| 123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258 | from __future__ import unicode_literalsimport jsonimport operatorimport refrom .utils import (    ExtractorError,)_OPERATORS = [    ('|', operator.or_),    ('^', operator.xor),    ('&', operator.and_),    ('>>', operator.rshift),    ('<<', operator.lshift),    ('-', operator.sub),    ('+', operator.add),    ('%', operator.mod),    ('/', operator.truediv),    ('*', operator.mul),]_ASSIGN_OPERATORS = [(op + '=', opfunc) for op, opfunc in _OPERATORS]_ASSIGN_OPERATORS.append(('=', lambda cur, right: right))_NAME_RE = r'[a-zA-Z_$][a-zA-Z_$0-9]*'class JSInterpreter(object):    def __init__(self, code, objects=None):        if objects is None:            objects = {}        self.code = code        self._functions = {}        self._objects = objects    def interpret_statement(self, stmt, local_vars, allow_recursion=100):        if allow_recursion < 0:            raise ExtractorError('Recursion limit reached')        should_abort = False        stmt = stmt.lstrip()        stmt_m = re.match(r'var\s', stmt)        if stmt_m:            expr = stmt[len(stmt_m.group(0)):]        else:            return_m = re.match(r'return(?:\s+|$)', stmt)            if return_m:                expr = stmt[len(return_m.group(0)):]                should_abort = True            else:                # Try interpreting it as an expression                expr = stmt        v = self.interpret_expression(expr, local_vars, allow_recursion)        return v, should_abort    def interpret_expression(self, expr, local_vars, allow_recursion):        expr = expr.strip()        if expr == '':  # Empty expression            return None        if expr.startswith('('):            parens_count = 0            for m in re.finditer(r'[()]', expr):                if m.group(0) == '(':                    parens_count += 1                else:                    parens_count -= 1                    if parens_count == 0:                        sub_expr = expr[1:m.start()]                        sub_result = self.interpret_expression(                            sub_expr, local_vars, allow_recursion)                        remaining_expr = expr[m.end():].strip()                        if not remaining_expr:                            return sub_result                        else:                            expr = json.dumps(sub_result) + remaining_expr                        break            else:                raise ExtractorError('Premature end of parens in %r' % expr)        for op, opfunc in _ASSIGN_OPERATORS:            m = re.match(r'''(?x)                (?P<out>%s)(?:\[(?P<index>[^\]]+?)\])?                \s*%s                (?P<expr>.*)$''' % (_NAME_RE, re.escape(op)), expr)            if not m:                continue            right_val = self.interpret_expression(                m.group('expr'), local_vars, allow_recursion - 1)            if m.groupdict().get('index'):                lvar = local_vars[m.group('out')]                idx = self.interpret_expression(                    m.group('index'), local_vars, allow_recursion)                assert isinstance(idx, int)                cur = lvar[idx]                val = opfunc(cur, right_val)                lvar[idx] = val                return val            else:                cur = local_vars.get(m.group('out'))                val = opfunc(cur, right_val)                local_vars[m.group('out')] = val                return val        if expr.isdigit():            return int(expr)        var_m = re.match(            r'(?!if|return|true|false)(?P<name>%s)$' % _NAME_RE,            expr)        if var_m:            return local_vars[var_m.group('name')]        try:            return json.loads(expr)        except ValueError:            pass        m = re.match(            r'(?P<var>%s)\.(?P<member>[^(]+)(?:\(+(?P<args>[^()]*)\))?$' % _NAME_RE,            expr)        if m:            variable = m.group('var')            member = m.group('member')            arg_str = m.group('args')            if variable in local_vars:                obj = local_vars[variable]            else:                if variable not in self._objects:                    self._objects[variable] = self.extract_object(variable)                obj = self._objects[variable]            if arg_str is None:                # Member access                if member == 'length':                    return len(obj)                return obj[member]            assert expr.endswith(')')            # Function call            if arg_str == '':                argvals = tuple()            else:                argvals = tuple([                    self.interpret_expression(v, local_vars, allow_recursion)                    for v in arg_str.split(',')])            if member == 'split':                assert argvals == ('',)                return list(obj)            if member == 'join':                assert len(argvals) == 1                return argvals[0].join(obj)            if member == 'reverse':                assert len(argvals) == 0                obj.reverse()                return obj            if member == 'slice':                assert len(argvals) == 1                return obj[argvals[0]:]            if member == 'splice':                assert isinstance(obj, list)                index, howMany = argvals                res = []                for i in range(index, min(index + howMany, len(obj))):                    res.append(obj.pop(index))                return res            return obj[member](argvals)        m = re.match(            r'(?P<in>%s)\[(?P<idx>.+)\]$' % _NAME_RE, expr)        if m:            val = local_vars[m.group('in')]            idx = self.interpret_expression(                m.group('idx'), local_vars, allow_recursion - 1)            return val[idx]        for op, opfunc in _OPERATORS:            m = re.match(r'(?P<x>.+?)%s(?P<y>.+)' % re.escape(op), expr)            if not m:                continue            x, abort = self.interpret_statement(                m.group('x'), local_vars, allow_recursion - 1)            if abort:                raise ExtractorError(                    'Premature left-side return of %s in %r' % (op, expr))            y, abort = self.interpret_statement(                m.group('y'), local_vars, allow_recursion - 1)            if abort:                raise ExtractorError(                    'Premature right-side return of %s in %r' % (op, expr))            return opfunc(x, y)        m = re.match(            r'^(?P<func>%s)\((?P<args>[a-zA-Z0-9_$,]+)\)$' % _NAME_RE, expr)        if m:            fname = m.group('func')            argvals = tuple([                int(v) if v.isdigit() else local_vars[v]                for v in m.group('args').split(',')])            if fname not in self._functions:                self._functions[fname] = self.extract_function(fname)            return self._functions[fname](argvals)        raise ExtractorError('Unsupported JS expression %r' % expr)    def extract_object(self, objname):        obj = {}        obj_m = re.search(            (r'(?:var\s+)?%s\s*=\s*\{' % re.escape(objname)) +            r'\s*(?P<fields>([a-zA-Z$0-9]+\s*:\s*function\(.*?\)\s*\{.*?\}(?:,\s*)?)*)' +            r'\}\s*;',            self.code)        fields = obj_m.group('fields')        # Currently, it only supports function definitions        fields_m = re.finditer(            r'(?P<key>[a-zA-Z$0-9]+)\s*:\s*function'            r'\((?P<args>[a-z,]+)\){(?P<code>[^}]+)}',            fields)        for f in fields_m:            argnames = f.group('args').split(',')            obj[f.group('key')] = self.build_function(argnames, f.group('code'))        return obj    def extract_function(self, funcname):        func_m = re.search(            r'''(?x)                (?:function\s+%s|[{;,]\s*%s\s*=\s*function|var\s+%s\s*=\s*function)\s*                \((?P<args>[^)]*)\)\s*                \{(?P<code>[^}]+)\}''' % (                re.escape(funcname), re.escape(funcname), re.escape(funcname)),            self.code)        if func_m is None:            raise ExtractorError('Could not find JS function %r' % funcname)        argnames = func_m.group('args').split(',')        return self.build_function(argnames, func_m.group('code'))    def call_function(self, funcname, *args):        f = self.extract_function(funcname)        return f(args)    def build_function(self, argnames, code):        def resf(args):            local_vars = dict(zip(argnames, args))            for stmt in code.split(';'):                res, abort = self.interpret_statement(stmt, local_vars)                if abort:                    break            return res        return resf
 |