#!/usr/bin/python
# coding: utf-8

"""Lexical analysis of formal languages (i.e. code) using Pygments."""

# :Author: Georg Brandl; Felix Wiemann; Günter Milde
# :Date: $Date: 2015-04-20 16:05:27 +0200 (Mo, 20 Apr 2015) $
# :Copyright: This module has been placed in the public domain.

from docutils import ApplicationError
try:
    import pygments
    from pygments.lexers import get_lexer_by_name
    from pygments.formatters.html import _get_ttype_class
    with_pygments = True
except (ImportError, SyntaxError): # pygments 2.0.1 fails with Py 3.1 and 3.2
    with_pygments = False

# Filter the following token types from the list of class arguments:
unstyled_tokens = ['token', # Token (base token type)
                   'text',  # Token.Text
                   '']      # short name for Token and Text
# (Add, e.g., Token.Punctuation with ``unstyled_tokens += 'punctuation'``.)

class LexerError(ApplicationError): 
    pass

class Lexer(object):
    """Parse `code` lines and yield "classified" tokens.

    Arguments

      code       -- string of source code to parse,
      language   -- formal language the code is written in,
      tokennames -- either 'long', 'short', or '' (see below).

    Merge subsequent tokens of the same token-type.

    Iterating over an instance yields the tokens as ``(tokentype, value)``
    tuples. The value of `tokennames` configures the naming of the tokentype:

      'long':  downcased full token type name,
      'short': short name defined by pygments.token.STANDARD_TYPES
               (= class argument used in pygments html output),
      'none':      skip lexical analysis.
    """

    def __init__(self, code, language, tokennames='short'):
        """
        Set up a lexical analyzer for `code` in `language`.
        """
        self.code = code
        self.language = language
        self.tokennames = tokennames
        self.lexer = None
        # get lexical analyzer for `language`:
        if language in ('', 'text') or tokennames == 'none':
            return
        if not with_pygments:
            raise LexerError('Cannot analyze code. '
                                    'Pygments package not found.')
        try:
            self.lexer = get_lexer_by_name(self.language)
        except pygments.util.ClassNotFound:
            raise LexerError('Cannot analyze code. '
                'No Pygments lexer found for "%s".' % language)

    # Since version 1.2. (released Jan 01, 2010) Pygments has a
    # TokenMergeFilter. However, this requires Python >= 2.4. When Docutils
    # requires same minimal version,  ``self.merge(tokens)`` in __iter__ can
    # be replaced by ``self.lexer.add_filter('tokenmerge')`` in __init__.
    def merge(self, tokens):
        """Merge subsequent tokens of same token-type.

           Also strip the final newline (added by pygments).
        """
        tokens = iter(tokens)
        (lasttype, lastval) = next(tokens)
        for ttype, value in tokens:
            if ttype is lasttype:
                lastval += value
            else:
                yield(lasttype, lastval)
                (lasttype, lastval) = (ttype, value)
        if lastval.endswith('\n'):
            lastval = lastval[:-1]
        if lastval:
            yield(lasttype, lastval)

    def __iter__(self):
        """Parse self.code and yield "classified" tokens.
        """
        if self.lexer is None:
            yield ([], self.code)
            return
        tokens = pygments.lex(self.code, self.lexer)
        for tokentype, value in self.merge(tokens):
            if self.tokennames == 'long': # long CSS class args
                classes = str(tokentype).lower().split('.')
            else: # short CSS class args
                classes = [_get_ttype_class(tokentype)]
            classes = [cls for cls in classes if cls not in unstyled_tokens]
            yield (classes, value)


class NumberLines(object):
    """Insert linenumber-tokens at the start of every code line.

    Arguments

       tokens    -- iterable of ``(classes, value)`` tuples
       startline -- first line number
       endline   -- last line number

    Iterating over an instance yields the tokens with a
    ``(['ln'], '<the line number>')`` token added for every code line.
    Multi-line tokens are splitted."""

    def __init__(self, tokens, startline, endline):
        self.tokens = tokens
        self.startline = startline
        # pad linenumbers, e.g. endline == 100 -> fmt_str = '%3d '
        self.fmt_str = '%%%dd ' % len(str(endline))

    def __iter__(self):
        lineno = self.startline
        yield (['ln'], self.fmt_str % lineno)
        for ttype, value in self.tokens:
            lines = value.split('\n')
            for line in lines[:-1]:
                yield (ttype, line + '\n')
                lineno += 1
                yield (['ln'], self.fmt_str % lineno)
            yield (ttype, lines[-1])