Backend support for partial history update
[gc-dialer] / src / backends / gvoice / gvoice.py
index 582c6f4..b0825ef 100755 (executable)
@@ -87,7 +87,7 @@ class Message(object):
                return "%s (%s): %s" % (
                        self.whoFrom,
                        self.when,
-                       "".join(str(part) for part in self.body)
+                       "".join(unicode(part) for part in self.body)
                )
 
        def to_dict(self):
@@ -170,7 +170,6 @@ class GVoiceBackend(object):
 
                SECURE_URL_BASE = "https://www.google.com/voice/"
                SECURE_MOBILE_URL_BASE = SECURE_URL_BASE + "mobile/"
-               self._forwardURL = SECURE_MOBILE_URL_BASE + "phones"
                self._tokenURL = SECURE_URL_BASE + "m"
                self._callUrl = SECURE_URL_BASE + "call/connect"
                self._callCancelURL = SECURE_URL_BASE + "call/cancel"
@@ -188,6 +187,8 @@ class GVoiceBackend(object):
                self._XML_ACCOUNT_URL = SECURE_URL_BASE + "contacts/"
                # HACK really this redirects to the main pge and we are grabbing some javascript
                self._XML_CONTACTS_URL = "http://www.google.com/voice/inbox/search/contact"
+               self._CSV_CONTACTS_URL = "http://mail.google.com/mail/contacts/data/export"
+               self._JSON_CONTACTS_URL = SECURE_URL_BASE + "b/0/request/user"
                self._XML_RECENT_URL = SECURE_URL_BASE + "inbox/recent/"
 
                self.XML_FEEDS = (
@@ -201,17 +202,15 @@ class GVoiceBackend(object):
                self._XML_TRASH_URL = SECURE_URL_BASE + "inbox/recent/trash"
                self._XML_VOICEMAIL_URL = SECURE_URL_BASE + "inbox/recent/voicemail/"
                self._XML_SMS_URL = SECURE_URL_BASE + "inbox/recent/sms/"
+               self._JSON_SMS_URL = SECURE_URL_BASE + "b/0/request/messages/"
+               self._JSON_SMS_COUNT_URL = SECURE_URL_BASE + "b/0/request/unread"
                self._XML_RECORDED_URL = SECURE_URL_BASE + "inbox/recent/recorded/"
                self._XML_PLACED_URL = SECURE_URL_BASE + "inbox/recent/placed/"
                self._XML_RECEIVED_URL = SECURE_URL_BASE + "inbox/recent/received/"
                self._XML_MISSED_URL = SECURE_URL_BASE + "inbox/recent/missed/"
 
                self._galxRe = re.compile(r"""<input.*?name="GALX".*?value="(.*?)".*?/>""", re.MULTILINE | re.DOTALL)
-               self._tokenRe = re.compile(r"""<input.*?name="_rnr_se".*?value="(.*?)"\s*/>""")
-               self._accountNumRe = re.compile(r"""<b class="ms\d">(.{14})</b></div>""")
-               self._callbackRe = re.compile(r"""\s+(.*?):\s*(.*?)<br\s*/>\s*$""", re.M)
 
-               self._contactsBodyRe = re.compile(r"""gcData\s*=\s*({.*?});""", re.MULTILINE | re.DOTALL)
                self._seperateVoicemailsRegex = re.compile(r"""^\s*<div id="(\w+)"\s* class=".*?gc-message.*?">""", re.MULTILINE | re.DOTALL)
                self._exactVoicemailTimeRegex = re.compile(r"""<span class="gc-message-time">(.*?)</span>""", re.MULTILINE)
                self._relativeVoicemailTimeRegex = re.compile(r"""<span class="gc-message-relative">(.*?)</span>""", re.MULTILINE)
@@ -227,31 +226,21 @@ class GVoiceBackend(object):
 
        def is_quick_login_possible(self):
                """
-               @returns True then is_authed might be enough to login, else full login is required
+               @returns True then refresh_account_info might be enough to login, else full login is required
                """
                return self._loadedFromCookies or 0.0 < self._lastAuthed
 
-       def is_authed(self, force = False):
-               """
-               Attempts to detect a current session
-               @note Once logged in try not to reauth more than once a minute.
-               @returns If authenticated
-               """
-               isRecentledAuthed = (time.time() - self._lastAuthed) < 120
-               isPreviouslyAuthed = self._token is not None
-               if isRecentledAuthed and isPreviouslyAuthed and not force:
-                       return True
-
+       def refresh_account_info(self):
                try:
-                       page = self._get_page(self._forwardURL)
-                       self._grab_account_info(page)
+                       page = self._get_page(self._JSON_CONTACTS_URL)
+                       accountData = self._grab_account_info(page)
                except Exception, e:
                        _moduleLogger.exception(str(e))
-                       return False
+                       return None
 
                self._browser.save_cookies()
                self._lastAuthed = time.time()
-               return True
+               return accountData
 
        def _get_token(self):
                tokenPage = self._get_page(self._tokenURL)
@@ -273,7 +262,7 @@ class GVoiceBackend(object):
                        "btmpl": "mobile",
                        "PersistentCookie": "yes",
                        "GALX": token,
-                       "continue": self._forwardURL,
+                       "continue": self._JSON_CONTACTS_URL,
                }
 
                loginSuccessOrFailurePage = self._get_page(self._loginURL, loginData)
@@ -283,36 +272,46 @@ class GVoiceBackend(object):
                """
                Attempt to login to GoogleVoice
                @returns Whether login was successful or not
+               @blocks
                """
                self.logout()
                galxToken = self._get_token()
                loginSuccessOrFailurePage = self._login(username, password, galxToken)
 
                try:
-                       self._grab_account_info(loginSuccessOrFailurePage)
+                       accountData = self._grab_account_info(loginSuccessOrFailurePage)
                except Exception, e:
                        # Retry in case the redirect failed
-                       # luckily is_authed does everything we need for a retry
-                       loggedIn = self.is_authed(True)
-                       if not loggedIn:
+                       # luckily refresh_account_info does everything we need for a retry
+                       accountData = self.refresh_account_info()
+                       if accountData is None:
                                _moduleLogger.exception(str(e))
-                               return False
+                               return None
                        _moduleLogger.info("Redirection failed on initial login attempt, auto-corrected for this")
 
                self._browser.save_cookies()
                self._lastAuthed = time.time()
-               return True
+               return accountData
 
        def persist(self):
                self._browser.save_cookies()
 
+       def shutdown(self):
+               self._browser.save_cookies()
+               self._token = None
+               self._lastAuthed = 0.0
+
        def logout(self):
                self._browser.clear_cookies()
                self._browser.save_cookies()
                self._token = None
                self._lastAuthed = 0.0
+               self._callbackNumbers = {}
 
        def is_dnd(self):
+               """
+               @blocks
+               """
                isDndPage = self._get_page(self._isDndURL)
 
                dndGroup = self._isDndRe.search(isDndPage)
@@ -323,6 +322,9 @@ class GVoiceBackend(object):
                return isDnd
 
        def set_dnd(self, doNotDisturb):
+               """
+               @blocks
+               """
                dndPostData = {
                        "doNotDisturb": 1 if doNotDisturb else 0,
                }
@@ -332,6 +334,7 @@ class GVoiceBackend(object):
        def call(self, outgoingNumber):
                """
                This is the main function responsible for initating the callback
+               @blocks
                """
                outgoingNumber = self._send_validation(outgoingNumber)
                subscriberNumber = None
@@ -357,6 +360,7 @@ class GVoiceBackend(object):
                """
                Cancels a call matching outgoing and forwarding numbers (if given). 
                Will raise an error if no matching call is being placed
+               @blocks
                """
                page = self._get_page_with_token(
                        self._callCancelURL,
@@ -369,6 +373,9 @@ class GVoiceBackend(object):
                self._parse_with_validation(page)
 
        def send_sms(self, phoneNumbers, message):
+               """
+               @blocks
+               """
                validatedPhoneNumbers = [
                        self._send_validation(phoneNumber)
                        for phoneNumber in phoneNumbers
@@ -378,7 +385,7 @@ class GVoiceBackend(object):
                        self._sendSmsURL,
                        {
                                'phoneNumber': flattenedPhoneNumbers,
-                               'text': message
+                               'text': unicode(message).encode("utf-8"),
                        },
                )
                self._parse_with_validation(page)
@@ -387,6 +394,7 @@ class GVoiceBackend(object):
                """
                Search your Google Voice Account history for calls, voicemails, and sms
                Returns ``Folder`` instance containting matching messages
+               @blocks
                """
                page = self._get_page(
                        self._XML_SEARCH_URL,
@@ -396,6 +404,9 @@ class GVoiceBackend(object):
                return json
 
        def get_feed(self, feed):
+               """
+               @blocks
+               """
                actualFeed = "_XML_%s_URL" % feed.upper()
                feedUrl = getattr(self, actualFeed)
 
@@ -404,19 +415,21 @@ class GVoiceBackend(object):
 
                return json
 
-       def download(self, messageId, adir):
+       def recording_url(self, messageId):
+               url = self._downloadVoicemailURL+messageId
+               return url
+
+       def download(self, messageId, targetPath):
                """
                Download a voicemail or recorded call MP3 matching the given ``msg``
                which can either be a ``Message`` instance, or a SHA1 identifier. 
-               Saves files to ``adir`` (defaults to current directory). 
                Message hashes can be found in ``self.voicemail().messages`` for example. 
-               Returns location of saved file.
+               @returns location of saved file.
+               @blocks
                """
-               page = self._get_page(self._downloadVoicemailURL, {"id": messageId})
-               fn = os.path.join(adir, '%s.mp3' % messageId)
-               with open(fn, 'wb') as fo:
+               page = self._get_page(self.recording_url(messageId))
+               with open(targetPath, 'wb') as fo:
                        fo.write(page)
-               return fn
 
        def is_valid_syntax(self, number):
                """
@@ -435,8 +448,6 @@ class GVoiceBackend(object):
                @returns a dictionary mapping call back numbers to descriptions
                @note These results are cached for 30 minutes.
                """
-               if not self.is_authed():
-                       return {}
                return self._callbackNumbers
 
        def set_callback_number(self, callbacknumber):
@@ -454,56 +465,73 @@ class GVoiceBackend(object):
                """
                return self._callbackNumber
 
-       def get_recent(self):
+       def get_received_calls(self):
+               """
+               @returns Iterable of (personsName, phoneNumber, exact date, relative date, action)
+               @blocks
+               """
+               return self._parse_recent(self._get_page(self._XML_RECEIVED_URL))
+
+       def get_missed_calls(self):
                """
                @returns Iterable of (personsName, phoneNumber, exact date, relative date, action)
+               @blocks
                """
-               for action, url in (
-                       ("Received", self._XML_RECEIVED_URL),
-                       ("Missed", self._XML_MISSED_URL),
-                       ("Placed", self._XML_PLACED_URL),
-               ):
-                       flatXml = self._get_page(url)
-
-                       allRecentHtml = self._grab_html(flatXml)
-                       allRecentData = self._parse_history(allRecentHtml)
-                       for recentCallData in allRecentData:
-                               recentCallData["action"] = action
-                               yield recentCallData
-
-       def get_contacts(self):
-               """
-               @returns Iterable of (contact id, contact name)
-               """
-               page = self._get_page(self._XML_CONTACTS_URL)
-               contactsBody = self._contactsBodyRe.search(page)
-               if contactsBody is None:
-                       raise RuntimeError("Could not extract contact information")
-               accountData = _fake_parse_json(contactsBody.group(1))
-               for contactId, contactDetails in accountData["contacts"].iteritems():
-                       # A zero contact id is the catch all for unknown contacts
-                       if contactId != "0":
-                               if "name" in contactDetails:
-                                       contactDetails["name"] = unescape(contactDetails["name"])
-                               yield contactId, contactDetails
+               return self._parse_recent(self._get_page(self._XML_MISSED_URL))
+
+       def get_placed_calls(self):
+               """
+               @returns Iterable of (personsName, phoneNumber, exact date, relative date, action)
+               @blocks
+               """
+               return self._parse_recent(self._get_page(self._XML_PLACED_URL))
+
+       def get_csv_contacts(self):
+               data = {
+                       "groupToExport": "mine",
+                       "exportType": "ALL",
+                       "out": "OUTLOOK_CSV",
+               }
+               encodedData = urllib.urlencode(data)
+               contacts = self._get_page(self._CSV_CONTACTS_URL+"?"+encodedData)
+               return contacts
 
        def get_voicemails(self):
+               """
+               @blocks
+               """
                voicemailPage = self._get_page(self._XML_VOICEMAIL_URL)
                voicemailHtml = self._grab_html(voicemailPage)
                voicemailJson = self._grab_json(voicemailPage)
+               if voicemailJson is None:
+                       return ()
                parsedVoicemail = self._parse_voicemail(voicemailHtml)
                voicemails = self._merge_conversation_sources(parsedVoicemail, voicemailJson)
                return voicemails
 
        def get_texts(self):
+               """
+               @blocks
+               """
                smsPage = self._get_page(self._XML_SMS_URL)
                smsHtml = self._grab_html(smsPage)
                smsJson = self._grab_json(smsPage)
+               if smsJson is None:
+                       return ()
                parsedSms = self._parse_sms(smsHtml)
                smss = self._merge_conversation_sources(parsedSms, smsJson)
                return smss
 
+       def get_unread_counts(self):
+               countPage = self._get_page(self._JSON_SMS_COUNT_URL)
+               counts = parse_json(countPage)
+               counts = counts["unreadCounts"]
+               return counts
+
        def mark_message(self, messageId, asRead):
+               """
+               @blocks
+               """
                postData = {
                        "read": 1 if asRead else 0,
                        "id": messageId,
@@ -512,6 +540,9 @@ class GVoiceBackend(object):
                markPage = self._get_page(self._markAsReadURL, postData)
 
        def archive_message(self, messageId):
+               """
+               @blocks
+               """
                postData = {
                        "id": messageId,
                }
@@ -532,32 +563,26 @@ class GVoiceBackend(object):
                return flatHtml
 
        def _grab_account_info(self, page):
-               tokenGroup = self._tokenRe.search(page)
-               if tokenGroup is None:
-                       raise RuntimeError("Could not extract authentication token from GoogleVoice")
-               self._token = tokenGroup.group(1)
-
-               anGroup = self._accountNumRe.search(page)
-               if anGroup is not None:
-                       self._accountNum = anGroup.group(1)
-               else:
-                       _moduleLogger.debug("Could not extract account number from GoogleVoice")
-
-               self._callbackNumbers = {}
-               for match in self._callbackRe.finditer(page):
-                       callbackNumber = match.group(2)
-                       callbackName = match.group(1)
-                       self._callbackNumbers[callbackNumber] = callbackName
+               accountData = parse_json(page)
+               self._token = accountData["r"]
+               self._accountNum = accountData["number"]["raw"]
+               for callback in accountData["phones"].itervalues():
+                       self._callbackNumbers[callback["phoneNumber"]] = callback["name"]
                if len(self._callbackNumbers) == 0:
                        _moduleLogger.debug("Could not extract callback numbers from GoogleVoice (the troublesome page follows):\n%s" % page)
+               return accountData
 
        def _send_validation(self, number):
                if not self.is_valid_syntax(number):
                        raise ValueError('Number is not valid: "%s"' % number)
-               elif not self.is_authed():
-                       raise RuntimeError("Not Authenticated")
                return number
 
+       def _parse_recent(self, recentPage):
+               allRecentHtml = self._grab_html(recentPage)
+               allRecentData = self._parse_history(allRecentHtml)
+               for recentCallData in allRecentData:
+                       yield recentCallData
+
        def _parse_history(self, historyHtml):
                splitVoicemail = self._seperateVoicemailsRegex.split(historyHtml)
                for messageId, messageHtml in itergroup(splitVoicemail[1:], 2):
@@ -634,7 +659,11 @@ class GVoiceBackend(object):
                        message = Message()
                        message.body = messageParts
                        message.whoFrom = conv.name
-                       message.when = conv.time.strftime("%I:%M %p")
+                       try:
+                               message.when = conv.time.strftime("%I:%M %p")
+                       except ValueError:
+                               _moduleLogger.exception("Confusing time provided: %r" % conv.time)
+                               message.when = "Unknown"
                        conv.messages = (message, )
 
                        yield conv
@@ -724,9 +753,27 @@ class GVoiceBackend(object):
 
        def _parse_with_validation(self, page):
                json = parse_json(page)
-               validate_response(json)
+               self._validate_response(json)
                return json
 
+       def _validate_response(self, response):
+               """
+               Validates that the JSON response is A-OK
+               """
+               try:
+                       assert response is not None, "Response not provided"
+                       assert 'ok' in response, "Response lacks status"
+                       assert response['ok'], "Response not good"
+               except AssertionError:
+                       try:
+                               if response["data"]["code"] == 20:
+                                       raise RuntimeError(
+"""Ambiguous error 20 returned by Google Voice.
+Please verify you have configured your callback number (currently "%s").  If it is configured some other suspected causes are: non-verified callback numbers, and Gizmo5 callback numbers.""" % self._callbackNumber)
+                       except KeyError:
+                               pass
+                       raise RuntimeError('There was a problem with GV: %s' % response)
+
 
 _UNESCAPE_ENTITIES = {
  "&quot;": '"',
@@ -748,7 +795,7 @@ def google_strptime(time):
        """
        abbrevTime = time[:-3]
        parsedTime = datetime.datetime.strptime(abbrevTime, "%m/%d/%y %I:%M")
-       if time[-2] == "PN":
+       if time.endswith("PM"):
                parsedTime += datetime.timedelta(hours=12)
        return parsedTime
 
@@ -789,9 +836,16 @@ def itergroup(iterator, count, padValue = None):
 def safe_eval(s):
        _TRUE_REGEX = re.compile("true")
        _FALSE_REGEX = re.compile("false")
+       _COMMENT_REGEX = re.compile("^\s+//.*$", re.M)
        s = _TRUE_REGEX.sub("True", s)
        s = _FALSE_REGEX.sub("False", s)
-       return eval(s, {}, {})
+       s = _COMMENT_REGEX.sub("#", s)
+       try:
+               results = eval(s, {}, {})
+       except SyntaxError:
+               _moduleLogger.exception("Oops")
+               results = None
+       return results
 
 
 def _fake_parse_json(flattened):
@@ -821,16 +875,6 @@ def extract_payload(flatXml):
        return jsonTree, flatHtml
 
 
-def validate_response(response):
-       """
-       Validates that the JSON response is A-OK
-       """
-       try:
-               assert 'ok' in response and response['ok']
-       except AssertionError:
-               raise RuntimeError('There was a problem with GV: %s' % response)
-
-
 def guess_phone_type(number):
        if number.startswith("747") or number.startswith("1747") or number.startswith("+1747"):
                return GVoiceBackend.PHONE_TYPE_GIZMO
@@ -906,15 +950,18 @@ def grab_debug_info(username, password):
        browser = backend._browser
 
        _TEST_WEBPAGES = [
-               ("forward", backend._forwardURL),
                ("token", backend._tokenURL),
                ("login", backend._loginURL),
                ("isdnd", backend._isDndURL),
                ("account", backend._XML_ACCOUNT_URL),
-               ("contacts", backend._XML_CONTACTS_URL),
+               ("html_contacts", backend._XML_CONTACTS_URL),
+               ("contacts", backend._JSON_CONTACTS_URL),
+               ("csv", backend._CSV_CONTACTS_URL),
 
                ("voicemail", backend._XML_VOICEMAIL_URL),
-               ("sms", backend._XML_SMS_URL),
+               ("html_sms", backend._XML_SMS_URL),
+               ("sms", backend._JSON_SMS_URL),
+               ("count", backend._JSON_SMS_COUNT_URL),
 
                ("recent", backend._XML_RECENT_URL),
                ("placed", backend._XML_PLACED_URL),
@@ -945,8 +992,8 @@ def grab_debug_info(username, password):
                backend._grab_account_info(loginSuccessOrFailurePage)
        except Exception:
                # Retry in case the redirect failed
-               # luckily is_authed does everything we need for a retry
-               loggedIn = backend.is_authed(True)
+               # luckily refresh_account_info does everything we need for a retry
+               loggedIn = backend.refresh_account_info() is not None
                if not loggedIn:
                        raise
 
@@ -972,6 +1019,21 @@ def grab_debug_info(username, password):
                )
 
 
+def grab_voicemails(username, password):
+       cookieFile = os.path.join(".", "raw_cookies.txt")
+       try:
+               os.remove(cookieFile)
+       except OSError:
+               pass
+
+       backend = GVoiceBackend(cookieFile)
+       backend.login(username, password)
+       voicemails = list(backend.get_voicemails())
+       for voicemail in voicemails:
+               print voicemail.id
+               backend.download(voicemail.id, ".")
+
+
 def main():
        import sys
        logging.basicConfig(level=logging.DEBUG)
@@ -981,6 +1043,7 @@ def main():
                password = args[2]
 
        grab_debug_info(username, password)
+       grab_voicemails(username, password)
 
 
 if __name__ == "__main__":