From 868ad8eab888fd4938003ad97659b54662c27998 Mon Sep 17 00:00:00 2001 From: Ben Kurtovic Date: Sun, 14 Aug 2011 17:52:02 -0400 Subject: [PATCH] Reorganizing functions in config.py; store User-Agent and IRC version in config; exception, docstring, and comment cleanup. --- bot/blowfish.py | 37 ++++++----- bot/commands/ctcp.py | 6 +- bot/config.py | 173 ++++++++++++++++++++++++-------------------------- bot/main.py | 18 +++--- bot/wiki/constants.py | 12 ++-- bot/wiki/functions.py | 14 ++-- bot/wiki/site.py | 37 ++++++----- earwigbot.py | 4 +- 8 files changed, 154 insertions(+), 147 deletions(-) diff --git a/bot/blowfish.py b/bot/blowfish.py index b02bd84..d0862a9 100644 --- a/bot/blowfish.py +++ b/bot/blowfish.py @@ -442,8 +442,8 @@ class Blowfish(object): def encrypt(self, data): if not len(data) == 8: - raise BlockSizeError("blocks must be 8 bytes long, but tried to " + - "encrypt one {0} bytes long".format(len(data))) + e = "blocks must be 8 bytes long, but tried to encrypt one {0} bytes long" + raise BlockSizeError(e.format(len(data))) # Use big endianess since that's what everyone else uses xl = ord (data[3]) | (ord (data[2]) << 8) | (ord (data[1]) << 16) | (ord (data[0]) << 24) @@ -458,8 +458,8 @@ class Blowfish(object): def decrypt(self, data): if not len(data) == 8: - raise BlockSizeError("blocks must be 8 bytes long, but tried to " + - "decrypt one {0} bytes long".format(len(data))) + e = "blocks must be 8 bytes long, but tried to decrypt one {0} bytes long" + raise BlockSizeError(e.format(len(data))) # Use big endianess since that's what everyone else uses cl = ord (data[3]) | (ord (data[2]) << 8) | (ord (data[1]) << 16) | (ord (data[0]) << 24) @@ -482,16 +482,18 @@ class Blowfish(object): return 56 * 8 def verify_key(self, key): - """Make sure our key is not too short or too long; if there's a - problem, raise KeyTooShortError() or KeyTooLongError().""" + """Make sure our key is not too short or too long. + + If there's a problem, raise KeyTooShortError() or KeyTooLongError(). + """ if not key: raise KeyLengthError("no key given") if len(key) < 8: - raise KeyLengthError(("key is {0} bytes long, but it must be at " + - "least 8").format(len(key))) + e = "key is {0} bytes long, but it must be at least 8" + raise KeyLengthError(e.format(len(key))) if len(key) > 56: - raise KeyLengthError(("key is {0} bytes long, but it must be " + - "less than 56").format(len(key))) + e = "key is {0} bytes long, but it must be less than 56" + raise KeyLengthError(e.format(len(key))) def encrypt(key, plaintext): """Encrypt any length of plaintext using a given key that must be between @@ -518,24 +520,25 @@ def decrypt(key, cyphertext): try: cyphertext = cyphertext.decode("hex") - except TypeError as error: + except (TypeError, AttributeError) as error: e = error.message raise DecryptionError("cyphertext could not be decoded: " + e.lower()) if len(cyphertext) % 8 > 0: - raise DecryptionError("cyphertext cannot be broken into " + - "8-byte blocks evenly") + e = "cyphertext cannot be broken into 8-byte blocks evenly" + raise DecryptionError(e) blocks = [cyphertext[f:f+8] for f in range(0, len(cyphertext), 8)] msg = ''.join(map(cypher.decrypt, blocks)) - if not msg.startswith("TRUE"): # sanity check to ensure valid decryption - raise DecryptionError("the given key is incorrect, or part of the " + - "cyphertext is malformed") + # Sanity check to ensure valid decryption: + if not msg.startswith("TRUE"): + e = "the given key is incorrect, or part of the cyphertext is malformed" + raise DecryptionError(e) size, msg = msg[4:].split("|", 1) while len(msg) > int(size): - msg = msg[:-1] # remove the padding that we applied earlier + msg = msg[:-1] # Remove the padding that we applied earlier return msg diff --git a/bot/commands/ctcp.py b/bot/commands/ctcp.py index cc98441..3993279 100644 --- a/bot/commands/ctcp.py +++ b/bot/commands/ctcp.py @@ -4,6 +4,7 @@ import platform import time from classes import BaseCommand +import config class Command(BaseCommand): """Not an actual command, this module is used to respond to the CTCP @@ -40,6 +41,7 @@ class Command(BaseCommand): self.connection.notice(target, "\x01TIME {0}\x01".format(ts)) elif command == "VERSION": - vers = "EarwigBot - 0.1-dev - Python/{0} https://github.com/earwig/earwigbot" - vers = vers.format(platform.python_version()) + default = "EarwigBot - 0.1-dev - Python/$1 https://github.com/earwig/earwigbot" + vers = config.metadata.get("ircVersion", default) + vers = vers.replace("$1", platform.python_version()) self.connection.notice(target, "\x01VERSION {0}\x01".format(vers)) diff --git a/bot/config.py b/bot/config.py index 82969b6..8a99c26 100644 --- a/bot/config.py +++ b/bot/config.py @@ -8,13 +8,19 @@ including encrypting and decrypting passwords and making a new config file from scratch at the inital bot run. Usually you'll just want to do "from core import config" and access config data -from within config's three global variables and one function: - -* config.components - a list of enabled components -* config.wiki - a dict of config information for wiki-editing -* config.irc - a dict of config information for IRC -* config.schedule() - returns a list of tasks scheduled to run at a given - time +from within config's four global variables and one function: + +* config.components - a list of enabled components +* config.wiki - a dict of information about wiki-editing +* config.irc - a dict of information about IRC +* config.metadata - a dict of miscellaneous information +* config.schedule() - returns a list of tasks scheduled to run at a given time + +Additionally, functions related to config loading: +* config.load() - loads and parses our config file, returning True if + passwords are stored encrypted or False otherwise +* config.decrypt() - given a key, decrypts passwords inside our config + variables; won't work if passwords aren't encrypted """ import json @@ -26,19 +32,12 @@ script_dir = path.dirname(path.abspath(__file__)) root_dir = path.split(script_dir)[0] config_path = path.join(root_dir, "config.json") -_config = None # holds data loaded from our config file - -# set our three easy-config-access global variables to None -components, wiki, irc = (None, None, None) +_config = None # Holds data loaded from our config file -def is_config_loaded(): - """Return True if our config file has already been loaded, and False if it - hasn't.""" - if _config is not None: - return True - return False +# Set our four easy-config-access global variables to None +components, wiki, irc, metadata = None, None, None, None -def load_config(): +def _load(): """Load data from our JSON config file (config.json) into _config.""" global _config with open(config_path, 'r') as fp: @@ -49,86 +48,92 @@ def load_config(): print error exit(1) -def verify_config(): - """Check to see if we have a valid config file, and if not, notify the - user. If there is no config file at all, offer to make one; otherwise, - exit. If everything goes well, return True if stored passwords are - encrypted in the file, or False if they are not.""" - if path.exists(config_path): - load_config() - try: - return _config["encryptPasswords"] # are passwords encrypted? - except KeyError: - return False # assume passwords are not encrypted by default +def _make_new(): + """Make a new config file based on the user's input.""" + encrypt = raw_input("Would you like to encrypt passwords stored in config.json? [y/n] ") + if encrypt.lower().startswith("y"): + is_encrypted = True else: + is_encrypted = False + + return is_encrypted + +def is_loaded(): + """Return True if our config file has been loaded, otherwise False.""" + return _config is not None + +def load(): + """Load, or reload, our config file. + + First, check if we have a valid config file, and if not, notify the user. + If there is no config file at all, offer to make one, otherwise exit. + + Store data from our config file in four global variables (components, wiki, + irc, metadata) for easy access (as well as the internal _config variable). + + If everything goes well, return True if stored passwords are + encrypted in the file, or False if they are not. + """ + global components, wiki, irc, metadata + + if not path.exists(config_path): print "You haven't configured the bot yet!" choice = raw_input("Would you like to do this now? [y/n] ") if choice.lower().startswith("y"): - return make_new_config() + return _make_new() else: exit(1) -def parse_config(key): - """Store data from our config file in three global variables for easy - access, and use the key to unencrypt passwords. Catch password decryption - errors and report them to the user.""" - global components, wiki, irc + _load() - load_config() # we might be re-loading unnecessarily here, but no harm in - # that! - try: - components = _config["components"] - except KeyError: - components = [] - try: - wiki = _config["wiki"] - except KeyError: - wiki = {} - try: - irc = _config["irc"] - except KeyError: - irc = {} + components = _config.get("components", []) + wiki = _config.get("wiki", {}) + irc = _config.get("irc", {}) + metadata = _config.get("metadata", {}) + + # Are passwords encrypted? + return metadata.get("encryptPasswords", False) + +def decrypt(key): + """Use the key to decrypt passwords in our config file. + + Call this if load() returns True. Catch password decryption errors and + report them to the user. + """ + global irc, wiki try: - try: - if _config["encryptPasswords"]: - decrypt(key, "wiki['password']") - decrypt(key, "irc['frontend']['nickservPassword']") - decrypt(key, "irc['watcher']['nickservPassword']") - except KeyError: - pass + item = wiki.get("password") + if item: + wiki["password"] = blowfish.decrypt(key, item) + + item = irc.get("frontend").get("nickservPassword") + if item: + irc["frontend"]["nickservPassword"] = blowfish.decrypt(key, item) + + item = irc.get("watcher").get("nickservPassword") + if item: + irc["watcher"]["nickservPassword"] = blowfish.decrypt(key, item) + except blowfish.BlowfishError as error: print "\nError decrypting passwords:" print "{0}: {1}.".format(error.__class__.__name__, error) exit(1) -def decrypt(key, item): - """Decrypt 'item' with blowfish.decrypt() using the given key and set it to - the decrypted result. 'item' should be a string, like - decrypt(key, "wiki['password']"), NOT decrypt(key, wiki['password'), - because that won't work.""" - global irc, wiki - try: - result = blowfish.decrypt(key, eval(item)) - except KeyError: - return - exec "{0} = result".format(item) - def schedule(minute, hour, month_day, month, week_day): - """Return a list of tasks that are scheduled to run at the time specified - by the function arguments. The schedule data comes from our config file's - 'schedule' field, which is stored as _config["schedule"]. Call this - function with config.schedule(args).""" - tasks = [] # tasks to run this turn, each as a tuple of either (task_name, - # kwargs), or just task_name + """Return a list of tasks scheduled to run at the specified time. + + The schedule data comes from our config file's 'schedule' field, which is + stored as _config["schedule"]. Call this function as config.schedule(args). + """ + # Tasks to run this turn, each as a list of either [task_name, kwargs], or + # just the task_name: + tasks = [] now = {"minute": minute, "hour": hour, "month_day": month_day, "month": month, "week_day": week_day} - try: - data = _config["schedule"] - except KeyError: - return [] # nothing is in our schedule + data = _config.get("schedule", []) for event in data: do = True for key, value in now.items(): @@ -146,15 +151,3 @@ def schedule(minute, hour, month_day, month, week_day): pass return tasks - -def make_new_config(): - """Make a new config file based on the user's input.""" - - encrypt = raw_input("Would you like to encrypt passwords stored in " + - "config.json? [y/n] ") - if encrypt.lower().startswith("y"): - is_encrypted = True - else: - is_encrypted = False - - return is_encrypted diff --git a/bot/main.py b/bot/main.py index f0ae18b..0e51d67 100644 --- a/bot/main.py +++ b/bot/main.py @@ -4,9 +4,9 @@ """ EarwigBot's Core -This (should) not be run directly; the wrapper in "earwigbot.py" is preferred, +This should not be run directly; the wrapper in "earwigbot.py" is preferred, but it should work fine alone, as long as you enter the password-unlock key at -the initial hidden prompt. +the initial hidden prompt if one is needed. The core is essentially responsible for starting the various bot components (irc, scheduler, etc) and making sure they are all happy. An explanation of the @@ -103,12 +103,14 @@ def irc_frontend(): f_conn.close() def run(): + config.load() try: - key = raw_input() # wait for our password unlock key from the bot's - except EOFError: # wrapper - key = None - config.parse_config(key) # load data from the config file and parse it - # using the unlock key + key = raw_input() # wait for our password decrypt key from the bot's + except EOFError: # wrapper, then decrypt passwords + pass + else: + config.decrypt(key) + enabled = config.components if "irc_frontend" in enabled: # make the frontend run on our primary @@ -120,7 +122,7 @@ def run(): tasks.load() # watcher on another thread iff it if "irc_watcher" in enabled: # is enabled print "\nStarting IRC watcher..." - t_watcher = threading.Thread(target=irc_watcher, args=()) + t_watcher = threading.Thread(target=irc_watcher) t_watcher.name = "irc-watcher" t_watcher.daemon = True t_watcher.start() diff --git a/bot/wiki/constants.py b/bot/wiki/constants.py index 62884b5..5705cc2 100644 --- a/bot/wiki/constants.py +++ b/bot/wiki/constants.py @@ -3,18 +3,18 @@ """ EarwigBot's Wiki Toolset: Constants -This module defines some useful constants, such as default namespace IDs for -easy lookup and our user agent. +This module defines some useful constants: +* USER_AGENT - our default User Agent when making API queries +* NS_* - default namespace IDs for easy lookup -Import with `from wiki.constants import *`. +Import with `from wiki import constants` or `from wiki.constants import *`. """ +# Default User Agent when making API queries: import platform - -# User agent when making API queries USER_AGENT = "EarwigBot/0.1-dev (Python/{0}; https://github.com/earwig/earwigbot)".format(platform.python_version()) -# Default namespace IDs +# Default namespace IDs: NS_MAIN = 0 NS_TALK = 1 NS_USER = 2 diff --git a/bot/wiki/functions.py b/bot/wiki/functions.py index 6fce5a1..a9c611b 100644 --- a/bot/wiki/functions.py +++ b/bot/wiki/functions.py @@ -14,6 +14,7 @@ from cookielib import LWPCookieJar, LoadError import errno from getpass import getpass from os import chmod, path +import platform import stat import config @@ -30,12 +31,10 @@ def _load_config(): directly from Python's interpreter and not the bot itself, because earwigbot.py or core/main.py will already call these functions. """ - is_encrypted = config.verify_config() + is_encrypted = config.load() if is_encrypted: # passwords in the config file are encrypted key = getpass("Enter key to unencrypt bot passwords: ") - config.parse_config(key) - else: - config.parse_config(None) + config.decrypt(key) def _get_cookiejar(): """Returns a LWPCookieJar object loaded from our .cookies file. The same @@ -87,6 +86,10 @@ def _get_site_object_from_dict(name, d): namespaces = d.get("namespaces", {}) login = (config.wiki.get("username"), config.wiki.get("password")) cookiejar = _get_cookiejar() + user_agent = config.metadata.get("userAgent") + + if user_agent: + user_agent = user_agent.replace("$1", platform.python_version()) for key, value in namespaces.items(): # Convert string keys to integers del namespaces[key] @@ -98,7 +101,8 @@ def _get_site_object_from_dict(name, d): return Site(name=name, project=project, lang=lang, base_url=base_url, article_path=article_path, script_path=script_path, sql=sql, - namespaces=namespaces, login=login, cookiejar=cookiejar) + namespaces=namespaces, login=login, cookiejar=cookiejar, + user_agent=user_agent) def get_site(name=None, project=None, lang=None): """Returns a Site instance based on information from our config file. diff --git a/bot/wiki/site.py b/bot/wiki/site.py index f5d77e5..eecf2c3 100644 --- a/bot/wiki/site.py +++ b/bot/wiki/site.py @@ -40,7 +40,8 @@ class Site(object): def __init__(self, name=None, project=None, lang=None, base_url=None, article_path=None, script_path=None, sql=(None, None), - namespaces=None, login=(None, None), cookiejar=None): + namespaces=None, login=(None, None), cookiejar=None, + user_agent=None): """Constructor for new Site instances. This probably isn't necessary to call yourself unless you're building a @@ -57,8 +58,8 @@ class Site(object): the API, and then log in if a username/pass was given and we aren't already logged in. """ - # attributes referring to site information, filled in by an API query - # if they are missing (and an API url can be determined) + # Attributes referring to site information, filled in by an API query + # if they are missing (and an API url can be determined): self._name = name self._project = project self._lang = lang @@ -68,19 +69,21 @@ class Site(object): self._sql = sql self._namespaces = namespaces - # set up cookiejar and URL opener for making API queries + # Set up cookiejar and URL opener for making API queries: if cookiejar is not None: self._cookiejar = cookiejar else: self._cookiejar = CookieJar() + if user_agent is None: + user_agent = USER_AGENT # Set default UA from wiki.constants self._opener = build_opener(HTTPCookieProcessor(self._cookiejar)) - self._opener.addheaders = [("User-Agent", USER_AGENT), + self._opener.addheaders = [("User-Agent", user_agent), ("Accept-Encoding", "gzip")] - # get all of the above attributes that were not specified as arguments + # Get all of the above attributes that were not specified as arguments: self._load_attributes() - # if we have a name/pass and the API says we're not logged in, log in + # If we have a name/pass and the API says we're not logged in, log in: self._login_info = name, password = login if name is not None and password is not None: logged_in_as = self._get_username_from_cookies() @@ -112,7 +115,7 @@ class Site(object): raise SiteAPIError(e) url = ''.join((self._base_url, self._script_path, "/api.php")) - params["format"] = "json" # this is the only format we understand + params["format"] = "json" # This is the only format we understand data = urlencode(params) print url, data # debug code @@ -135,7 +138,7 @@ class Site(object): stream = StringIO(result) gzipper = GzipFile(fileobj=stream) result = gzipper.read() - return loads(result) # parse as a JSON object + return loads(result) # Parse as a JSON object def _load_attributes(self, force=False): """Load data about our Site from the API. @@ -147,8 +150,8 @@ class Site(object): Additionally, you can call this with `force=True` to forcibly reload all attributes. """ - # all attributes to be loaded, except _namespaces, which is a special - # case because it requires additional params in the API query + # All attributes to be loaded, except _namespaces, which is a special + # case because it requires additional params in the API query: attrs = [self._name, self._project, self._lang, self._base_url, self._article_path, self._script_path] @@ -158,9 +161,9 @@ class Site(object): params["siprop"] = "general|namespaces|namespacealiases" result = self._api_query(params) self._load_namespaces(result) - elif all(attrs): # everything is already specified and we're not told + elif all(attrs): # Everything is already specified and we're not told return # to force a reload, so do nothing - else: # we're only loading attributes other than _namespaces + else: # We're only loading attributes other than _namespaces params["siprop"] = "general" result = self._api_query(params) @@ -240,9 +243,9 @@ class Site(object): continue if cookie.name != name: continue - # build a regex that will match domains this cookie affects + # Build a regex that will match domains this cookie affects: search = ''.join(("(.*?)", re_escape(cookie.domain))) - if re_match(search, domain): # test it against our site + if re_match(search, domain): # Test it against our site user_name = self._get_cookie("centralauth_User", cookie.domain) if user_name is not None: return user_name.value @@ -402,7 +405,7 @@ class Site(object): """ lname = name.lower() for ns_id, names in self._namespaces.items(): - lnames = [n.lower() for n in names] # be case-insensitive + lnames = [n.lower() for n in names] # Be case-insensitive if lname in lnames: return ns_id @@ -421,7 +424,7 @@ class Site(object): """ prefixes = self.namespace_id_to_name(NS_CATEGORY, all=True) prefix = title.split(":", 1)[0] - if prefix != title: # avoid a page that is simply "Category" + if prefix != title: # Avoid a page that is simply "Category" if prefix in prefixes: return Category(self, title, follow_redirects) return Page(self, title, follow_redirects) diff --git a/earwigbot.py b/earwigbot.py index bf6d9aa..0fcb314 100755 --- a/earwigbot.py +++ b/earwigbot.py @@ -33,7 +33,7 @@ bot_script = path.join(path.dirname(path.abspath(__file__)), "bot", "main.py") def main(): print "EarwigBot v{0}\n".format(__version__) - is_encrypted = config.verify_config() + is_encrypted = config.load() if is_encrypted: # passwords in the config file are encrypted key = getpass("Enter key to unencrypt bot passwords: ") else: @@ -41,7 +41,7 @@ def main(): while 1: bot = Popen([executable, bot_script], stdin=PIPE) - bot.communicate(key) # give the key to core.config.load_config() + bot.communicate(key) # give the key to core.config.decrypt() return_code = bot.wait() if return_code == 1: exit() # let critical exceptions in the subprocess cause us to