###
# Copyright (c) 2014-2015, James Lu
# All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are met:
#
#   * Redistributions of source code must retain the above copyright notice,
#     this list of conditions, and the following disclaimer.
#   * Redistributions in binary form must reproduce the above copyright notice,
#     this list of conditions, and the following disclaimer in the
#     documentation and/or other materials provided with the distribution.
#   * Neither the name of the author of this software nor the name of
#     contributors to this software may be used to endorse or promote products
#     derived from this software without specific prior written consent.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
# AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
# ARE DISCLAIMED.  IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
# LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
# CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
# SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
# INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
# CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
# ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
# POSSIBILITY OF SUCH DAMAGE.

###

from __future__ import unicode_literals
import supybot.utils as utils
from supybot.commands import *
import supybot.plugins as plugins
import supybot.ircutils as ircutils
import supybot.callbacks as callbacks
import supybot.log as log

from collections import OrderedDict, defaultdict
try:  # Python 3
    from urllib.parse import urlencode, quote
except ImportError:  # Python 2
    from urllib import urlencode, quote
import json
import re
try:
    from bs4 import BeautifulSoup
except ImportError:
    raise ImportError("Beautiful Soup 4 is required for this plugin: get it"
                      " at http://www.crummy.com/software/BeautifulSoup/bs4/"
                      "doc/#installing-beautiful-soup")

try:
    from supybot.i18n import PluginInternationalization
    _ = PluginInternationalization('PkgInfo')
except ImportError:
    # Placeholder that allows to run the plugin on a bot
    # without the i18n module
    _ = lambda x: x

class MadisonParser():
    def parse(self, pkg, dist, archs, codenames='', suite='', reverse=False, verbose=False):
        """Parser for the madison API at https://qa.debian.org/madison.php."""
        # This arch value implies 'all' (architecture-independent packages)
        # and 'source' (source packages), in order to prevent misleading
        # "Not found" errors.
        self.arg = {'package': pkg, 'table': dist, 'a': archs, 'c': codenames,
                    's': suite}
        self.arg = urlencode(self.arg)
        url = 'https://qa.debian.org/madison.php?text=on&' + self.arg
        log.debug("PkgInfo: Using url %s for 'vlist' command", url)
        d = OrderedDict()
        fd = utils.web.getUrlFd(url)
        for line in fd.readlines():
            L = line.decode("utf-8").split("|")
            try:
                L = map(unicode.strip, L)
            except:
                L = map(str.strip, L)
            name, version, release, archs = L
            d[release] = (version, archs)
        if d:
            if reverse:
                # *sigh*... I wish there was a better way to do this
                d = OrderedDict(reversed(tuple(d.items())))
            if verbose:
                items = ["{name} \x02({version} [{archs}])\x02".format(name=k,
                         version=v[0], archs=v[1]) for (k, v) in d.items()]
            else:
                items = ["{name} \x02({version})\x02".format(name=k,
                         version=v[0]) for (k, v) in d.items()]
            s = format('Found %n: %L', (len(d), 'result'), items)
            return s
        else:
            log.debug("PkgInfo: No results found for URL %s", url)


class PkgInfo(callbacks.Plugin):
    """Fetches package information from the repositories of
    Debian, Arch Linux, Linux Mint, Fedora, CentOS, and Ubuntu."""
    threaded = True

    def __init__(self, irc):
        self.__parent = super(PkgInfo, self)
        self.__parent.__init__(irc)
        self.addrs = {'ubuntu': 'http://packages.ubuntu.com/',
                      'debian': 'https://packages.debian.org/',
                      # This site is very, VERY slow, but it still works..
                      'debian-archive': 'http://archive.debian.net/'}
        self.unknowndist = _("Unknown distribution. This command only supports "
                             "package lookup for Debian and Ubuntu. For "
                             "commands for other distros' packages, use "
                             "'list PkgInfo'.")

    def _getDistro(self, release):
        """<release>

        Guesses the distribution from the release name."""
        release = release.lower()
        debian = ("oldoldstable", "squeeze", "oldstable", "wheezy", "stable",
                  "jessie", "testing", "sid", "unstable", "stretch", "buster",
                  "experimental")
        debian_archive = ("bo", "hamm", "slink", "potato", "woody", "sarge",
                          "etch", "lenny")
        ubuntu = ("lucid", "precise", "trusty", "utopic", "vivid", "wily")
        if release.startswith(debian):
            return "debian"
        elif release.startswith(ubuntu):
            return "ubuntu"
        elif release.startswith(debian_archive):
            return "debian-archive"

    _dependencyColor = utils.str.MultipleReplacer({'rec:': '\x0312rec:\x03',
                                                   'dep:': '\x0304dep:\x03',
                                                   'sug:': '\x0309sug:\x03',
                                                   'adep:': '\x0305adep:\x03',
                                                   'idep:': '\x0302idep:\x03',
                                                   'enh:': '\x0308enh:\x03'})
    def package(self, irc, msg, args, release, pkg, opts):
        """<release> <package> [--depends] [--source]

        Fetches information for <package> from Debian or Ubuntu's repositories.
        <release> is the codename/release name (e.g. 'trusty', 'squeeze'). If
        --depends is given, fetches dependency info for <package>. If --source
        is given, look up the source package instead of a binary."""
        pkg = pkg.lower()
        distro = self._getDistro(release)
        opts = dict(opts)
        source = 'source' in opts
        try:
            url = self.addrs[distro]
        except KeyError:
            irc.error(self.unknowndist, Raise=True)
        if source:
            url += 'source/'
        url += "{}/{}".format(release, pkg)
        try:
            fd = utils.web.getUrl(url).decode("utf-8")
        except utils.web.Error as e:
            irc.error(str(e), Raise=True)
        soup = BeautifulSoup(fd)
        if "Error" in soup.title.string:
            err = soup.find('div', attrs={"id": "content"}).find('p').string
            if "two or more packages specified" in err:
                irc.error("Unknown distribution/release.", Raise=True)
            irc.reply(err)
            return
        # If we're using the  --depends option, handle that separately.
        if 'depends' in opts:
            items = soup.find('div', {'id': 'pdeps'}).find_all('dt')
            res = []
            for item in items:
                # Get package name and related versions and architectures:
                # <packagename> (>= 1.0) [arch1, arch2]
                try:
                    deptype = item.span.text if item.find('span') else ''
                    name = '%s %s %s' % (deptype, ircutils.bold(item.a.text),
                            item.a.next_sibling.replace('\n', '').strip())
                    name = utils.str.normalizeWhitespace(name).strip()
                    if item.find('span'):
                        res.append(name)
                    else:
                        # OR dependency
                        res[-1] += " or %s" % name
                except AttributeError:
                    continue
            if res:
                s = format("Package \x02%s\x02 dependencies: %s, View more at %u", pkg,
                           ', '.join(res), url)
                s = self._dependencyColor(s)
                irc.reply(s)
            else:
                irc.error("%s doesn't seem to have any dependencies." % pkg)
            return
        desc = soup.find('meta', attrs={"name": "Description"})["content"]
        # Override description if we selected source lookup, since the meta
        # tag Description should be empty for those. Replace this with a list
        # of binary packages that the source package builds.
        if source:
            binaries = soup.find('div', {'id': "pbinaries"})
            binaries = [ircutils.bold(obj.a.text) for obj in binaries.find_all('dt')]
            desc = format('Built packages: %L', binaries)
        # Get package information from the meta tags
        keywords = soup.find('meta', attrs={"name": "Keywords"})["content"]
        keywords = keywords.replace(",", "").split()
        version = keywords[-1]
        # Handle virtual packages, showing a list of packages that provide it
        if version == "virtual":
            providing = [ircutils.bold(obj.a.text) for obj in soup.find_all('dt')]
            desc = "Virtual package provided by: %s" % ', '.join(providing[:10])
            if len(providing) > 10:
                desc += " and %s others" % (ircutils.bold(len(providing) - 10))
        s = format("Package: \x02%s (%s)\x02 in %s - %s, View more at: %u",
                   pkg, version, keywords[1], desc, url)
        irc.reply(s)
    pkg = wrap(package, ['somethingWithoutSpaces', 'somethingWithoutSpaces',
               getopts({'depends': '', 'source': ''})])

    def vlist(self, irc, msg, args, distro, pkg, opts):
        """<distribution> <package> [--reverse]

        Fetches all available version of <package> in <distribution>, if
        such package exists. Supported entries for <distribution>
        include 'debian', 'ubuntu', 'derivatives', and 'all'. If
        --reverse is given, show the newest package versions first."""
        pkg, distro = map(str.lower, (pkg, distro))
        supported = ("debian", "ubuntu", "derivatives", "all")
        if distro not in supported:
            distro = self._getDistro(distro)
            if distro is None:
                irc.error(self.unknowndist, Raise=True)
        opts = dict(opts)
        reverse = 'reverse' in opts
        archs = self.registryValue("archs") + ['all', 'source']
        archs = ','.join(set(archs))
        parser = MadisonParser()
        d = parser.parse(pkg, distro, archs, reverse=reverse,
                         verbose=self.registryValue("verbose"))
        if not d:
            irc.error("No results found.", Raise=True)
        try:
            url = "{}search?keywords={}".format(self.addrs[distro], pkg)
            d += format(" View more at: %u", url)
        except KeyError:
            pass
        irc.reply(d)
    vlist = wrap(vlist, ['somethingWithoutSpaces', 'somethingWithoutSpaces',
                 getopts({'reverse': ''})])

    def archlinux(self, irc, msg, args, pkg, opts):
        """<package> [--exact]

        Looks up <package> in the Arch Linux package repositories.
        If --exact is given, will output only exact matches.
        """
        pkg = pkg.lower()
        baseurl = 'https://www.archlinux.org/packages/search/json/?'
        if 'exact' in dict(opts):
            url = baseurl + urlencode({'name': pkg})
        else:
            url = baseurl + urlencode({'q': pkg})
        self.log.debug("PkgInfo: using url %s for 'archlinux' command", url)
        fd = utils.web.getUrl(url)
        data = json.loads(fd.decode("utf-8"))
        if data['valid'] and data['results']:
            # We want one entry per package, but the API gives one
            # entry per architecture! Remove duplicates with a set:
            results = set()
            archs = defaultdict(list)
            for x in data['results']:
                s = "\x02{name}\x02 - {desc} \x02({version})\x02".format(
                    name=x['pkgname'], desc=x['pkgdesc'], version=x['pkgver'])
                results.add(s)
                archs[s].append(x['arch'])
            count = len(results)
            items = [format("%s \x02[%s]\x02", s, ', '.join(archs[s])) for s
                     in results]
            irc.reply(format('Found %n: %L', (len(results), 'result'),
                             list(results)))
        else:
            irc.error("No results found.", Raise=True)
    archlinux = wrap(archlinux, ['somethingWithoutSpaces', getopts({'exact': ''})])

    def archaur(self, irc, msg, args, pkg):
        """<package>

        Looks up <package> in the Arch Linux AUR."""
        pkg = pkg.lower()
        baseurl = 'https://aur.archlinux.org/rpc.php?type=search&'
        url = baseurl + urlencode({'arg': pkg})
        self.log.debug("PkgInfo: using url %s for 'archaur' command", url)
        fd = utils.web.getUrl(url)
        data = json.loads(fd.decode("utf-8"))
        if data["type"] == "error":
            irc.error(data["results"], Raise=True)
        count = data["resultcount"]
        if count:
            # We want this to be limited to prevent overflow warnings
            # in the bot.
            if count > 150:
                count = '150+'
            s = format("Found %n: ", (data["resultcount"], 'result'))
            for x in data['results'][:150]:
                verboseInfo = ''
                if self.registryValue("verbose"):
                    verboseInfo = format("[ID: %s Votes: %s]", x['ID'],
                                         x['NumVotes'])
                s += "{name} - {desc} \x02({version} {verbose})\x02, " \
                    .format(name=x['Name'], desc=x['Description'],
                            version=x['Version'], verbose=verboseInfo)
            irc.reply(s[:-2])  # cut off the ", " at the end
        else:
            irc.error("No results found.", Raise=True)
    archaur = wrap(archaur, ['somethingWithoutSpaces'])

    def pkgsearch(self, irc, msg, args, distro, query):
        """<distro> <query>

        Looks up <query> in <distro>'s website. Valid <distro>'s include
        'debian', 'ubuntu', and 'debian-archive'."""
        distro = distro.lower()
        if distro not in self.addrs.keys():
            distro = self._getDistro(distro)
        try:
            url = '%ssearch?keywords=%s' % (self.addrs[distro], quote(query))
        except KeyError:
            irc.error(self.unknowndist, Raise=True)
        try:
            fd = utils.web.getUrl(url).decode("utf-8")
        except utils.web.Error as e:
            irc.error(str(e), Raise=True)
        soup = BeautifulSoup(fd)
        # Debian/Ubuntu use h3 for result names in the format 'Package abcd'
        results = [pkg.string.split()[1] for pkg in soup.find_all('h3')]
        if results:
            s = format("Found %n: \x02%L\x02, View more at: %u",
                       (len(results), 'result'), results, url)
            irc.reply(s)
        else:
            try:
                # Look for "too many results" errors and others reported by the
                # web interface.
                if distro == "debian":
                    errorParse = soup.find("div", class_="note").p
                else:
                    errorParse = soup.find("p", attrs={"id":
                                                       "psearchtoomanyhits"})
                if errorParse:
                    for br in errorParse.findAll('br'):
                        br.replace_with(" ")
                    e = errorParse.text.strip()
            except AttributeError:
                e = "No results found."
            irc.error(e)
    pkgsearch = wrap(pkgsearch, ['somethingWithoutSpaces',
                                 'somethingWithoutSpaces'])

    def linuxmint(self, irc, msg, args, release, query, opts):
        """<release> <package> [--exact]

        Looks up <package> in Linux Mint's repositories. If --exact is given,
        look up packages by the exact package name. Otherwise, look it up
        as a simple glob pattern."""
        addr = 'http://packages.linuxmint.com/list.php?release=' + \
            quote(release)
        try:
            fd = utils.web.getUrl(addr).decode("utf-8")
        except utils.web.Error as e:
            irc.error(str(e), Raise=True)
        soup = BeautifulSoup(fd)
        # Linux Mint puts their package lists in tables
        results = soup.find_all("td")
        found = OrderedDict()
        query = query.lower()
        exact = 'exact' in dict(opts)
        for result in results:
            name = result.contents[0].string  # Package name
            if query == name or (query in name and not exact):
                # This feels like really messy code, but we have to find tags
                # relative to our results.
                # Ascend to find the section name (in <h2>):
                section = result.parent.parent.parent.previous_sibling.\
                    previous_sibling.string
                # Find the package version in the next <td>; for some reason we
                # have to go two siblings further, as the first .next_sibling
                # returns '\n'. This is mentioned briefly in Beautiful Soup 4's
                # documentation...
                version = result.next_sibling.next_sibling.string
                # We format our found dictionary this way because the same
                # package can exist multiple times in different sections of
                # the repository (e.g. one in Main, one in Backports, etc.)
                found['%s [\x02%s\x02]' % (name, section)] = version
        if found:
            items = [format('%s \x02(%s)\x02', pkg, found[pkg]) for pkg in
                     found]
            s = format('Found %n: %L, %s %u', (len(found), 'result'), items,
                       _('View more at: '), addr)
            irc.reply(s)
        else:
            irc.error('No results found.')
    linuxmint = wrap(linuxmint, ['somethingWithoutSpaces',
                             'somethingWithoutSpaces',
                             getopts({'exact': ''})])

    @wrap(['positiveInt', 'somethingWithoutSpaces'])
    def fedora(self, irc, msg, args, release, query):
        """<release> <package name>

        Looks up <package> in Fedora's repositories. Globs (*, ?, etc.)
        are supported here. <release> is the release version (22, 21, etc.)"""
        url = 'https://admin.fedoraproject.org/pkgdb/api/packages/%s?' % quote(query)
        # Fedora uses f## in their API, where ## is the release version
        url += urlencode({'branches': 'f' + str(release), 'format': 'json'})
        self.log.debug("PkgInfo: using url %s for 'fedora' command", url)
        try:
            fd = utils.web.getUrl(url).decode("utf-8")
        except utils.web.Error as e:
            if '404' in str(e):
                e = 'No results found.'
                if '*' not in query:
                    e += " Try wrapping your query with *'s: '*%s*'" % query
            irc.error(e, Raise=True)
        data = json.loads(fd)
        def formatdesc(s):
            # Fedora's package descriptions have newlines inserted in them at strange positions,
            # sometimes even inside sentences. We'll break at the first sentence here:
            s = s.split('.')[0].strip()
            s = re.sub('\n+', ' ', s)
            return s
        results = ['%s: %s' % (ircutils.bold(pkg['name']), formatdesc(pkg['description']))
                   for pkg in data["packages"]]
        irc.reply('; '.join(results))

    @wrap(['positiveInt', 'somethingWithoutSpaces', 'somethingWithoutSpaces',
           getopts({'arch': 'somethingWithoutSpaces'})])
    def centos(self, irc, msg, args, release, repo, query, opts):
        """<release> <repository> <package name> [--arch <arch>]

        Looks up <package> in CentOS's repositories. <release> is the release
        version (6, 7, etc.), and <repository> is the repository name.
        You can find a list of possible repository names here:
        http://mirror.centos.org/centos/7/ (each folder is a repository).
        Supported values for <arch> include x86_64 and i386 (prior to CentOS 7)"""
        # CentOS doesn't have a package lookup interface, but only an autoindexed
        # file server...
        arch = dict(opts).get('arch') or 'x86_64'
        query = query.lower()
        # Different CentOS versions use different paths for their pool, ugh.
        for folder in ('Packages', 'RPMS', 'openstack-juno', 'openstack-kilo'):
            url = 'http://mirror.centos.org/centos/%s/%s/%s/%s/' % \
                (release, repo, arch, folder)
            self.log.debug("PkgInfo: trying url %s for 'centos' command", url)
            try:
                fd = utils.web.getUrl(url).decode("utf-8")
            except utils.web.Error:
                continue
            else:
                break
        else:
            irc.error('Unknown repository %r.' % repo, Raise=True)
        soup = BeautifulSoup(fd)
        # The first two tables are for the navigation bar; the third is the actual autoindex
        # content.
        res = []
        packagetable = soup.find_all('table')[2]
        exact = 'exact' in dict(opts)
        for tr in packagetable.find_all('tr'):
            try:
                package = tr.find_all('td')[1].a.text
            except IndexError:
                continue
            if not package.endswith('.rpm'):
                # Prevent things like "Parent Directory" from appearing in results
                continue
            package = ircutils.bold(package)
            if exact:
                if query == package.lower():
                    res.append(package)
            else:
                if query in package.lower():
                    res.append(package)
        if res:
            irc.reply(format('Available RPMs: %L', res))
        else:
            irc.error('No results found.')

Class = PkgInfo


# vim:set shiftwidth=4 softtabstop=4 expandtab textwidth=79: