gen = pycodegen.ModuleCodeGenerator(tree)
- code = gen.getCode()
- return code