"""
Functions that take strings and yield streams or dicts of `pugsql.lexer.Token`
objects, keeping track of source location.
"""
from . import context
from collections import namedtuple
import re


__pdoc__ = {}


Token = namedtuple('Token', ['tag', 'value', 'context'])
__pdoc__['Token'] = 'A tagged string produced by the lexer.'
__pdoc__['Token.tag'] = 'A character indicating the meaning of the `value`.'
__pdoc__['Token.value'] = 'The string value of the `Token`.'
__pdoc__['Token.context'] = ('A `pugsql.context.Context` for tracking source '
                             'code information.')


def lex(pugsql, ctx):
    """
    Splits the provided multiline PugSQL string into
    """
    def generate(pugsql, ctx):
        for l in pugsql.splitlines():
            ctx = context.advance(ctx, lines=1)
            yield _categorize(l, ctx)
    return list(generate(pugsql, ctx))


def _categorize(line, ctx):
    line, ctx = _whitespace_advance(line, ctx)

    if line.startswith('--'):
        return Token('C', line, ctx)
    return Token('Q', line, ctx)


def lex_comment(token):
    m = re.match(
        r'(?P<lead>--\s*)'
        r'(?P<keyword>\:[^ ]+)'
        r'(?P<internalws>\s+)?'
        r'(?P<rest>.*)?', token.value)

    if not m:
        return None

    d = m.groupdict()
    restbegin = sum(len(d[k] or '') for k in d.keys() if k != 'rest')

    return {
        'keyword': Token(
            'K',
            d['keyword'],
            context.advance(token.context, cols=len(d['lead']))),

        'rest': Token(
            'S',
            d['rest'],
            context.advance(token.context, cols=restbegin))
    }


def lex_name(token):
    line, ctx = _whitespace_advance(token.value, token.context)

    m = re.match(
        r'(?P<name>[^ ]+)'
        r'(?P<internalws>\s+)?'
        r'(?P<keyword>\:[^ ]+)?'
        r'(?P<internalws2>\s+)?'
        r'(?P<rest>.+)?', line)

    if not m:
        return None

    d = m.groupdict()

    kwbegin = len(d['name']) + len(d['internalws'] or '')
    kwctx = context.advance(ctx, cols=kwbegin)
    restbegin = sum(len(d[k] or '') for k in d.keys() if k != 'rest')
    restctx = context.advance(ctx, cols=restbegin)
    return {
        'name': Token('N', d['name'], ctx),
        'keyword': Token('K', d['keyword'], kwctx),
        'rest': Token('S', d['rest'], restctx),
    }


def lex_result(token):
    line, ctx = _whitespace_advance(token.value, token.context)
    m = re.match(
        r'(?P<keyword>\:[^ ]+)'
        r'(?P<rest>.+)?', line)

    if not m:
        return None

    d = m.groupdict()
    restctx = context.advance(ctx, cols=len(d['keyword']))

    return {
        'keyword': Token('K', d['keyword'], ctx),
        'rest': Token('S', d['rest'], restctx),
    }


def _whitespace_advance(line, ctx):
    ctx = context.advance(ctx, cols=len(line) - len(line.lstrip()))
    return line.strip(), ctx