# -*- coding: utf-8 -*-

# Copyright © 2012-2014 marmuta <marmvta@gmail.com>
#
# This file is part of Onboard.
#
# Onboard is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 3 of the License, or
# (at your option) any later version.
#
# Onboard is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.

from __future__ import division, print_function, unicode_literals

import os
import subprocess
import time
import re
import glob

from Onboard.utils import unicode_str

import Onboard.osk as osk

import logging
_logger = logging.getLogger(__name__)


class SpellChecker(object):
    MAX_QUERY_CACHE_SIZE = 100    # max number of cached queries

    def __init__(self, language_db = None):
        self._language_db = language_db
        self._backend = None
        self._cached_queries = {}

    def set_backend(self, backend):
        """ Switch spell check backend on the fly """
        if backend is None:
            if self._backend:
                self._backend.stop()
            self._backend = None
        else:
            if backend == 0:
                _class = hunspell
            else:
                _class = aspell_cmd

            if not self._backend or \
               not type(self._backend) == _class:
                if self._backend:
                    self._backend.stop()
                self._backend = _class()

        self.invalidate_query_cache()

    def set_dict_ids(self, dict_ids):
        success = False
        ids = self._find_matching_dicts(dict_ids)
        if self._backend and \
           not ids == self._backend.get_active_dict_ids():
            self._backend.stop()
            if ids:
                self._backend.start(ids)
                success = True
            else:
                _logger.info("No matching dictionaries for '{backend}' {dicts}" \
                             .format(backend=type(self._backend),
                                     dicts=dict_ids))
        self.invalidate_query_cache()
        return success

    def _find_matching_dicts(self, dict_ids):
        results = []
        for dict_id in dict_ids:
            id = self._find_matching_dict(dict_id)
            if id:
                results.append(id)
        return results

    def _find_matching_dict(self, dict_id):
        """
        Try to match up the given dict_id with the available dictionaries.
        Look for alternatives if there is no direct match.
        """
        available_dict_ids = self.get_supported_dict_ids()

        result = ""
        # try an exact match
        if dict_id in available_dict_ids:
            result = dict_id
        else:
            # try separator "-", common for myspell dicts
            alt_id = dict_id.replace("_", "-")
            if alt_id in available_dict_ids:
                result = alt_id
            elif self._language_db:
                # try the language code alone
                lang_code, country_code = self._language_db.split_lang_id(dict_id)
                if lang_code in available_dict_ids:
                    result = lang_code
                else:
                    # try adding the languages main country
                    lang_id = self._language_db.get_main_lang_id(lang_code)
                    if lang_id and lang_id in available_dict_ids:
                        result = lang_id

        return result

    def find_corrections(self, word, caret_offset):
        """
        Return spelling suggestions for word.
        Multiple result sets may be returned, as the spell
        checkers may return more than one result for certain tokens,
        e.g. before and after hyphens.
        """
        span = None
        suggestions = []
        if self._backend:
            results = self.query_cached(word)
            # hunspell splits words at underscores and then
            # returns results for multiple sub-words.
            # -> find the sub-word at the current caret offset.
            for result in results:
                if result[0][0] > caret_offset:
                    break
                suggestions = result[1]
                span = result[0]

        return span, suggestions

    def find_incorrect_spans(self, word):
        """
        Return misspelled spans (begin- to end-index) inside word.
        Multiple result sets may be returned, as the spell
        checkers may return more than one result for certain tokens,
        e.g. before and after hyphens.
        """
        spans = []
        if self._backend:
            results = self.query_cached(word)
            # hunspell splits words at underscores and then
            # returns results for multiple sub-words.
            # -> find the sub-word at the current caret offset.
            spans = [result[0] for result in results]
        return spans

    def query_cached(self, word):
        """
        Return cached query or ask the backend if necessary.
        """
        query = self._cached_queries.get(word)
        if query is None:
            # limit cache size
            size = len(self._cached_queries)
            if size >= self.MAX_QUERY_CACHE_SIZE:
                new_size = size // 2

                _logger.debug("shrinking query cache from {} to {} entries." \
                              .format(size, new_size))

                # discard the oldest entries
                queries = sorted(self._cached_queries.items(),
                                 key = lambda x: x[1][0])
                self._cached_queries = dict(queries[new_size:])

            # query backend
            results = self.query(word)
            query = [0.0, results]
            self._cached_queries[word] = query

        query[0] = time.time()

        return query[1]

    def query(self, word):
        return self._backend.query(word)

    def invalidate_query_cache(self):
        self._cached_queries = {}

    def get_supported_dict_ids(self):
        return self._backend.get_supported_dict_ids()


class SCBackend(object):
    """ Abstract base class of all spellchecker backends """

    def __init__(self, dict_ids = None):
        self._active_dicts = None
        self._p = None

    def start(self, dict_ids = None):
        self._active_dicts = dict_ids
        _logger.info("starting '{backend}' {dicts}" \
                     .format(backend=type(self), dicts=dict_ids))

    def stop(self):
        if self.is_running():
            _logger.info("stopping '{backend}'" \
                         .format(backend=type(self)))
            self._active_dicts = None

    def is_running(self):
        return NotImplementedError()

    def query(self, text):
        """
        Query for spelling suggestions.
        Text may contain one or more words. Each word generates its own
        list of suggestions. The spell checker backend decides about
        word boundaries.
        """
        results = []

        # Check if the process is still running, it might have
        # exited on start due to an unknown dictinary name.
        if self._p and not self._p.poll() is None:
            self._p = None

        if self._p:

            # unicode?
            if type(text) == type(""):
                line = "^" + text + "\n"
                line = line.encode("UTF-8")
            else: # already UTF-8 byte array
                line = b"^" + text + b"\n"

            self._p.stdin.write(line)
            self._p.stdin.flush()
            while True:
                s = self._p.stdout.readline().decode("UTF-8")
                s = s.strip()
                if not s:
                    break
                if s[:1] == "&":
                    sections = s.split(":")
                    a = sections[0].split()
                    begin = int(a[3]) - 1 # -1 for the prefixed ^
                    end   = begin + len(a[1])
                    span = [begin, end, a[1]] # begin, end, word
                    suggestions = sections[1].strip().split(', ')
                    results.append([span, suggestions])
                if s[:1] == "#":
                    sections = s.split(":")
                    a = sections[0].split()
                    begin = int(a[2]) - 1 # -1 for the prefixed ^
                    end   = begin + len(a[1])
                    span = [begin, end, a[1]] # begin, end, word
                    suggestions = []
                    results.append([span, suggestions])

        return results

    def get_supported_dict_ids(self):
        """
        Return raw supported dictionary ids.
        """
        raise NotImplementedError()

    def get_active_dict_ids(self):
        """
        Return active dictionary ids.
        """
        return self._active_dicts


class hunspell(SCBackend):
    """
    Hunspell backend using the C API.

    Doctests:
    # known word
    >>> sp = hunspell(["en_US"])
    >>> sp.query("test")
    []

    # unknown word
    >>> sp = hunspell(["en_US"])
    >>> sp.query("jdaskljasd")  # doctest: +ELLIPSIS
    [[...
    """
    def __init__(self, dict_ids = None):
        self._osk_hunspell = None
        SCBackend.__init__(self, dict_ids)
        if dict_ids:
            self.start(dict_ids)

    def start(self, dict_ids = None):
        super(hunspell, self).start(dict_ids)

        if dict_ids:
            dic, aff = self._search_dict_files(dict_ids[0])
            _logger.info("using hunspell files '{}', '{}'" \
                            .format(dic, aff))
            if dic:
                try:
                    self._osk_hunspell = osk.Hunspell(aff, dic)
                    _logger.info("dictionary encoding '{}'" \
                                 .format(self._osk_hunspell.get_encoding()))
                except Exception as e:
                    _logger.error("failed to create hunspell backend: " + \
                                  unicode_str(e))
                    self._osk_hunspell = None

    def stop(self):
        super(hunspell, self).stop()
        if self.is_running():
            self._osk_hunspell = None
            self._active_dicts = None

    def is_running(self):
        return not self._osk_hunspell is None

    SPLITWORDS = re.compile("[^-_\s]+", re.UNICODE|re.DOTALL)

    def query(self, text):
        """
        Query for spelling suggestions.
        Text may contain one or more words. Each word generates its own
        list of suggestions. Word boundaries are chosen to match the
        behavior of the hunspell command line tool, i.e. split at '-','_'
        and whitespace.

        Doctests:
        # one prediction token, two words for the spell checker
        >>> sp = hunspell(["en_US"])
        >>> q = sp.query("conter_trop")
        >>> q  # doctest: +ELLIPSIS
        [[[0, 6, 'conter'], [...
        >>> len(q)
        2

        # dashes
        >>> sp = hunspell(["en_US"])
        >>> sp.query("ubuntu-system")  # doctest: +ELLIPSIS
        [[[0, 6, 'ubuntu'], ['Ubuntu', 'Kubuntu', 'urbanite', 'urbanity']]]

        # unrecognized word returns error span with zero choices (# mark)
        >>> q = sp.query("ἄναρχος")
        >>> q  # doctest:
        [[[0, 7, 'ἄναρχος'], []]]
        >>> len(q)
        1

        # dictionaries come in various encodings, spot-check German
        >>> sp = hunspell(["de_DE"])
        >>> sp.query("Frühstück")
        []
        >>> sp.query("Früstück")   # doctest: +ELLIPSIS
        [[[0, 8, 'Früstück'], ['Frühstück', ...

        # dictionaries come in various encodings, spot-check Spanish
        >>> sp = hunspell(["es_ES"])
        >>> sp.query("situación")
        []
        >>> sp.query("situción")   # doctest: +ELLIPSIS
        [[[0, 8, 'situción'], ['situación', ...

        # dictionaries come in various encodings, spot-check Portuguese
        >>> sp = hunspell(["pt_PT"])
        >>> sp.query("questão")
        []
        >>> sp.query("quetão")   # doctest: +ELLIPSIS
        [[[0, 6, 'quetão'], ['questão', ...

        # dictionaries come in various encodings, spot-check French
        >>> sp = hunspell(["fr_FR"])
        >>> sp.query("garçon")
        []
        >>> sp.query("garçoon")   # doctest: +ELLIPSIS
        [[[0, 7, 'garçoon'], ['garçon', ...

        # dictionaries come in various encodings, spot-check Italian
        >>> sp = hunspell(["it_IT"])
        >>> sp.query("parlerò")
        []
        >>> sp.query("parllerò")   # doctest: +ELLIPSIS
        [[[0, 8, 'parllerò'], ['parlerò', ...

        # dictionaries come in various encodings, spot-check Russian
        >>> sp = hunspell(["ru_RU"])
        >>> sp.query("ОБЕДЕННЫЙ") # dinner
        []
        >>> sp.query("ОБЕДЕНЫЙ")   # doctest: +ELLIPSIS
        [[[0, 8, 'ОБЕДЕНЫЙ'], ['ОБЕДЕННЫЙ', ...

        # dictionaries come in various encodings, spot-check Greek
        >>> sp = hunspell(["el_GR"])
        >>> sp.query("Ωκεανού")
        []
        >>> sp.query("Ωκεαανού")   # doctest: +ELLIPSIS
        [[[0, 8, 'Ωκεαανού'], ['Ωκεανού', ...

        """
        results = []

        if self._osk_hunspell:
            matches = self.SPLITWORDS.finditer(text)
            for match in matches:
                word = match.group()
                begin = match.start()
                end   = match.end()
                span = [begin, end, word]

                try:
                    if self._osk_hunspell.spell(word) == 0:
                        suggestions = list(self._osk_hunspell.suggest(word))
                        results.append([span, suggestions])
                except UnicodeEncodeError:
                    # Assume the offending character isn't part of the
                    # target language and consider this word to be not
                    # in the dictionary, i.e. misspelled without suggestions.
                    results.append([span, []])

        return results

    def get_supported_dict_ids(self):
        """
        Return raw supported dictionary ids.
        They may not all be valid language ids, e.g. en-GB
        instead of en_GB for myspell dicts.

        Doctests:
        # Just check it does anything at all
        >>> sp = hunspell(["en_US"])
        >>> sp.get_supported_dict_ids()   # doctest: +ELLIPSIS
        ['...
        """
        dict_ids = []

        # search for dictionary files
        filenames = []
        paths = self._get_dict_paths()
        for path in paths:
            fn = os.path.join(path, "*" + os.path.extsep + "dic")
            filenames.extend(glob.glob(fn))

        # extract language ids
        for fn in filenames:
            lang_id, _ext = os.path.splitext(os.path.basename(fn))
            if not lang_id.lower().startswith("hyph"):
                dict_ids.append(lang_id)

        return dict_ids

    def _search_dict_files(self, dict_id):
        """ Locate dictionary and affix files for the given dict_id """
        paths = self._get_dict_paths()
        aff = self._find_file_in_paths(paths, dict_id, "aff")
        dic = self._find_file_in_paths(paths, dict_id, "dic")
        return dic, aff

    @staticmethod
    def _find_file_in_paths(paths, name, extension):
        """ Search for a file name in multiple paths. """
        for path in paths:
            file = os.path.join(path, name + os.path.extsep + extension)
            if os.path.exists(file):
                return file
        return None

    def _get_dict_paths(self):
        """
        Return all paths where dictionaries may be located.
        Path logic taken from the source of the hunspell command line tool.
        """
        paths = []

        pathsep = os.path.pathsep
        LIBDIRS = [
            "/usr/share/hunspell",
            "/usr/share/myspell",
            "/usr/share/myspell/dicts",
            "/Library/Spelling"]
        USEROOODIRS = [
            ".openoffice.org/3/user/wordbook",
            ".openoffice.org2/user/wordbook",
            ".openoffice.org2.0/user/wordbook",
            "Library/Spelling"]
        OOODIRS = [
            "/opt/openoffice.org/basis3.0/share/dict/ooo",
            "/usr/lib/openoffice.org/basis3.0/share/dict/ooo",
            "/opt/openoffice.org2.4/share/dict/ooo,"
            "/usr/lib/openoffice.org2.4/share/dict/ooo,",
            "/opt/openoffice.org2.3/share/dict/ooo",
            "/usr/lib/openoffice.org2.3/share/dict/ooo",
            "/opt/openoffice.org2.2/share/dict/ooo",
            "/usr/lib/openoffice.org2.2/share/dict/ooo",
            "/opt/openoffice.org2.1/share/dict/ooo,"
            "/usr/lib/openoffice.org2.1/share/dict/ooo",
            "/opt/openoffice.org2.0/share/dict/ooo",
            "/usr/lib/openoffice.org2.0/share/dict/ooo"]
        paths.append(".")

        dicpath = os.getenv("DICPATH")
        if dicpath:
            paths.extend(dicpath.split(pathsep))

        paths.extend(LIBDIRS)

        home = os.getenv("HOME")
        if home:
            paths.append(home)
            paths.extend(USEROOODIRS)

        paths.extend(OOODIRS)

        return paths


class SCBackend_cmd(SCBackend):
    """ Abstract base class of command line backends """

    def __init__(self, dict_ids = None):
        super(SCBackend_cmd, self).__init__(dict_ids)
        self._p = None

    def stop(self):
        super(SCBackend_cmd, self).__init__(dict_ids)
        if self.is_running():
            self._p.terminate()
            self._p.wait()
            self._p = None

    def is_running(self):
        return not self._p is None

    def query(self, text):
        """
        Query for spelling suggestions.
        Text may contain one or more words. Each word generates its own
        list of suggestions. The spell checker backend decides about
        word boundaries.
        """
        results = []

        # Check if the process is still running, it might have
        # exited on start due to an unknown dictinary name.
        if self._p and not self._p.poll() is None:
            self._p = None

        if self._p:

            # unicode?
            if type(text) == type(""):
                line = "^" + text + "\n"
                line = line.encode("UTF-8")
            else: # already UTF-8 byte array
                line = b"^" + text + b"\n"

            self._p.stdin.write(line)
            self._p.stdin.flush()
            while True:
                s = self._p.stdout.readline().decode("UTF-8")
                s = s.strip()
                if not s:
                    break
                if s[:1] == "&":
                    sections = s.split(":")
                    a = sections[0].split()
                    begin = int(a[3]) - 1 # -1 for the prefixed ^
                    end   = begin + len(a[1])
                    span = [begin, end, a[1]] # begin, end, word
                    suggestions = sections[1].strip().split(', ')
                    results.append([span, suggestions])
                if s[:1] == "#":
                    sections = s.split(":")
                    a = sections[0].split()
                    begin = int(a[2]) - 1 # -1 for the prefixed ^
                    end   = begin + len(a[1])
                    span = [begin, end, a[1]] # begin, end, word
                    suggestions = []
                    results.append([span, suggestions])

        return results

    def get_supported_dict_ids(self):
        """
        Return raw supported dictionary ids.
        """
        raise NotImplementedError()

    def get_active_dict_ids(self):
        """
        Return active dictionary ids.
        """
        return self._active_dicts


class hunspell_cmd(SCBackend_cmd):
    """
    Hunspell backend, using the "hunspell" command.

    Doctests:
    # known word
    >>> sp = hunspell_cmd(["en_US"])
    >>> sp.query("test")
    []

    # unknown word
    >>> sp = hunspell_cmd(["en_US"])
    >>> sp.query("jdaskljasd")  # doctest: +ELLIPSIS
    [[...
    """
    def __init__(self, dict_ids = None):
        SCBackend.__init__(self, dict_ids)
        if dict_ids:
            self.start(dict_ids)

    def start(self, dict_ids = None):
        super(hunspell_cmd, self).start(dict_ids)

        args = ["hunspell", "-a", "-i", "UTF-8"]
        if dict_ids:
            args += ["-d", ",".join(dict_ids)]

        try:
            self._p = subprocess.Popen(args,
                                       stdin=subprocess.PIPE,
                                       stdout=subprocess.PIPE,
                                       close_fds=True)
            self._p.stdout.readline() # skip header line
        except OSError as e:
            _logger.error(_format("Failed to execute '{}', {}", \
                            " ".join(args), e))
            self._p = None

    def get_supported_dict_ids(self):
        """
        Return raw supported dictionary ids.
        They may not all be valid language ids, e.g. en-GB for myspell dicts.
        """
        dict_ids = []
        args = ["hunspell", "-D"]

        try:
            p = subprocess.Popen(args,
                                 stdin=subprocess.PIPE,
                                 stdout=subprocess.PIPE,
                                 stderr=subprocess.PIPE,
                                 close_fds=True)
            out, err = p.communicate("") # send something to shut hunspell down

            # scrape the dict_ids from stderr
            in_dicts = False
            for line in err.decode("UTF-8").split("\n"):
                if in_dicts:
                    if not "/" in line:
                        break

                    # extract language id
                    lang_id = os.path.basename(line)
                    if not lang_id.lower().startswith("hyph"):
                        dict_ids.append(lang_id)

                if line.startswith("AVAILABLE DICTIONARIES"): # not translated?
                    in_dicts = True

        except OSError as e:
            _logger.error(_format("Failed to execute '{}', {}", \
                            " ".join(args), e))
        return dict_ids


class aspell_cmd(SCBackend):
    """
    Aspell backend.

    Doctests:
    # known word
    >>> sp = aspell_cmd(["en_US"])
    >>> sp.query("test")
    []

    # unknown word
    >>> sp = aspell_cmd(["en_US"])
    >>> sp.query("jdaskljasd")  # doctest: +ELLIPSIS
    [[...
    """
    def __init__(self, dict_ids = None):
        SCBackend.__init__(self, dict_ids)
        if dict_ids:
            self.start(dict_ids)

    def start(self, dict_ids = None):
        super(aspell_cmd, self).start(dict_ids)

        args = ["aspell", "-a"]
        if dict_ids:
            args += ["-l", ",".join(dict_ids)]

        try:
            self._p = subprocess.Popen(args,
                                       stdin=subprocess.PIPE,
                                       stdout=subprocess.PIPE,
                                       close_fds=True)
            self._p.stdout.readline() # skip header line
        except OSError as e:
            _logger.error(_format("Failed to execute '{}', {}", \
                            " ".join(args), e))
            self._p = None

    def get_supported_dict_ids(self):
        """
        Return raw supported dictionary ids.
        """
        dict_ids = []
        args = ["aspell", "dump", "dicts"]
        try:
            dict_ids = subprocess.check_output(args) \
                                .decode("UTF-8").split("\n")
        except OSError as e:
            _logger.error(_format("Failed to execute '{}', {}", \
                            " ".join(args), e))
        return [id for id in dict_ids if id]