#!/usr/bin/python
# encoding: utf-8

import base64
import sys
import os
import argparse
from collections import namedtuple
from subprocess import Popen, PIPE

from urllib import urlencode
from urllib2 import Request, urlopen
import json
from urlparse import urlparse

WHITELIST=[]
BLACKLIST=[]

LISTINGS_PER_PAGE = 100
LISTINGS_PER_PAGE_PARAM = 'per_page'
LISTING_PAGE_PARAM = 'page'
GITHUB_API_HOST = 'https://api.github.com'

GIT_CLONE_CMD = 'git clone %s %s %s'
GIT_CLONE_API_URL = 'https://%s@github.com/%s'
GIT_SHA_CMD = 'git rev-parse --short %s'
GIT_FETCH_CMD = 'git fetch --tags'
GIT_CHECK_REMOTE_CMD = 'git ls-remote'

USER_DETAILS_PATH = '/users/%s'
USER_ORG_DETAILS_PATH = '/user/orgs'

DEFAULT_TOKEN_FILE = os.path.expanduser('~/.config/ghtoken')

class Color:
    GREEN = "\033[1;32m"
    BLUE = "\033[1;34m"
    YELLOW = "\033[1;33m"
    RED = "\033[1;31m"
    END = "\033[0m"

def get_color_str(text, color):
    return color + str(text) + Color.END

def system_exec(command, directory=None, show_output=True, ignore_error=False):
    if not directory:
        directory = os.getcwd()

    try:
        process = Popen(command, stdout=PIPE, stderr=PIPE, shell=True, cwd=directory)
        (output, error) = process.communicate()
        output = output.strip()
        error = error.strip()

        if show_output and len(output) > 0:
            print output
            sys.stdout.flush()

        if process.returncode != 0 and not ignore_error:
            raise Exception(error)

        ReturnInfo = namedtuple('ReturnInfo', 'return_code output error')
        return ReturnInfo(process.returncode, output, error)

    except Exception as err:
        print >>sys.stderr, Color.RED + "Could not execute", command
        print >>sys.stderr, err, Color.END
        print >>sys.stderr, "Terminating early"
        exit(1)

class AttributeDict(dict):
    def __getattr__(self, attr):
        return self[attr]
    def __setattr__(self, attr, value):
        self[attr] = value

def read_api_uri(uri, config, page=1):
    params = {
        LISTINGS_PER_PAGE_PARAM: LISTINGS_PER_PAGE,
        LISTING_PAGE_PARAM: page,
    }
    uri += '?' + urlencode(params)

    creds = base64.b64encode(config.token.encode('utf-8')).decode('ascii')
    basic_auth = 'Basic {}'.format(creds)
    headers = {
        'Authorization': basic_auth,
    }

    if config.debug:
        print "Trying URI {} with headers {}".format(uri, headers)

    req = Request(uri, headers=headers)
    return urlopen(req).read()

def get_json(uri, config, obj_type=AttributeDict, page=1):
    return json.loads(read_api_uri(uri, config, page), object_hook=obj_type)

class GitHubRepo(AttributeDict):
    """Top-level class managing all content of a GitHub repository"""
    def __init__(self, *args, **kwargs):
        AttributeDict.__init__(self, *args, **kwargs)
        self.content = []

    def configure(self, config):
        self.config = config

        # Initialize the content repos
        self.content.append(CodeRepo(self, config))
        if self.config.full_backup:
            self.content.append(WikiRepo(self, config))
            if self.has_issues:
                self.content.append(IssuesRepo(self, config))
                self.content.append(MilestonesRepo(self, config))
            self.content.append(PullsRepo(self, config))
            # Teams are only available from user's repos
            if self.owner.login == self.config.username:
                self.content.append(TeamsRepo(self, config))
            self.content.append(CommentsRepo(self, config))
            self.content.append(ForksRepo(self, config))

    def update(self):
        [repo.update() for repo in self.content]

class CodeRepo(object):
    """GitHub code git repository"""
    def __init__(self, gh_repo, config):
        self.config = config

        # Copy the details we need from the JSON
        self.name = gh_repo.name
        self.full_name = gh_repo.full_name
        self.ssh_url = gh_repo.ssh_url
        self.default_branch = gh_repo.default_branch
        self.description = gh_repo.description or ""

        self.target_directory = os.path.join(self.config.cwd, self.name)
        if self.config.mirror:
            self.target_directory += '.git'

    def try_clone(self, ignore_error=False):
        print "Using", self.target_directory
        if os.path.isdir(self.target_directory):
            return False

        print "Need to clone", self.name
        if self.config.mirror:
            clone_opts = '--mirror'
        else:
            clone_opts = '--recurse-submodules'
        if self.config.ssh:
            clone_url = self.ssh_url
        else:
            clone_url = GIT_CLONE_API_URL % (self.config.token,
                                             self.full_name)
        clone_cmd = GIT_CLONE_CMD % (clone_opts, clone_url, \
                                     self.target_directory)

        # Create leading directories if necessary
        clone_dir = os.path.dirname(self.target_directory)
        if not os.path.isdir(clone_dir):
            os.makedirs(clone_dir)

        # Let the caller decide if errors should be ignored.
        return_code = system_exec(clone_cmd, ignore_error=ignore_error).return_code
        if ignore_error and return_code != 0:
            print "Repo for %s not initialized, skipping" % self.name
            return True

        print "Finished cloning"
        return True

    def get_sha_str(self, branch, directory):
        if self.config.mirror:
            sha_cmd = GIT_SHA_CMD % branch
        else:
            sha_cmd = GIT_SHA_CMD % 'origin/' + branch
        sha = system_exec(sha_cmd, directory, False, True).output
        return get_color_str(sha, Color.GREEN)

    def print_start_sha(self, branch, directory):
        print("- " + get_color_str(branch, Color.GREEN) + " @ " \
                + self.get_sha_str(branch, directory) + ' ..'),

    def update(self, fatal_remote_errors = True):
        if self.try_clone(False):
            return

        # Check if the remote is still there before fetching
        return_code = system_exec(GIT_CHECK_REMOTE_CMD, self.target_directory, show_output=False, ignore_error=True).return_code
        if return_code != 0:
            # Is it fatal?
            if fatal_remote_errors:
                error_message = get_color_str("ERROR: Repo for %s seems to have been deleted. Skipping update." % self.full_name, Color.RED)
                print error_message
                raise Exception(error_message)
            else:
                print get_color_str("WARNING: Repo for %s seems to have been deleted. Skipping update." % self.full_name, Color.GREEN)
                return

        self.print_start_sha(self.default_branch, self.target_directory)
        system_exec(GIT_FETCH_CMD, self.target_directory, False)
        print self.get_sha_str(self.default_branch, self.target_directory)

class WikiRepo(CodeRepo):
    """GitHub wiki git repository"""
    def __init__(self, gh_repo, config):
        CodeRepo.__init__(self, gh_repo, config)

        # Put the wiki under the project.extras/wiki directory
        self.target_directory = os.path.join(self.config.cwd,
                                             self.name + '.extras',
                                             'wiki')
        if self.config.mirror:
            self.target_directory += '.git'

        # Adjust the name and hardcode master as the default branch
        self.name += '.wiki'
        self.full_name += '.wiki'
        self.ssh_url = self.ssh_url.rpartition('.git')[0] \
                       + '.wiki.git'
        self.default_branch = 'master'

        if self.description == None:
            self.description = gh_repo.name

        self.description += ' - Wiki'

    def try_clone(self, ignore_error=True):
        # This is sad. Github will tell us a repo has a wiki, but if the
        # wiki hasn't actually been initialized, it won't exist. Reuse the
        # GitRepo clone code, but allow it to fail.
        return CodeRepo.try_clone(self, True)

    def update(self):
        super(WikiRepo, self).update(fatal_remote_errors = False)

class JsonRepo(object):
    def __init__(self, gh_repo, api_url, config, content=None):
        self.config = config

        # Copy the details we need from the JSON
        self.name = gh_repo.name
        self.full_name = gh_repo.full_name
        self.repo_url = gh_repo.url
        self.description = gh_repo.description or ""

        # Sanitize the content url to strip the {/number} type markers
        self.content_url = api_url.split('{')[0]

        # Put downloaded content in project.extras directory
        self.content_directory = os.path.join(self.config.cwd,
                                              self.name + '.extras')

        # If the content type was supplied, adjust the names further
        if content is not None:
            self.name += '.%s' % content
            self.full_name += '.%s' % content
            self.description += ' - %s' % content.title()
            self.content_directory = os.path.join(self.content_directory,
                                                  content)

    def download_file(self, url, suffix='.json'):
        if url == self.content_url:
            # Just use the content directory
            target = self.content_directory
        else:
            # Strip off base content URL path to get target
            base = os.path.relpath(urlparse(url).path,
                                   urlparse(self.content_url).path)
            target = os.path.join(self.content_directory, base)

        # Add suffix if specified
        if suffix:
            target += suffix

        target_directory = os.path.dirname(target)
        if not os.path.isdir(target_directory):
            os.makedirs(target_directory)

        if self.config.debug:
            print "Downloading %s" % target
        data = read_api_uri(url, self.config)
        with open(target, 'wb') as dl_file:
            dl_file.write(data)

class IssuesRepo(JsonRepo):
    def __init__(self, gh_repo, config):
        JsonRepo.__init__(self, gh_repo, gh_repo.issues_url, config, 'issues')

    def update(self):
        print "Using", self.content_directory
        for issue in get_json(self.content_url, self.config):
            self.download_file(issue.url)
            self.download_file(issue.comments_url)
            self.download_file(issue.events_url)

class PullsRepo(JsonRepo):
    def __init__(self, gh_repo, config):
        JsonRepo.__init__(self, gh_repo, gh_repo.pulls_url, config, 'pulls')

    def update(self):
        print "Using", self.content_directory
        for pull in get_json(self.content_url, self.config):
            self.download_file(pull.url)
            self.download_file(pull.commits_url)
            self.download_file(pull.comments_url)

class MilestonesRepo(JsonRepo):
    def __init__(self, gh_repo, config):
        JsonRepo.__init__(self, gh_repo, gh_repo.milestones_url, config,
                          'milestones')

    def update(self):
        print "Using", self.content_directory
        for milestone in get_json(self.content_url, self.config):
            self.download_file(milestone.url)

class TeamsRepo(JsonRepo):
    def __init__(self, gh_repo, config):
        JsonRepo.__init__(self, gh_repo, gh_repo.teams_url, config, 'teams')

    def update(self):
        print "Using", self.content_directory + '.json'
        self.download_file(self.content_url)

class CommentsRepo(JsonRepo):
    def __init__(self, gh_repo, config):
        JsonRepo.__init__(self, gh_repo, gh_repo.comments_url, config,
                          'comments')

    def update(self):
        print "Using", self.content_directory + '.json'
        self.download_file(self.content_url)

class ForksRepo(JsonRepo):
    def __init__(self, gh_repo, config):
        JsonRepo.__init__(self, gh_repo, gh_repo.forks_url, config, 'forks')

    def update(self):
        print "Using", self.content_directory + '.json'
        self.download_file(self.content_url)

class RepoUpdater(object):
    def __init__(self, args):
        # Keep a copy of the arguments dictionary
        self.args = args

        # Use current directory if unspecified
        self.args.cwd = args.directory if args.directory else os.getcwd()

        # Read a token file if necessary
        if self.args.token is None:
            try:
                with open(self.args.token_file, 'r') as tfile:
                    self.args.token = tfile.read().strip()
            except IOError as err:
                print >>sys.stderr, Color.RED + \
                    "Could not open token file %s: %s" \
                    % (self.args.token_file, err), \
                    Color.END
                print >>sys.stderr, "Terminating early"
                exit(1)

        if self.args.debug:
            print "Current configuration: "
            for arg in self.args:
                # We don't want this printed
                if arg != 'token':
                    print arg, ":", get_color_str(args[arg], Color.GREEN)

    def update(self):
        if not os.path.isdir(self.args.cwd):
            os.makedirs(self.args.cwd)

        print "User:", get_color_str(self.args.username, Color.GREEN)
        user_data = self.get_user_data()

        repos, excluded_repos = self.get_repos(user_data.repos_url, GITHUB_API_HOST + USER_ORG_DETAILS_PATH)
        repos = self.filter_repo_names(repos, excluded_repos)
        for repo in repos:
            print get_color_str('{:-^60}'.format(repo.name), Color.YELLOW)
            repo.update()

    def get_user_data(self):
        return get_json(GITHUB_API_HOST + USER_DETAILS_PATH
                        % self.args.username, self.args)

    def get_own_repos(self, repos_url):
        repos = self._get_paged_repos(repos_url, self.args, GitHubRepo)

        [repo.configure(self.args) for repo in repos]
        repo_count = len(repos)

        if self.args.exclude_forks:
            repos = [repo for repo in repos if not repo.fork]

        return repos, repo_count

    def get_org_repos(self, orgs_url):
        all_orgs_repos = []
        repo_count = 0

        orgs = get_json(orgs_url, self.args)
        for org in orgs:
            org_repos = self._get_paged_repos(org.repos_url, self.args, GitHubRepo)

            [repo.configure(self.args) for repo in org_repos]
            repo_count += len(org_repos)

            if not self.args.include_public_org_repos:
                org_repos = [repo for repo in org_repos if repo.private]
            if not self.args.include_org_forks:
                org_repos = [repo for repo in org_repos if not repo.fork]

            all_orgs_repos += org_repos

        return all_orgs_repos, repo_count

    def _get_paged_repos(self, url, args, clazz):
        repos_page = get_json(url, args, clazz)
        repos = repos_page

        if self.args.debug:
            print "Retrieved %d entries from %s" % (len(repos), url )

        page = 2
        while len(repos_page) >= LISTINGS_PER_PAGE:
            repos_page = get_json(url, args, clazz, page)

            if self.args.debug:
                print "Retrieved %d entries from %s" % (len(repos_page), url)

            repos += repos_page
            page += 1

        return repos

    def get_repos(self, repos_url, orgs_url):
        repos = []
        if self.args.debug:
            print "Getting repo data from", get_color_str(repos_url, Color.GREEN)

        own_repos, repo_count = self.get_own_repos(repos_url)
        if not self.args.exclude_own:
            repos += own_repos

        org_repos, org_repo_count = self.get_org_repos(orgs_url)
        if not self.args.exclude_orgs:
            repos += org_repos

        if self.args.debug:
            print "Available repos:", get_color_str(str(len(repos)), Color.GREEN)
            for repo in repos:
                owner = repo.owner.login
                print " -", get_color_str(repo.name, Color.YELLOW)
                print " " * 5, repo.description

        excluded_repos = repo_count + org_repo_count - len(repos)
        return repos, excluded_repos

    def filter_repo_names(self, repos, excluded_repos):
        original_repos = len(repos)

        if BLACKLIST:
            repos = [repo for repo in repos if not repo.name in BLACKLIST]

        if WHITELIST:
            repos = [repo for repo in repos if repo.name in WHITELIST]

        filtered_repos = original_repos - len(repos)

        ignored_repos_str = " (" + str(filtered_repos) + " filtered, " + str(excluded_repos) + " excluded)"
        repo_count_str = str(original_repos - filtered_repos) + " / " + str(original_repos)
        print "Fetching repos:", get_color_str(repo_count_str + ignored_repos_str, Color.GREEN)

        for repo in repos:
            owner = repo.owner.login
            print " -", Color.YELLOW + repo.name, Color.END
            print " " * 5, repo.description

        return repos

if __name__ == '__main__':
    parser = argparse.ArgumentParser(description='Retrieve and/or update local copies of GitHub-hosted repos')

    parser.add_argument('username', \
            help='GitHub username that will be used for cloning and fetching')
    parser.add_argument('token', \
            nargs='?', \
            help='GitHub auth token for that username. \
            You can create one at https://github.com/settings/applications')
    parser.add_argument('--version', \
            action='version', \
            version='%(prog)s v0.7')

    parser.add_argument('-d', '--directory', \
            help='Target directory for cloning and fetching')

    parser.add_argument('-t', '--token-file', \
            default=DEFAULT_TOKEN_FILE, \
            help='File containing the github token')
    parser.add_argument('-s', '--ssh', \
            help='Fetch repositories using ssh', \
            action='store_true')
    parser.add_argument('-m', '--mirror', \
            help='Mirror bare repositories instead of making full checkouts', \
            action='store_true')
    parser.add_argument('-x', '--exclude-own', \
            help='Exclude own repositories in the updates', \
            action='store_true')
    parser.add_argument('--exclude-forks', \
            help='Exclude forked repositories from the updates', \
            action='store_true')
    parser.add_argument('--exclude-orgs', \
            help='Exclude repos that are in user\'s orgs (this does not filter the ones you have)', \
            action='store_true')
    parser.add_argument('--include-org-forks', \
            help='Include forked repos that are in user\'s orgs', \
            action='store_true')
    parser.add_argument('--include-public-org-repos', \
            help='Include public repos that are in user\'s orgs', \
            action='store_true')
    parser.add_argument('--full-backup', \
            help='Include all repository content', \
            action='store_true')

    parser.add_argument('--debug', \
            help='Enable debugging output', \
            action='store_true')

    args = AttributeDict(vars(parser.parse_args()))

    RepoUpdater(args).update()
