# Yapps 2 Runtime, part of Yapps 2 - yet another python parser system
# Copyright 1999-2003 by Amit J. Patel <amitp@cs.stanford.edu>
# Enhancements copyright 2003-2004 by Matthias Urlichs <smurf@debian.org>
#
# This version of the Yapps 2 Runtime can be distributed under the
# terms of the MIT open source license, either found in the LICENSE file
# included with the Yapps distribution
# <http://theory.stanford.edu/~amitp/yapps/> or at
# <http://www.opensource.org/licenses/mit-license.php>
#

"""Run time libraries needed to run parsers generated by Yapps.

This module defines parse-time exception classes, a scanner class, a
base class for parsers produced by Yapps, and a context class that
keeps track of the parse stack.

"""

import re
import sys

DEBUG = False
# File lookup window


class SyntaxError(Exception):
    """When we run into an unexpected token, this is the exception to use"""
    def __init__(self, pos=None, msg="Bad Token", context=None):
        Exception.__init__(self)
        self.pos = pos
        self.msg = msg
        self.context = context

    def __str__(self):
        if not self.pos:
            return 'SyntaxError'
        else:
            return 'SyntaxError@%s(%s)' % (repr(self.pos), self.msg)


class NoMoreTokens(Exception):
    """Another exception object, for when we run out of tokens"""
    pass


class Token(object):
    """Yapps token.

    This is a container for a scanned token.
    """

    def __init__(self, type, value, pos=None):
        """Initialize a token."""
        self.type = type
        self.value = value
        self.pos = pos

    def __repr__(self):
        output = '<%s: %s' % (self.type, repr(self.value))
        if self.pos:
            output += " @ "
            if self.pos[0]:
                output += "%s:" % self.pos[0]
            if self.pos[1]:
                output += "%d" % self.pos[1]
            if self.pos[2] is not None:
                output += ".%d" % self.pos[2]
        output += ">"
        return output


class Scanner(object):
    """Yapps scanner.

    The Yapps scanner can work in context sensitive or context
    insensitive modes.  The token(i) method is used to retrieve the
    i-th token.  It takes a restrict set that limits the set of tokens
    it is allowed to return.  In context sensitive mode, this restrict
    set guides the scanner.  In context insensitive mode, there is no
    restriction (the set is always the full set of tokens).

    """
    MIN_WINDOW = 4096
    in_name = 0

    def __init__(self, patterns, ignore, input="",
            file=None, filename=None):
        """Initialize the scanner.

        Parameters:
          patterns : [(terminal, uncompiled regex), ...] or None
          ignore : {terminal:None, ...}
          input : string

        If patterns is None, we assume that the subclass has
        defined self.patterns : [(terminal, compiled regex), ...].
        Note that the patterns parameter expects uncompiled regexes,
        whereas the self.patterns field expects compiled regexes.

        The 'ignore' value is either None or a callable, which is called
        with the scanner and the to-be-ignored match object; this can
        be used for include file or comment handling.
        """

        if not filename:
            filename = "<f.%d>" % self.__class__.in_name
            self.__class__.in_name += 1

        self.reset(input, file, filename)
        self.ignore = ignore

        if patterns is not None:
            # Compile the regex strings into regex objects
            self.patterns = []
            for terminal, regex in patterns:
                self.patterns.append((terminal, re.compile(regex)))

    def reset(self, input="", file=None, filename=None):
        self.restrictions = []
        self.input = input
        self.file = file
        self.filename = filename
        self.pos = 0
        self.del_pos = 0  # skipped
        self.line = 1
        self.del_line = 0  # skipped
        self.col = 0
        self.tokens = []

        self.last_read_token = None
        self.last_token = None
        self.last_types = None

    def __repr__(self):
        """
        Print the last 10 tokens that have been scanned in
        """
        output = ''
        for t in self.tokens[-10:]:
            output = "%s\n  (@%s)  %s  =  %s" % (output, t[0], t[2], repr(t[3]))
        return output

    def get_pos(self):
        """Return a file/line/char tuple."""
        return (self.filename, self.line + self.del_line, self.col)

    def print_line_with_pointer(self, pos, length=0, out=sys.stderr):
        """Print the line of 'text' that includes position 'p',
        along with a second line with a single caret (^) at position p"""

        file, line, p = pos
        if file != self.filename:
            print >>out, "(%s: not in input buffer)" % file
            return

        text = self.input
        p += length - 1  # starts at pos 1

        origline = line
        line -= self.del_line
        spos = 0
        if line > 0:
            while 1:
                line = line - 1
                try:
                    cr = text.index("\n", spos)
                except ValueError:
                    if line:
                        text = ""
                    break
                if line == 0:
                    text = text[spos:cr]
                    break
                spos = cr+1
        else:
            print >>out, "(%s:%d not in input buffer)" % (file, origline)
            return

        # Now try printing part of the line
        text = text[max(p - 80, 0):p + 80]
        p = p - max(p - 80, 0)

        # Strip to the left
        i = text[:p].rfind('\n')
        j = text[:p].rfind('\r')
        if i < 0 or (0 <= j < i):
            i = j
        if 0 <= i < p:
            p = p - i - 1
            text = text[i+1:]

        # Strip to the right
        i = text.find('\n', p)
        j = text.find('\r', p)
        if i < 0 or (0 <= j < i):
            i = j
        if i >= 0:
            text = text[:i]

        # Now shorten the text
        while len(text) > 70 and p > 60:
            # Cut off 10 chars
            text = "..." + text[10:]
            p = p - 7

        # Now print the string, along with an indicator
        print >>out, '> ', text
        print >>out, '> ', ' ' * p + '^'

    def grab_input(self):
        """Get more input if possible."""
        if not self.file:
            return
        if len(self.input) - self.pos >= self.MIN_WINDOW:
            return

        data = self.file.read(self.MIN_WINDOW)
        if data is None or data == "":
            self.file = None

        # Drop bytes from the start, if necessary.
        if self.pos > 2 * self.MIN_WINDOW:
            self.del_pos += self.MIN_WINDOW
            self.del_line += self.input[:self.MIN_WINDOW].count("\n")
            self.pos -= self.MIN_WINDOW
            self.input = self.input[self.MIN_WINDOW:] + data
        else:
            self.input = self.input + data

    def getchar(self):
        """Return the next character."""
        self.grab_input()

        c = self.input[self.pos]
        self.pos += 1
        return c

    def _scan(self, restrict, context=None):
        """
        Should scan another token and add it to the list, self.tokens,
        and add the restriction to self.restrictions
        """
        token = None
        # Keep looking for a token, ignoring any in self.ignore
        while True:
            tok = None

            self.grab_input()

            # Search the patterns for the longest match, with earlier
            # tokens in the list having preference
            best_pat_len = -1
            best_pat = None
            for tok, regex in self.patterns:
                if DEBUG:
                    print("\tTrying %s: %s at pos %d -> %s" % (repr(tok), repr(regex.pattern), self.pos, repr(self.input)))
                # First check to see if we're ignoring this token
                if restrict and tok not in restrict and tok not in self.ignore:
                    if DEBUG:
                        print "\tSkipping %s!" % repr(tok)
                    continue
                m = regex.match(self.input, self.pos)
                if m and m.end() - m.start() > best_pat_len:
                    # We got a match that's better than the previous one
                    best_pat = tok
                    best_pat_len = m.end() - m.start()
                    if DEBUG:
                        print("Match OK! %s: %s at pos %d" % (repr(tok), repr(regex.pattern), self.pos))
                    break

            # If we didn't find anything, raise an error
            if best_pat is None or best_pat_len < 0:
                msg = "Bad token: %s" % ("???" if tok is None else repr(tok),)
                if restrict:
                    msg = "%s found while trying to find one of the restricted tokens: %s" % ("???" if tok is None else repr(tok), ", ".join(repr(r) for r in restrict))
                raise SyntaxError(self.get_pos(), msg, context=context)

            ignore = best_pat in self.ignore  # Should this token be ignored?
            start_pos = self.pos
            end_pos = start_pos + best_pat_len
            self.pos = end_pos

            # If we found something that isn't to be ignored, return it
            if not ignore:
                value = self.input[start_pos:end_pos]
                # token = Token(type=best_pat, value=value, pos=self.get_pos())
                token = (
                    start_pos,
                    end_pos,
                    best_pat,
                    value,
                )
                # print repr(token)
                if not self.tokens or token != self.last_read_token:
                    # Only add this token if it's not in the list (to prevent looping)
                    self.last_read_token = token
                    self.tokens.append(token)
                    self.restrictions.append(restrict)
                    return 1
                return 0

    def token(self, i, restrict=None, **kwargs):
        """
        Get the i'th token, and if i is one past the end, then scan
        for another token; restrict is a list of tokens that
        are allowed, or 0 for any token.
        """
        context = kwargs.get("context")
        tokens_len = len(self.tokens)
        if i == tokens_len:  # We are at the end, get the next...
            tokens_len += self._scan(restrict, context)
        elif i >= 0 and i < tokens_len:
            if restrict and self.restrictions[i] and restrict > self.restrictions[i]:
                raise NotImplementedError("Unimplemented: restriction set changed")
        if i >= 0 and i < tokens_len:
            return self.tokens[i]
        raise NoMoreTokens()

    def rewind(self, i):
        tokens_len = len(self.tokens)
        if i <= tokens_len:
            token = self.tokens[i]
            self.tokens = self.tokens[:i]
            self.restrictions = self.restrictions[:i]
            self.pos = token[0]


class Parser(object):
    """Base class for Yapps-generated parsers.

    """

    def __init__(self, scanner):
        self._scanner = scanner
        self._pos = 0

    def reset(self, input):
        self._scanner.reset(input)
        self._pos = 0

    def _peek(self, types, **kwargs):
        """Returns the token type for lookahead; if there are any args
        then the list of args is the set of token types to allow"""
        try:
            tok = self._scanner.token(self._pos, types)
            return tok[2]
        except SyntaxError:
            return None

    def _scan(self, type, **kwargs):
        """Returns the matched text, and moves to the next token"""
        tok = self._scanner.token(self._pos, set([type]))
        if tok[2] != type:
            raise SyntaxError("SyntaxError[@ char %s: %s]" % (repr(tok[0]), "Trying to find " + type))
        self._pos += 1
        return tok[3]

    def _rewind(self, n=1):
        self._pos -= min(n, self._pos)
        self._scanner.rewind(self._pos)


class Context(object):
    """Class to represent the parser's call stack.

    Every rule creates a Context that links to its parent rule.  The
    contexts can be used for debugging.

    """

    def __init__(self, parent, scanner, rule, args=()):
        """Create a new context.

        Args:
        parent: Context object or None
        scanner: Scanner object
        rule: string (name of the rule)
        args: tuple listing parameters to the rule

        """
        self.parent = parent
        self.scanner = scanner
        self.rule = rule
        self.args = args
        self.token = scanner.last_read_token

    def __str__(self):
        output = ''
        if self.parent:
            output = str(self.parent) + ' > '
        output += self.rule
        return output


def print_error(err, scanner, max_ctx=None):
    """Print error messages, the parser stack, and the input text -- for human-readable error messages."""
    # NOTE: this function assumes 80 columns :-(
    # Figure out the line number
    pos = err.pos
    if not pos:
        pos = scanner.get_pos()

    file_name, line_number, column_number = pos
    print >>sys.stderr, '%s:%d:%d: %s' % (file_name, line_number, column_number, err.msg)

    scanner.print_line_with_pointer(pos)

    context = err.context
    token = None
    while context:
        print >>sys.stderr, 'while parsing %s%s:' % (context.rule, tuple(context.args))
        if context.token:
            token = context.token
        if token:
            scanner.print_line_with_pointer(token.pos, length=len(token.value))
        context = context.parent
        if max_ctx:
            max_ctx = max_ctx-1
            if not max_ctx:
                break


def wrap_error_reporter(parser, rule, *args, **kwargs):
    try:
        return getattr(parser, rule)(*args, **kwargs)
    except SyntaxError, e:
        print_error(e, parser._scanner)
    except NoMoreTokens:
        print >>sys.stderr, 'Could not complete parsing; stopped around here:'
        print >>sys.stderr, parser._scanner