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

import re
import sys
import json
from datetime import datetime
from waybackpy.exceptions import WaybackError
from waybackpy.__version__ import __version__

if sys.version_info >= (3, 0):  # If the python ver >= 3
    from urllib.request import Request, urlopen
    from urllib.error import URLError
else:  # For python2.x
    from urllib2 import Request, urlopen, URLError

default_UA = "waybackpy python package - https://github.com/akamhy/waybackpy"


def _archive_url_parser(header):
    """Parse out the archive from header."""
    # Regex1
    arch = re.search(
        r"rel=\"memento.*?(web\.archive\.org/web/[0-9]{14}/.*?)>", str(header)
    )
    if arch:
        return arch.group(1)
    # Regex2
    arch = re.search(r"X-Cache-Key:\shttps(.*)[A-Z]{2}", str(header))
    if arch:
        return arch.group(1)
    raise WaybackError(
        "No archive URL found in the API response. "
        "This version of waybackpy (%s) is likely out of date. Visit "
        "https://github.com/akamhy/waybackpy for the latest version "
        "of waybackpy.\nHeader:\n%s" % (__version__, str(header))
    )


def _wayback_timestamp(**kwargs):
    """Return a formatted timestamp."""
    return "".join(
        str(kwargs[key]).zfill(2) for key in ["year", "month", "day", "hour", "minute"]
    )


def _get_response(req):
    """Get response for the supplied request."""
    try:
        response = urlopen(req)  # nosec
    except Exception:
        try:
            response = urlopen(req)  # nosec
        except Exception as e:
            raise WaybackError(e)
    return response


class Url:
    """waybackpy Url object"""

    def __init__(self, url, user_agent=default_UA):
        self.url = url
        self.user_agent = user_agent
        self._url_check()  # checks url validity on init.

    def __repr__(self):
        return "waybackpy.Url(url=%s, user_agent=%s)" % (self.url, self.user_agent)

    def __str__(self):
        return "%s" % self._clean_url()

    def __len__(self):
        return len(self._clean_url())

    def _url_check(self):
        """Check for common URL problems."""
        if "." not in self.url:
            raise URLError("'%s' is not a vaild URL." % self.url)

    def _clean_url(self):
        """Fix the URL, if possible."""
        return str(self.url).strip().replace(" ", "_")

    def save(self):
        """Create a new Wayback Machine archive for this URL."""
        request_url = "https://web.archive.org/save/" + self._clean_url()
        hdr = {"User-Agent": "%s" % self.user_agent}  # nosec
        req = Request(request_url, headers=hdr)  # nosec
        header = _get_response(req).headers
        return "https://" + _archive_url_parser(header)

    def _get(self, url="", user_agent="", encoding=""):
        """Return the source code of the supplied URL.
        If encoding is not supplied, it is auto-detected from the response.
        """
        if not url:
            url = self._clean_url()
        if not user_agent:
            user_agent = self.user_agent

        hdr = {"User-Agent": "%s" % user_agent}
        req = Request(url, headers=hdr)  # nosec
        response = _get_response(req)
        if not encoding:
            try:
                encoding = response.headers["content-type"].split("charset=")[-1]
            except AttributeError:
                encoding = "UTF-8"
        return response.read().decode(encoding.replace("text/html", "UTF-8", 1))

    def near(self, year=None, month=None, day=None, hour=None, minute=None):
        """Return the closest Wayback Machine archive to the time supplied.

        Supported params are year, month, day, hour and minute.
        Any non-supplied parameters default to the current time.
        """
        now = datetime.utcnow().timetuple()
        timestamp = _wayback_timestamp(
            year=year if year else now.tm_year,
            month=month if month else now.tm_mon,
            day=day if day else now.tm_mday,
            hour=hour if hour else now.tm_hour,
            minute=minute if minute else now.tm_min,
        )

        request_url = "https://archive.org/wayback/available?url=%s&timestamp=%s" % (
            self._clean_url(),
            timestamp,
        )
        hdr = {"User-Agent": "%s" % self.user_agent}
        req = Request(request_url, headers=hdr)  # nosec
        response = _get_response(req)
        data = json.loads(response.read().decode("UTF-8"))
        if not data["archived_snapshots"]:
            raise WaybackError(
                "'%s' is not yet archived. Use wayback.Url(url, user_agent).save() "
                "to create a new archive." % self._clean_url()
            )
        archive_url = data["archived_snapshots"]["closest"]["url"]
        # wayback machine returns http sometimes, idk why? But they support https
        archive_url = archive_url.replace(
            "http://web.archive.org/web/", "https://web.archive.org/web/", 1
        )
        return archive_url

    def oldest(self, year=1994):
        """Return the oldest Wayback Machine archive for this URL."""
        return self.near(year=year)

    def newest(self):
        """Return the newest Wayback Machine archive available for this URL.

        Due to Wayback Machine database lag, this may not always be the
        most recent archive.
        """
        return self.near()

    def total_archives(self):
        """Returns the total number of Wayback Machine archives for this URL."""
        hdr = {"User-Agent": "%s" % self.user_agent}
        request_url = (
            "https://web.archive.org/cdx/search/cdx?url=%s&output=json&fl=statuscode"
            % self._clean_url()
        )
        req = Request(request_url, headers=hdr)  # nosec
        response = _get_response(req)
        # Most efficient method to count number of archives (yet)
        return str(response.read()).count(",")