# -*- coding: utf-8 -*-
""" harvest_tools.base

"""
from exception import ToolException
from invenio_tools import REG_AUTHOR
from plugin_dbui import get_id, UNDEF_ID


DRY_RUN = "dry run"

MSG_CRASH = "Crash: %s"
MSG_FIX_ORIGIN = "Fixed the origin field"
MSG_IN_DB = "Already in the database"
MSG_LOAD = "Load in the database"
MSG_NO_ENTRY = "Reject %s is not defined"
MSG_TOOMANY_SYNONYM = "Reject too many %s synonyms"


def family_name_fr(full_name):
    """Extract the family name when the full name is encoded as ``J. Doe``.

    Args:
        full_name (unicode): author name encoded according to French
            typographic rules.

    Returns:
        unicode: family name

    """
    return full_name[full_name.find(' ') + 1:]


def format_author_fr(name):
    """Format the author name according to French typographic rules.

    Note:
        The name stays unchanged when the formatting failed.

    Args:
        name (unicode): full name. Possible patterns are:

            * ``Family, L``
            * ``Family, P L``
            * ``Family, M -H``
            * ``Family Name, J``
            * ``Family-Name, J``
            * ``Family, F Name``
            * ``Family, First``

    Returns:
        unicode: the author name encode as ``J.-P. Doe``.

    """
    # protection
    if name == '' or name is None:
        return name

    # name are encoded Family, L
    #                  Family, P L
    #                  Family, M -H
    #                  Family Name, J
    #                  Family-Name, J
    #                  Family, F Name
    #                  Family, First
    # To avoid to deal with unicode character
    # look for non empty string \S
    match = REG_AUTHOR.match(name)

    # reformat the name as L. Family
    # or keep it as it is
    if match:
        if match.group(3) and match.group(4):
            result = '%s.%s%s. %s' % (match.group(2)[0], match.group(3)[0],
                                      match.group(4)[0], match.group(1))

        elif "-" in match.group(2):
            li = [el[0] for el in match.group(2).split("-")]
            li.append(match.group(1))
            result = "%s.-%s. %s" % tuple(li)

        else:
            result = '%s. %s' % (match.group(2)[0], match.group(1))
    else:
        result = name

    # avoid author name in upper case (R. LE FOO --> R. Le Foo)
    result = result.title()

    return result


def learn_my_authors(db,
                     authors=None,
                     id_project=None,
                     id_team=None,
                     year=None):
    """Train the rescue list of the authors of my institute.
    Authors which are not in the rescue list, are added.
    The rescue list is defined by the project, the team identifier and
    by the year.

    Warning:
        all keyword arguments have to be defined.

    Args:
        db (gluon.dal.DAL): database connection.
        authors (list): authors names
        id_project (int): the identifier of the project in the database.
        id_team (int): the identifier of the team in the database.
        year (int): the year

    """
    # get the list of authors store in the database
    row = db.my_authors(id_projects=id_project,
                        id_teams=id_team,
                        year=year)

    # no entry in the database
    if not row:
        db.my_authors[0] = dict(authors=authors,
                                id_projects=id_project,
                                id_teams=id_team,
                                year=year)
        return

    database_authors = row.authors.split(', ')

    # compare with the input list
    # and extract authors which are not in the db
    new = set(authors.split(', '))
    ref = set(database_authors)
    diff = new.difference(ref)

    # update the database
    if diff:

        # NOTE1: be careful with the string encoding
        # NOTE2: handle the case J. Foo and J. M. Foo are the same person
        elems = []
        for elem in diff:
            if isinstance(elem, unicode):
                elem = elem.encode('utf8')
                family_name = elem[elem.rfind('. ') + 2:]

                if family_name not in row.authors:
                    elems.append(elem)

        database_authors.extend(elems)
        database_authors.sort(key=family_name_fr)
        db.my_authors[row.id] = dict(authors=', '.join(database_authors))


def search_synonym(table, fieldname, value, create=False):
    """Get the database identifier for the record having the database field
    or the synonyms field matching the value.

    Note:
        The database table must have a field name *synonyms*.
        It contains a list of strings.
    Args:
        table (gluon.DAL.Table): database table.
        fieldname (unicode): field of the database table
            identified by its name.
        value (unicode): value to be matched.
        create(bool): create a new entry in the database table when
            it is ``True``

    Returns:
        int:
            * the id of the database record.
            * UNDEF_ID if value is not defined.

    Raises:
        ToolException: when more than one synonym is found.

    """
    if not value:
        return UNDEF_ID

    db = table._db

    kwargs = {}
    kwargs[fieldname] = value

    id_rec = get_id(table, **kwargs)
    if id_rec is not None:
        return id_rec

    # nothing found, have a look to the synonyms field
    query = table.synonyms.contains(value)
    setrows = db(query)

    # no synonym found, create the entry
    ncount = setrows.count()
    if ncount == 0:
        if create:
            return table.insert(**kwargs)
        else:
            msg = MSG_NO_ENTRY % table._tablename
            raise ToolException(msg)

    # one synonym found
    elif ncount == 1:
        return setrows.select(table.id).first().id

    # more than one synonyms - don't know what to choose
    else:
        msg = MSG_TOOMANY_SYNONYM % table._tablename
        raise ToolException(msg)