# XML "function" nodes that transform their contents

import datetime

import re

import sre_constants

import dates

import exc

import format

import maps

import strings

import term

import util

import xml_dom

import xpath

##### Exceptions

class SyntaxException(Exception):

    def __init__(self, cause):

        Exception.__init__(self, 'Invalid XML function syntax: '

            +exc.str_(cause))

class FormatException(SyntaxException): pass

##### Functions

funcs = {}

def process(node, on_error=exc.raise_):

    for child in xml_dom.NodeElemIter(node): process(child, on_error)

    name = node.tagName

    if name.startswith('_') and name in funcs:

        try:

            value = funcs[name](xml_dom.NodeTextEntryIter(node))

            xml_dom.replace_with_text(node, value)

        except Exception, e: # also catch XML func internal errors

            # Save in case another exception raised, overwriting sys.exc_info()

            exc.add_traceback(e)

            str_ = str(node)

            exc.add_msg(e, 'function:\n'+str_)

            xml_dom.replace(node, node.ownerDocument.createComment(

                '\n'+term.emph_multiline(str_).replace('--','-')))

                # comments can't contain '--'

            on_error(e)

def map_items(func, items):

    return [(name, func(value)) for name, value in items]

def cast(type_, val):

    '''Throws SyntaxException if can't cast'''

    try: return type_(val)

    except ValueError, e: raise SyntaxException(e)

def conv_items(type_, items):

    return map_items(lambda val: cast(type_, val),

        xml_dom.TextEntryOnlyIter(items))

#### XML functions

# Function names must start with _ to avoid collisions with real tags

# Functions take arguments (items)

def _ignore(items):

    '''Used to "comment out" an XML subtree'''

    return None

funcs['_ignore'] = _ignore

def _eq(items):

    items = dict(items)

    try:

        left = items['left']

        right = items['right']

    except KeyError: return '' # a value was None

    return util.bool2str(left == right)

funcs['_eq'] = _eq

def _if(items):

    items = dict(items)

    try:

        cond = items['cond']

        then = items['then']

    except KeyError, e: raise SyntaxException(e)

    else_ = items.get('else', None)

    cond = bool(cast(str, cond))

    if cond: return then

    else: return else_

funcs['_if'] = _if

def _alt(items):

    items = list(items)

    items.sort()

    try: return items[0][1] # value of lowest-numbered item

    except IndexError: return None # input got removed by e.g. SyntaxException

funcs['_alt'] = _alt

def _merge(items):

    items = list(conv_items(strings.ustr, items))

        # get *once* from iter and check types

    items.sort()

    return maps.merge_values(*[v for k, v in items])

funcs['_merge'] = _merge

def _label(items):

    items = dict(conv_items(strings.ustr, items))

        # get *once* from iter and check types

    try:

        label = items['label']

        value = items['value']

    except KeyError, e: raise SyntaxException(e)

    return label+': '+value

funcs['_label'] = _label

def _nullIf(items):

    items = dict(conv_items(str, items))

    try:

        null = items['null']

        value = items['value']

    except KeyError, e: raise SyntaxException(e)

    type_str = items.get('type', None)

    type_ = str

    if type_str == 'float': type_ = float

    return util.none_if(value, type_(null))

funcs['_nullIf'] = _nullIf

def _map(items):

    items = conv_items(str, items) # get *once* from iter and check types

    try: value = items.pop()[1] # value is last entry's value

    except IndexError, e: raise SyntaxException(e)

    map_ = dict(items)

    closed = bool(map_.pop('_closed', False))

    try: return map_[value]

    except KeyError, e:

        if closed: raise SyntaxException(e)

        else: return value

funcs['_map'] = _map

def _replace(items):

    items = conv_items(str, items) # get *once* from iter and check types

    try: value = items.pop()[1] # last entry contains value

    except IndexError, e: raise SyntaxException(e)

    try:

        for repl, with_ in items:

            if re.match(r'^\w+$', repl):

                repl = r'(?<![^\W_])'+repl+r'(?![^\W_])' # match whole word

            value = re.sub(repl, with_, value)

    except sre_constants.error, e: raise SyntaxException(e)

    return util.none_if(value, u'') # empty strings always mean None

funcs['_replace'] = _replace

def _units(items):

    items = dict(conv_items(str, items))

    try: value = items['value']

    except KeyError: return None # input is empty

    default_units = items.get('units', None)

    # DB unit conversion isn't ready yet, so just return number

    try: return format.cleanup_units(value, default_units).split(' ')[0]

    except format.MissingUnitsException, e: raise SyntaxException(e)

funcs['_units'] = _units

def _range(items):

    items = dict(conv_items(float, items))

    from_ = items.get('from', None)

    to = items.get('to', None)

    if from_ == None or to == None: return None

    return str(to - from_)

funcs['_range'] = _range

def parse_range(str_, range_sep='-'):

    default = (str_, None)

    start, sep, end = str_.partition(range_sep)

    if sep == '': return default # not a range

    if start == '' and range_sep == '-': return default # negative number

    return tuple(d.strip() for d in (start, end))

def _rangeStart(items):

    items = dict(conv_items(str, items))

    try: value = items['value']

    except KeyError: return None # input is empty

    return parse_range(value)[0]

funcs['_rangeStart'] = _rangeStart

def _rangeEnd(items):

    items = dict(conv_items(str, items))

    try: value = items['value']

    except KeyError: return None # input is empty

    return parse_range(value)[1]

funcs['_rangeEnd'] = _rangeEnd

def _avg(items):

    count = 0

    sum_ = 0.

    for name, value in conv_items(float, items):

        count += 1

        sum_ += value

    return str(sum_/count)

funcs['_avg'] = _avg

class CvException(Exception):

    def __init__(self):

        Exception.__init__(self, 'CV (coefficient of variation) values are only'

            ' allowed for ratio scale data '

            '(see <http://en.wikipedia.org/wiki/Coefficient_of_variation>)')

def _noCV(items):

    try: name, value = items.next()

    except StopIteration: return None

    if re.match('^(?i)CV *\d+$', value): raise SyntaxException(CvException())

    return value

funcs['_noCV'] = _noCV

def _date(items):

    items = conv_items(str, items) # get *once* from iter and check types

    try: str_ = dict(items)['date']

    except KeyError:

        items = dict(conv_items(int, items))

        try: items['year'] # year is required

        except KeyError, e:

            if items == {}: return None # entire date is empty

            else: raise SyntaxException(e)

        items.setdefault('month', 1)

        items.setdefault('day', 1)

        try: date = datetime.date(**items)

        except ValueError, e: raise SyntaxException(e)

    else:

        try: year = float(str_)

        except ValueError:

            try: date = dates.strtotime(str_)

            except ImportError: return str_

            except ValueError, e: raise SyntaxException(e)

        else: date = (datetime.date(int(year), 1, 1) +

            datetime.timedelta(round((year % 1.)*365)))

    try: return dates.strftime('%Y-%m-%d', date)

    except ValueError, e: raise FormatException(e)

funcs['_date'] = _date

def _dateRangeStart(items):

    items = dict(conv_items(str, items))

    try: value = items['value']

    except KeyError: return None # input is empty

    return dates.parse_date_range(value)[0]

funcs['_dateRangeStart'] = _dateRangeStart

def _dateRangeEnd(items):

    items = dict(conv_items(str, items))

    try: value = items['value']

    except KeyError: return None # input is empty

    return dates.parse_date_range(value)[1]

funcs['_dateRangeEnd'] = _dateRangeEnd

_name_parts_slices_items = [

    ('first', slice(None, 1)),

    ('middle', slice(1, -1)),

    ('last', slice(-1, None)),

]

name_parts_slices = dict(_name_parts_slices_items)

name_parts = [name for name, slice_ in _name_parts_slices_items]

def _name(items):

    items = dict(items)

    parts = []

    for part in name_parts:

        if part in items: parts.append(items[part])

    return ' '.join(parts)

funcs['_name'] = _name

def _namePart(items):

    out_items = []

    for part, value in items:

        try: slice_ = name_parts_slices[part]

        except KeyError, e: raise SyntaxException(e)

        out_items.append((part, ' '.join(value.split(' ')[slice_])))

    return _name(out_items)

funcs['_namePart'] = _namePart

def _simplifyPath(items):

    items = dict(items)

    try:

        next = cast(str, items['next'])

        require = cast(str, items['require'])

        root = items['path']

    except KeyError, e: raise SyntaxException(e)

    node = root

    while node != None:

        new_node = xpath.get_1(node, next, allow_rooted=False)

        if xpath.get_1(node, require, allow_rooted=False) == None: # empty elem

            xml_dom.replace(node, new_node) # remove current elem

            if node is root: root = new_node # also update root

        node = new_node

    return root

funcs['_simplifyPath'] = _simplifyPath