frappe/frappe/utils/data.py

# Copyright (c) 2015, Frappe Technologies Pvt. Ltd. and Contributors
# MIT License. See license.txt

from __future__ import unicode_literals

# IMPORTANT: only import safe functions as this module will be included in jinja environment
import frappe
import operator
import re, urllib, datetime, math, time
import babel.dates
from babel.core import UnknownLocaleError
from dateutil import parser
from num2words import num2words
from six.moves import html_parser as HTMLParser
from six.moves.urllib.parse import quote, urljoin
from html2text import html2text
from six import iteritems, text_type, string_types, integer_types

DATE_FORMAT = "%Y-%m-%d"
TIME_FORMAT = "%H:%M:%S.%f"
DATETIME_FORMAT = DATE_FORMAT + " " + TIME_FORMAT

# datetime functions
def getdate(string_date=None):
	"""
		 Coverts string date (yyyy-mm-dd) to datetime.date object
	"""
	if not string_date:
		return get_datetime().date()

	if isinstance(string_date, datetime.datetime):
		return string_date.date()

	elif isinstance(string_date, datetime.date):
		return string_date

	# dateutil parser does not agree with dates like 0000-00-00
	if not string_date or string_date=="0000-00-00":
		return None

	return parser.parse(string_date).date()

def get_datetime(datetime_str=None):
	if not datetime_str:
		return now_datetime()

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

	elif isinstance(datetime_str, (list, tuple)):
		return datetime.datetime(datetime_str)

	elif isinstance(datetime_str, datetime.date):
		return datetime.datetime.combine(datetime_str, datetime.time())

	# dateutil parser does not agree with dates like 0000-00-00
	if not datetime_str or (datetime_str or "").startswith("0000-00-00"):
		return None

	try:
		return datetime.datetime.strptime(datetime_str, DATETIME_FORMAT)
	except ValueError:
		return parser.parse(datetime_str)

def to_timedelta(time_str):
	if isinstance(time_str, string_types):
		t = parser.parse(time_str)
		return datetime.timedelta(hours=t.hour, minutes=t.minute, seconds=t.second, microseconds=t.microsecond)

	else:
		return time_str

def add_to_date(date, years=0, months=0, days=0, hours=0, as_string=False, as_datetime=False):
	"""Adds `days` to the given date"""
	from dateutil.relativedelta import relativedelta

	if date==None:
		date = now_datetime()

	if hours:
		as_datetime = True

	if isinstance(date, string_types):
		as_string = True
		if " " in date:
			as_datetime = True
		date = parser.parse(date)

	date = date + relativedelta(years=years, months=months, days=days, hours=hours)

	if as_string:
		if as_datetime:
			return date.strftime(DATETIME_FORMAT)
		else:
			return date.strftime(DATE_FORMAT)
	else:
		return date

def add_days(date, days):
	return add_to_date(date, days=days)

def add_months(date, months):
	return add_to_date(date, months=months)

def add_years(date, years):
	return add_to_date(date, years=years)

def date_diff(string_ed_date, string_st_date):
	return (getdate(string_ed_date) - getdate(string_st_date)).days

def time_diff(string_ed_date, string_st_date):
	return get_datetime(string_ed_date) - get_datetime(string_st_date)

def time_diff_in_seconds(string_ed_date, string_st_date):
	return time_diff(string_ed_date, string_st_date).total_seconds()

def time_diff_in_hours(string_ed_date, string_st_date):
	return round(float(time_diff(string_ed_date, string_st_date).total_seconds()) / 3600, 6)

def now_datetime():
	dt = convert_utc_to_user_timezone(datetime.datetime.utcnow())
	return dt.replace(tzinfo=None)

def get_timestamp(date):
	return time.mktime(getdate(date).timetuple())

def get_eta(from_time, percent_complete):
	diff = time_diff(now_datetime(), from_time).total_seconds()
	return str(datetime.timedelta(seconds=(100 - percent_complete) / percent_complete * diff))

def _get_time_zone():
	return frappe.db.get_system_setting('time_zone') or 'Asia/Kolkata'

def get_time_zone():
	if frappe.local.flags.in_test:
		return _get_time_zone()

	return frappe.cache().get_value("time_zone", _get_time_zone)

def convert_utc_to_user_timezone(utc_timestamp):
	from pytz import timezone, UnknownTimeZoneError
	utcnow = timezone('UTC').localize(utc_timestamp)
	try:
		return utcnow.astimezone(timezone(get_time_zone()))
	except UnknownTimeZoneError:
		return utcnow

def now():
	"""return current datetime as yyyy-mm-dd hh:mm:ss"""
	if frappe.flags.current_date:
		return getdate(frappe.flags.current_date).strftime(DATE_FORMAT) + " " + \
			now_datetime().strftime(TIME_FORMAT)
	else:
		return now_datetime().strftime(DATETIME_FORMAT)

def nowdate():
	"""return current date as yyyy-mm-dd"""
	return now_datetime().strftime(DATE_FORMAT)

def today():
	return nowdate()

def nowtime():
	"""return current time in hh:mm"""
	return now_datetime().strftime(TIME_FORMAT)

def get_first_day(dt, d_years=0, d_months=0):
	"""
	 Returns the first day of the month for the date specified by date object
	 Also adds `d_years` and `d_months` if specified
	"""
	dt = getdate(dt)

	# d_years, d_months are "deltas" to apply to dt
	overflow_years, month = divmod(dt.month + d_months - 1, 12)
	year = dt.year + d_years + overflow_years

	return datetime.date(year, month + 1, 1)

def get_last_day(dt):
	"""
	 Returns last day of the month using:
	 `get_first_day(dt, 0, 1) + datetime.timedelta(-1)`
	"""
	return get_first_day(dt, 0, 1) + datetime.timedelta(-1)


def get_time(time_str):
	if isinstance(time_str, datetime.datetime):
		return time_str.time()
	elif isinstance(time_str, datetime.time):
		return time_str
	else:
		if isinstance(time_str, datetime.timedelta):
			time_str = str(time_str)
		return parser.parse(time_str).time()

def get_datetime_str(datetime_obj):
	if isinstance(datetime_obj, string_types):
		datetime_obj = get_datetime(datetime_obj)

	return datetime_obj.strftime(DATETIME_FORMAT)

def get_user_format():
	if getattr(frappe.local, "user_format", None) is None:
		frappe.local.user_format = frappe.db.get_default("date_format")

	return frappe.local.user_format or "yyyy-mm-dd"

def formatdate(string_date=None, format_string=None):
	"""
	 	Convers the given string date to :data:`user_format`
		User format specified in defaults

		 Examples:

		 * dd-mm-yyyy
		 * mm-dd-yyyy
		 * dd/mm/yyyy
	"""

	if not string_date:
		return ''

	date = getdate(string_date)
	if not format_string:
		format_string = get_user_format().replace("mm", "MM")

	try:
		formatted_date = babel.dates.format_date(date, format_string, locale=(frappe.local.lang or "").replace("-", "_"))
	except UnknownLocaleError:
		formatted_date = date.strftime("%Y-%m-%d")
	return formatted_date

def format_time(txt):
	try:
		formatted_time = babel.dates.format_time(get_time(txt), locale=(frappe.local.lang or "").replace("-", "_"))
	except UnknownLocaleError:
		formatted_time = get_time(txt).strftime("%H:%M:%S")
	return formatted_time

def format_datetime(datetime_string, format_string=None):
	if not datetime_string:
		return

	datetime = get_datetime(datetime_string)
	if not format_string:
		format_string = get_user_format().replace("mm", "MM") + " HH:mm:ss"

	try:
		formatted_datetime = babel.dates.format_datetime(datetime, format_string, locale=(frappe.local.lang or "").replace("-", "_"))
	except UnknownLocaleError:
		formatted_datetime = datetime.strftime('%Y-%m-%d %H:%M:%S')
	return formatted_datetime

def global_date_format(date):
	"""returns localized date in the form of January 1, 2012"""
	date = getdate(date)
	formatted_date = babel.dates.format_date(date, locale=(frappe.local.lang or "en").replace("-", "_"), format="long")
	return formatted_date

def has_common(l1, l2):
	"""Returns truthy value if there are common elements in lists l1 and l2"""
	return set(l1) & set(l2)

def flt(s, precision=None):
	"""Convert to float (ignore commas)"""
	if isinstance(s, string_types):
		s = s.replace(',','')

	try:
		num = float(s)
		if precision is not None:
			num = rounded(num, precision)
	except Exception:
		num = 0

	return num

def cint(s):
	"""Convert to integer"""
	try: num = int(float(s))
	except: num = 0
	return num

def cstr(s, encoding='utf-8'):
	return frappe.as_unicode(s, encoding)

def rounded(num, precision=0):
	"""round method for round halfs to nearest even algorithm aka banker's rounding - compatible with python3"""
	precision = cint(precision)
	multiplier = 10 ** precision

	# avoid rounding errors
	num = round(num * multiplier if precision else num, 8)

	floor = math.floor(num)
	decimal_part = num - floor

	if not precision and decimal_part == 0.5:
		num = floor if (floor % 2 == 0) else floor + 1
	else:
		num = round(num)

	return (num / multiplier) if precision else num

def remainder(numerator, denominator, precision=2):
	precision = cint(precision)
	multiplier = 10 ** precision

	if precision:
		_remainder = ((numerator * multiplier) % (denominator * multiplier)) / multiplier
	else:
		_remainder = numerator % denominator

	return flt(_remainder, precision);

def round_based_on_smallest_currency_fraction(value, currency, precision=2):
	smallest_currency_fraction_value = flt(frappe.db.get_value("Currency",
		currency, "smallest_currency_fraction_value"))

	if smallest_currency_fraction_value:
		remainder_val = remainder(value, smallest_currency_fraction_value, precision)
		if remainder_val > (smallest_currency_fraction_value / 2):
			value += smallest_currency_fraction_value - remainder_val
		else:
			value -= remainder_val
	else:
		value = rounded(value)

	return flt(value, precision)

def encode(obj, encoding="utf-8"):
	if isinstance(obj, list):
		out = []
		for o in obj:
			if isinstance(o, text_type):
				out.append(o.encode(encoding))
			else:
				out.append(o)
		return out
	elif isinstance(obj, text_type):
		return obj.encode(encoding)
	else:
		return obj

def parse_val(v):
	"""Converts to simple datatypes from SQL query results"""
	if isinstance(v, (datetime.date, datetime.datetime)):
		v = text_type(v)
	elif isinstance(v, datetime.timedelta):
		v = ":".join(text_type(v).split(":")[:2])
	elif isinstance(v, integer_types):
		v = int(v)
	return v

def fmt_money(amount, precision=None, currency=None):
	"""
	Convert to string with commas for thousands, millions etc
	"""
	number_format = frappe.db.get_default("number_format") or "#,###.##"
	if precision is None:
		precision = cint(frappe.db.get_default('currency_precision')) or None

	decimal_str, comma_str, number_format_precision = get_number_format_info(number_format)

	if precision is None:
		precision = number_format_precision

	# 40,000 -> 40,000.00
	# 40,000.00000 -> 40,000.00
	# 40,000.23000 -> 40,000.23
	if decimal_str:
		parts = str(amount).split(decimal_str)
		decimals = parts[1] if len(parts) > 1 else ''
		if precision > 2:
			if len(decimals) < 3:
				precision = 2
			elif len(decimals) < precision:
				precision = len(decimals)

	amount = '%.*f' % (precision, flt(amount))
	if amount.find('.') == -1:
		decimals = ''
	else:
		decimals = amount.split('.')[1]

	parts = []
	minus = ''
	if flt(amount) < 0:
		minus = '-'

	amount = cstr(abs(flt(amount))).split('.')[0]

	if len(amount) > 3:
		parts.append(amount[-3:])
		amount = amount[:-3]

		val = number_format=="#,##,###.##" and 2 or 3

		while len(amount) > val:
			parts.append(amount[-val:])
			amount = amount[:-val]

	parts.append(amount)

	parts.reverse()

	amount = comma_str.join(parts) + ((precision and decimal_str) and (decimal_str + decimals) or "")
	amount = minus + amount

	if currency and frappe.defaults.get_global_default("hide_currency_symbol") != "Yes":
		symbol = frappe.db.get_value("Currency", currency, "symbol") or currency
		amount = symbol + " " + amount

	return amount

number_format_info = {
	"#,###.##": (".", ",", 2),
	"#.###,##": (",", ".", 2),
	"# ###.##": (".", " ", 2),
	"# ###,##": (",", " ", 2),
	"#'###.##": (".", "'", 2),
	"#, ###.##": (".", ", ", 2),
	"#,##,###.##": (".", ",", 2),
	"#,###.###": (".", ",", 3),
	"#.###": ("", ".", 0),
	"#,###": ("", ",", 0)
}

def get_number_format_info(format):
	return number_format_info.get(format) or (".", ",", 2)

#
# convet currency to words
#
def money_in_words(number, main_currency = None, fraction_currency=None):
	"""
	Returns string in words with currency and fraction currency.
	"""
	from frappe.utils import get_defaults
	_ = frappe._

	try:
		# note: `flt` returns 0 for invalid input and we don't want that
		number = float(number)
	except ValueError:
		return ""

	number = flt(number)
	if number < 0:
		return ""

	d = get_defaults()
	if not main_currency:
		main_currency = d.get('currency', 'INR')
	if not fraction_currency:
		fraction_currency = frappe.db.get_value("Currency", main_currency, "fraction") or _("Cent")

	number_format = frappe.db.get_value("Currency", main_currency, "number_format", cache=True) or \
		frappe.db.get_default("number_format") or "#,###.##"

	fraction_length = get_number_format_info(number_format)[2]

	n = "%.{0}f".format(fraction_length) % number

	numbers = n.split('.')
	main, fraction =  numbers if len(numbers) > 1 else [n, '00']

	if len(fraction) < fraction_length:
		zeros = '0' * (fraction_length - len(fraction))
		fraction += zeros

	in_million = True
	if number_format == "#,##,###.##": in_million = False

	# 0.00
	if main == '0' and fraction in ['00', '000']:
		out = "{0} {1}".format(main_currency, _('Zero'))
	# 0.XX
	elif main == '0':
		out = _(in_words(fraction, in_million).title()) + ' ' + fraction_currency
	else:
		out = main_currency + ' ' + _(in_words(main, in_million).title())
		if cint(fraction):
			out = out + ' ' + _('and') + ' ' + _(in_words(fraction, in_million).title()) + ' ' + fraction_currency

	return out + ' ' + _('only.')

#
# convert number to words
#
def in_words(integer, in_million=True):
	"""
	Returns string in words for the given integer.
	"""
	locale = 'en_IN' if not in_million else frappe.local.lang
	integer = int(integer)
	try:
		ret = num2words(integer, lang=locale)
	except NotImplementedError:
		ret = num2words(integer, lang='en')
	return ret.replace('-', ' ')

def is_html(text):
	out = False
	for key in ["<br>", "<p", "<img", "<div"]:
		if key in text:
			out = True
			break
	return out


# from Jinja2 code
_striptags_re = re.compile(r'(<!--.*?-->|<[^>]*>)')
def strip_html(text):
	"""removes anything enclosed in and including <>"""
	return _striptags_re.sub("", text)

def escape_html(text):
	html_escape_table = {
		"&": "&amp;",
		'"': "&quot;",
		"'": "&apos;",
		">": "&gt;",
		"<": "&lt;",
	}

	return "".join(html_escape_table.get(c,c) for c in text)

def pretty_date(iso_datetime):
	"""
		Takes an ISO time and returns a string representing how
		long ago the date represents.
		Ported from PrettyDate by John Resig
	"""
	from frappe import _
	if not iso_datetime: return ''
	import math

	if isinstance(iso_datetime, string_types):
		iso_datetime = datetime.datetime.strptime(iso_datetime, DATETIME_FORMAT)
	now_dt = datetime.datetime.strptime(now(), DATETIME_FORMAT)
	dt_diff = now_dt - iso_datetime

	# available only in python 2.7+
	# dt_diff_seconds = dt_diff.total_seconds()

	dt_diff_seconds = dt_diff.days * 86400.0 + dt_diff.seconds

	dt_diff_days = math.floor(dt_diff_seconds / 86400.0)

	# differnt cases
	if dt_diff_seconds < 60.0:
		return _('just now')
	elif dt_diff_seconds < 120.0:
		return _('1 minute ago')
	elif dt_diff_seconds < 3600.0:
		return _('{0} minutes ago').format(cint(math.floor(dt_diff_seconds / 60.0)))
	elif dt_diff_seconds < 7200.0:
		return _('1 hour ago')
	elif dt_diff_seconds < 86400.0:
		return _('{0} hours ago').format(cint(math.floor(dt_diff_seconds / 3600.0)))
	elif dt_diff_days == 1.0:
		return _('Yesterday')
	elif dt_diff_days < 7.0:
		return _('{0} days ago').format(cint(dt_diff_days))
	elif dt_diff_days < 12:
		return _('1 weeks ago')
	elif dt_diff_days < 31.0:
		return _('{0} weeks ago').format(cint(math.ceil(dt_diff_days / 7.0)))
	elif dt_diff_days < 46:
		return _('1 month ago')
	elif dt_diff_days < 365.0:
		return _('{0} months ago').format(cint(math.ceil(dt_diff_days / 30.0)))
	elif dt_diff_days < 550.0:
		return _('1 year ago')
	else:
		return '{0} years ago'.format(cint(math.floor(dt_diff_days / 365.0)))

def comma_or(some_list):
	return comma_sep(some_list, frappe._("{0} or {1}"))

def comma_and(some_list):
	return comma_sep(some_list, frappe._("{0} and {1}"))

def comma_sep(some_list, pattern):
	if isinstance(some_list, (list, tuple)):
		# list(some_list) is done to preserve the existing list
		some_list = [text_type(s) for s in list(some_list)]
		if not some_list:
			return ""
		elif len(some_list) == 1:
			return some_list[0]
		else:
			some_list = ["'%s'" % s for s in some_list]
			return pattern.format(", ".join(frappe._(s) for s in some_list[:-1]), some_list[-1])
	else:
		return some_list

def new_line_sep(some_list):
	if isinstance(some_list, (list, tuple)):
		# list(some_list) is done to preserve the existing list
		some_list = [text_type(s) for s in list(some_list)]
		if not some_list:
			return ""
		elif len(some_list) == 1:
			return some_list[0]
		else:
			some_list = ["%s" % s for s in some_list]
			return format("\n ".join(some_list))
	else:
		return some_list


def filter_strip_join(some_list, sep):
	"""given a list, filter None values, strip spaces and join"""
	return (cstr(sep)).join((cstr(a).strip() for a in filter(None, some_list)))

def get_url(uri=None, full_address=False):
	"""get app url from request"""
	host_name = frappe.local.conf.host_name or frappe.local.conf.hostname

	if uri and (uri.startswith("http://") or uri.startswith("https://")):
		return uri

	if not host_name:
		if hasattr(frappe.local, "request") and frappe.local.request and frappe.local.request.host:
			protocol = 'https://' if 'https' == frappe.get_request_header('X-Forwarded-Proto', "") else 'http://'
			host_name = protocol + frappe.local.request.host

		elif frappe.local.site:
			protocol = 'http://'

			if frappe.local.conf.ssl_certificate:
				protocol = 'https://'

			elif frappe.local.conf.wildcard:
				domain = frappe.local.conf.wildcard.get('domain')
				if domain and frappe.local.site.endswith(domain) and frappe.local.conf.wildcard.get('ssl_certificate'):
					protocol = 'https://'

			host_name = protocol + frappe.local.site

		else:
			host_name = frappe.db.get_value("Website Settings", "Website Settings",
				"subdomain")

			if not host_name:
				host_name = "http://localhost"

	if host_name and not (host_name.startswith("http://") or host_name.startswith("https://")):
		host_name = "http://" + host_name

	if not uri and full_address:
		uri = frappe.get_request_header("REQUEST_URI", "")

	if frappe.conf.http_port:
		host_name = host_name + ':' + str(frappe.conf.http_port)

	url = urljoin(host_name, uri) if uri else host_name

	return url

def get_host_name():
	return get_url().rsplit("//", 1)[-1]

def get_link_to_form(doctype, name, label=None):
	if not label: label = name

	return """<a href="{0}">{1}</a>""".format(get_url_to_form(doctype, name), label)

def get_url_to_form(doctype, name):
	return get_url(uri = "desk#Form/{0}/{1}".format(quoted(doctype), quoted(name)))

def get_url_to_list(doctype):
	return get_url(uri = "desk#List/{0}".format(quoted(doctype)))

def get_url_to_report(name, report_type = None, doctype = None):
	if report_type == "Report Builder":
		return get_url(uri = "desk#Report/{0}/{1}".format(quoted(doctype), quoted(name)))
	else:
		return get_url(uri = "desk#query-report/{0}".format(quoted(name)))

operator_map = {
	# startswith
	"^": lambda a, b: (a or "").startswith(b),

	# in or not in a list
	"in": lambda a, b: operator.contains(b, a),
	"not in": lambda a, b: not operator.contains(b, a),

	# comparison operators
	"=": lambda a, b: operator.eq(a, b),
	"!=": lambda a, b: operator.ne(a, b),
	">": lambda a, b: operator.gt(a, b),
	"<": lambda a, b: operator.lt(a, b),
	">=": lambda a, b: operator.ge(a, b),
	"<=": lambda a, b: operator.le(a, b),
	"not None": lambda a, b: a and True or False,
	"None": lambda a, b: (not a) and True or False
}

def evaluate_filters(doc, filters):
	'''Returns true if doc matches filters'''
	if isinstance(filters, dict):
		for key, value in iteritems(filters):
			f = get_filter(None, {key:value})
			if not compare(doc.get(f.fieldname), f.operator, f.value):
				return False

	elif isinstance(filters, (list, tuple)):
		for d in filters:
			f = get_filter(None, d)
			if not compare(doc.get(f.fieldname), f.operator, f.value):
				return False

	return True


def compare(val1, condition, val2):
	ret = False
	if condition in operator_map:
		ret = operator_map[condition](val1, val2)

	return ret

def get_filter(doctype, f):
	"""Returns a _dict like

		{
			"doctype":
			"fieldname":
			"operator":
			"value":
		}
	"""
	from frappe.model import default_fields, optional_fields

	if isinstance(f, dict):
		key, value = next(iter(f.items()))
		f = make_filter_tuple(doctype, key, value)

	if not isinstance(f, (list, tuple)):
		frappe.throw(frappe._("Filter must be a tuple or list (in a list)"))

	if len(f) == 3:
		f = (doctype, f[0], f[1], f[2])

	elif len(f) != 4:
		frappe.throw(frappe._("Filter must have 4 values (doctype, fieldname, operator, value): {0}").format(str(f)))

	f = frappe._dict(doctype=f[0], fieldname=f[1], operator=f[2], value=f[3])

	if not f.operator:
		# if operator is missing
		f.operator = "="

	valid_operators = ("=", "!=", ">", "<", ">=", "<=", "like", "not like", "in", "not in", "between")
	if f.operator.lower() not in valid_operators:
		frappe.throw(frappe._("Operator must be one of {0}").format(", ".join(valid_operators)))


	if f.doctype and (f.fieldname not in default_fields + optional_fields):
		# verify fieldname belongs to the doctype
		meta = frappe.get_meta(f.doctype)
		if not meta.has_field(f.fieldname):

			# try and match the doctype name from child tables
			for df in meta.get_table_fields():
				if frappe.get_meta(df.options).has_field(f.fieldname):
					f.doctype = df.options
					break

	return f

def make_filter_tuple(doctype, key, value):
	'''return a filter tuple like [doctype, key, operator, value]'''
	if isinstance(value, (list, tuple)):
		return [doctype, key, value[0], value[1]]
	else:
		return [doctype, key, "=", value]

def make_filter_dict(filters):
	'''convert this [[doctype, key, operator, value], ..]
	to this { key: (operator, value), .. }
	'''
	_filter = frappe._dict()
	for f in filters:
		_filter[f[1]] = (f[2], f[3])

	return _filter

def scrub_urls(html):
	html = expand_relative_urls(html)
	# encoding should be responsibility of the composer
	# html = quote_urls(html)
	return html

def expand_relative_urls(html):
	# expand relative urls
	url = get_url()
	if url.endswith("/"): url = url[:-1]

	def _expand_relative_urls(match):
		to_expand = list(match.groups())

		if not to_expand[2].startswith('mailto') and not to_expand[2].startswith('data:'):
			if not to_expand[2].startswith("/"):
				to_expand[2] = "/" + to_expand[2]
			to_expand.insert(2, url)

		if 'url' in to_expand[0] and to_expand[1].startswith('(') and to_expand[-1].endswith(')'):
			# background-image: url('/assets/...') - workaround for wkhtmltopdf print-media-type
			to_expand.append(' !important')

		return "".join(to_expand)

	html = re.sub('(href|src){1}([\s]*=[\s]*[\'"]?)((?!http)[^\'" >]+)([\'"]?)', _expand_relative_urls, html)

	# background-image: url('/assets/...')
	html = re.sub('(:[\s]?url)(\([\'"]?)([^\)]*)([\'"]?\))', _expand_relative_urls, html)

	return html

def quoted(url):
	return cstr(quote(encode(url), safe=b"~@#$&()*!+=:;,.?/'"))

def quote_urls(html):
	def _quote_url(match):
		groups = list(match.groups())
		groups[2] = quoted(groups[2])
		return "".join(groups)
	return re.sub('(href|src){1}([\s]*=[\s]*[\'"]?)((?:http)[^\'">]+)([\'"]?)',
		_quote_url, html)

def unique(seq):
	"""use this instead of list(set()) to preserve order of the original list.
	Thanks to Stackoverflow: http://stackoverflow.com/questions/480214/how-do-you-remove-duplicates-from-a-list-in-python-whilst-preserving-order"""

	seen = set()
	seen_add = seen.add
	return [ x for x in seq if not (x in seen or seen_add(x)) ]

def strip(val, chars=None):
	# \ufeff is no-width-break, \u200b is no-width-space
	return (val or "").replace("\ufeff", "").replace("\u200b", "").strip(chars)

def to_markdown(html):
	text = None
	try:
		text = html2text(html)
	except HTMLParser.HTMLParseError:
		pass

	return text