import re
import sys
import json
import os.path
import hashlib
import logging
import datetime
import dateutil.parser
import collections
from werkzeug.utils import cached_property
from piecrust.configuration import (
        Configuration, ConfigurationError,
        parse_config_header)


logger = logging.getLogger(__name__)


class PageConfiguration(Configuration):
    def __init__(self, values=None, validate=True):
        super(PageConfiguration, self).__init__(values, validate)

    def _validateAll(self, values):
        values.setdefault('title', 'Untitled Page')
        values.setdefault('content_type', 'html')
        ppp = values.get('posts_per_page')
        if ppp is not None:
            values.setdefault('items_per_page', ppp)
        pf = values.get('posts_filters')
        if pf is not None:
            values.setdefault('items_filters', pf)
        return values


FLAG_NONE = 0
FLAG_RAW_CACHE_VALID = 2**0


class Page(object):
    def __init__(self, source, source_metadata, rel_path):
        self.source = source
        self.source_metadata = source_metadata
        self.rel_path = rel_path
        self._config = None
        self._segments = None
        self._flags = FLAG_NONE
        self._datetime = None

    @property
    def app(self):
        return self.source.app

    @property
    def ref_spec(self):
        return '%s:%s' % (self.source.name, self.rel_path)

    @cached_property
    def path(self):
        path, _ = self.source.resolveRef(self.rel_path)
        return path

    @cached_property
    def path_mtime(self):
        return os.path.getmtime(self.path)

    @property
    def flags(self):
        return self._flags

    @property
    def config(self):
        self._load()
        return self._config

    @property
    def segments(self):
        self._load()
        return self._segments

    @property
    def datetime(self):
        if self._datetime is None:
            try:
                if 'datetime' in self.source_metadata:
                    # Get the date/time from the source.
                    self._datetime = self.source_metadata['datetime']
                elif 'date' in self.source_metadata:
                    # Get the date from the source. Potentially get the
                    # time from the page config.
                    page_date = self.source_metadata['date']
                    page_time = _parse_config_time(self.config.get('time'))
                    if page_time is not None:
                        self._datetime = datetime.datetime(
                                page_date.year,
                                page_date.month,
                                page_date.day) + page_time
                    else:
                        self._datetime = datetime.datetime(
                                page_date.year, page_date.month, page_date.day)
                elif 'date' in self.config:
                    # Get the date from the page config, and maybe the
                    # time too.
                    page_date = _parse_config_date(self.config.get('date'))
                    self._datetime = datetime.datetime(
                            page_date.year,
                            page_date.month,
                            page_date.day)
                    page_time = _parse_config_time(self.config.get('time'))
                    if page_time is not None:
                        self._datetime += page_time
                else:
                    # No idea what the date/time for this page is.
                    self._datetime = datetime.datetime.fromtimestamp(0)
            except Exception as ex:
                logger.exception(ex)
                raise Exception(
                        "Error computing time for page: %s" %
                        self.path) from ex
        return self._datetime

    @datetime.setter
    def datetime(self, value):
        self._datetime = value

    def getSegment(self, name='content'):
        return self.segments[name]

    def _load(self):
        if self._config is not None:
            return

        config, content, was_cache_valid = load_page(self.app, self.path,
                                                     self.path_mtime)
        if 'config' in self.source_metadata:
            config.merge(self.source_metadata['config'])

        self._config = config
        self._segments = content
        if was_cache_valid:
            self._flags |= FLAG_RAW_CACHE_VALID


def _parse_config_date(page_date):
    if page_date is None:
        return None

    if isinstance(page_date, str):
        try:
            parsed_d = dateutil.parser.parse(page_date)
        except Exception as ex:
            logger.exception(ex)
            raise ConfigurationError("Invalid date: %s" % page_date) from ex
        return datetime.date(
                year=parsed_d.year,
                month=parsed_d.month,
                day=parsed_d.day)

    raise ConfigurationError("Invalid date: %s" % page_date)


def _parse_config_time(page_time):
    if page_time is None:
        return None

    if isinstance(page_time, datetime.timedelta):
        return page_time

    if isinstance(page_time, str):
        try:
            parsed_t = dateutil.parser.parse(page_time)
        except Exception as ex:
            logger.exception(ex)
            raise ConfigurationError("Invalid time: %s" % page_time) from ex
        return datetime.timedelta(
                hours=parsed_t.hour,
                minutes=parsed_t.minute,
                seconds=parsed_t.second)

    if isinstance(page_time, int):
        # Total seconds... convert to a time struct.
        return datetime.timedelta(seconds=page_time)

    raise ConfigurationError("Invalid time: %s" % page_time)


class PageLoadingError(Exception):
    def __init__(self, path, inner=None):
        super(PageLoadingError, self).__init__(
                "Error loading page: %s" % path,
                inner)


class ContentSegment(object):
    debug_render_func = 'debug_render'

    def __init__(self):
        self.parts = []

    def debug_render(self):
        return '\n'.join([p.content for p in self.parts])


class ContentSegmentPart(object):
    def __init__(self, content, fmt=None, offset=-1, line=-1):
        self.content = content
        self.fmt = fmt
        self.offset = offset
        self.line = line

    def __str__(self):
        return '%s [%s]' % (self.content, self.fmt or '<default>')


def json_load_segments(data):
    segments = {}
    for key, seg_data in data.items():
        seg = ContentSegment()
        for p_data in seg_data:
            part = ContentSegmentPart(p_data['c'], p_data['f'], p_data['o'],
                                      p_data['l'])
            seg.parts.append(part)
        segments[key] = seg
    return segments


def json_save_segments(segments):
    data = {}
    for key, seg in segments.items():
        seg_data = []
        for part in seg.parts:
            p_data = {'c': part.content, 'f': part.fmt, 'o': part.offset,
                      'l': part.line}
            seg_data.append(p_data)
        data[key] = seg_data
    return data


def load_page(app, path, path_mtime=None):
    try:
        with app.env.timerScope('PageLoad'):
            return _do_load_page(app, path, path_mtime)
    except Exception as e:
        logger.exception(
                "Error loading page: %s" %
                os.path.relpath(path, app.root_dir))
        _, __, traceback = sys.exc_info()
        raise PageLoadingError(path, e).with_traceback(traceback)


def _do_load_page(app, path, path_mtime):
    # Check the cache first.
    cache = app.cache.getCache('pages')
    cache_path = hashlib.md5(path.encode('utf8')).hexdigest() + '.json'
    page_time = path_mtime or os.path.getmtime(path)
    if cache.isValid(cache_path, page_time):
        cache_data = json.loads(
                cache.read(cache_path),
                object_pairs_hook=collections.OrderedDict)
        config = PageConfiguration(
                values=cache_data['config'],
                validate=False)
        content = json_load_segments(cache_data['content'])
        return config, content, True

    # Nope, load the page from the source file.
    logger.debug("Loading page configuration from: %s" % path)
    with open(path, 'r', encoding='utf-8') as fp:
        raw = fp.read()
    header, offset = parse_config_header(raw)

    if 'format' not in header:
        auto_formats = app.config.get('site/auto_formats')
        name, ext = os.path.splitext(path)
        header['format'] = auto_formats.get(ext, None)

    config = PageConfiguration(header)
    content = parse_segments(raw, offset)
    config.set('segments', list(content.keys()))

    # Save to the cache.
    cache_data = {
            'config': config.getAll(),
            'content': json_save_segments(content)}
    cache.write(cache_path, json.dumps(cache_data))

    return config, content, False


segment_pattern = re.compile(
        r"""^\-\-\-\s*(?P<name>\w+)(\:(?P<fmt>\w+))?\s*\-\-\-\s*$""",
        re.M)
part_pattern = re.compile(
        r"""^<\-\-\s*(?P<fmt>\w+)\s*\-\->\s*$""",
        re.M)


def _count_lines(s):
    return len(s.split('\n'))


def _string_needs_parsing(txt, offset):
    txtlen = len(txt)
    index = txt.find('-', offset)
    while index >= 0 and index < txtlen - 8:
        # Look for a potential `<--format-->`
        if index > 0 and txt[index - 1] == '<' and txt[index + 1] == '-':
            return True
        # Look for a potential `---segment---`
        if txt[index + 1] == '-' and txt[index + 2] == '-':
            return True
        index = txt.find('-', index + 1)
    return False


def parse_segments(raw, offset=0):
    # Get the number of lines in the header.
    header_lines = _count_lines(raw[:offset].rstrip())
    current_line = header_lines

    # Figure out if we need any parsing.
    do_parse = _string_needs_parsing(raw, offset)
    if not do_parse:
        seg = ContentSegment()
        seg.parts = [
                ContentSegmentPart(raw[offset:], None, offset, current_line)]
        return {'content': seg}

    # Start parsing segments and parts.
    matches = list(segment_pattern.finditer(raw, offset))
    num_matches = len(matches)
    if num_matches > 0:
        contents = {}

        first_offset = matches[0].start()
        if first_offset > 0:
            # There's some default content segment at the beginning.
            seg = ContentSegment()
            seg.parts, current_line = parse_segment_parts(
                    raw, offset, first_offset, current_line)
            contents['content'] = seg

        for i in range(1, num_matches):
            m1 = matches[i - 1]
            m2 = matches[i]
            seg = ContentSegment()
            seg.parts, current_line = parse_segment_parts(
                    raw, m1.end() + 1, m2.start(), current_line,
                    m1.group('fmt'))
            contents[m1.group('name')] = seg

        # Handle text past the last match.
        lastm = matches[-1]
        seg = ContentSegment()
        seg.parts, current_line = parse_segment_parts(
                raw, lastm.end() + 1, len(raw), current_line,
                lastm.group('fmt'))
        contents[lastm.group('name')] = seg

        return contents
    else:
        # No segments, just content.
        seg = ContentSegment()
        seg.parts, current_line = parse_segment_parts(
                raw, offset, len(raw), current_line)
        return {'content': seg}


def parse_segment_parts(raw, start, end, line_offset, first_part_fmt=None):
    matches = list(part_pattern.finditer(raw, start, end))
    num_matches = len(matches)
    if num_matches > 0:
        parts = []

        # First part, before the first format change.
        part_text = raw[start:matches[0].start()]
        parts.append(
                ContentSegmentPart(part_text, first_part_fmt, start,
                                   line_offset))
        line_offset += _count_lines(part_text)

        for i in range(1, num_matches):
            m1 = matches[i - 1]
            m2 = matches[i]
            part_text = raw[m1.end() + 1:m2.start()]
            parts.append(
                    ContentSegmentPart(
                        part_text, m1.group('fmt'), m1.end() + 1,
                        line_offset))
            line_offset += _count_lines(part_text)

        lastm = matches[-1]
        part_text = raw[lastm.end() + 1:end]
        parts.append(ContentSegmentPart(
                part_text, lastm.group('fmt'), lastm.end() + 1,
                line_offset))

        return parts, line_offset
    else:
        part_text = raw[start:end]
        parts = [ContentSegmentPart(part_text, first_part_fmt, start,
                                    line_offset)]
        return parts, line_offset