From 80b7367a021ee4a084e6545fbff97a3ead6653fa Mon Sep 17 00:00:00 2001
From: Anton Grigoryev <grig.anton@gmail.com>
Date: Wed, 11 Mar 2015 19:55:55 +0300
Subject: [PATCH 1/2] TL Schema moved to JSON mode. Important comments added

---
 mtproto.py | 217 +++++++++++++++++++++++------------------------------
 1 file changed, 95 insertions(+), 122 deletions(-)

diff --git a/mtproto.py b/mtproto.py
index a9f1b136..d077fbbe 100644
--- a/mtproto.py
+++ b/mtproto.py
@@ -5,57 +5,33 @@ Created on Tue Sep  2 19:26:15 2014
 @author: agrigoryev
 """
 from binascii import crc32
-import struct
-import socket
-import re
 from datetime import datetime
-import sys
 import io
-import configparser
+import json
+import socket
+import struct
 
-current_module = sys.modules[__name__]
 
 def vis(bs):
-    l = len(bs)
-    n = len(bs) // 8
+    """
+    Function to visualize byte streams. Split into bytes, print to console.
+    :param bs: BYTE STRING
+    """
+    symbols_in_one_line = 8
+    n = len(bs) // symbols_in_one_line
     for i in range(n):
-        print(" ".join(["%02X" % b for b in bs[i*8:i*8+8]]))
-    print(" ".join(["%02X" % b for b in bs[i*8+8:]])+"\n")
-
-
-class TlElement:
-    def __init__(self, type_string):
-        tl_re = re.compile("""([a-z]\w*)           #name
-                              \#?([0-9a-f]{8})?\s+ #id
-                              (\{.*\}\s+)?         #subtype
-                              (.*)\s+              #arguments list
-                              =\s+([A-Z]\w*)       #result
-                              (\s+(\w+))?;         #subresult""", re.X)
-        assert isinstance(type_string, str)
-        x = tl_re.match(type_string)
-        if x is not None:
-            self.name = x.groups()[0]
-            if x.groups()[1] is not None:
-                self.id = int(x.groups()[1], 16)
-            else:
-                self.id = crc32(re.sub("(#[0-9a-f]{8})|([;\n{}])", "", type_string).encode())
-            self.args = self.get_arg_list(x.groups()[3])
-            self.result = x.groups()[4]
-        else:
-            raise SyntaxError
-
-    @staticmethod
-    def get_arg_list(arg_string):
-        arg_re = re.compile("([\w0-9]+)?:?([\w0-9]+)(<([\w0-0]+)>)?")
-        res = []
-        for s in arg_re.findall(arg_string):
-            d = {'name': s[0], 'type': s[1], 'subtype': s[3] if s[2] is not None else None}
-            res.append(d)
-        return res
+        print(" ".join(["%02X" % b for b in bs[i*symbols_in_one_line:(i+1)*symbols_in_one_line]])) # for every 8 symbols line
+    print(" ".join(["%02X" % b for b in bs[(i+1)*symbols_in_one_line:]])+"\n") # for last line
 
 
 class TL:
     def __init__(self):
+        with open("TL_schema.JSON", 'r') as f:
+            TL_dict = json.load(f)
+
+        self.methods = TL_dict['methods']
+        self.constructors = TL_dict['constructors']
+
         self.func_dict_id = {}
         self.func_dict_name = {}
         self.obj_dict_id = {}
@@ -63,83 +39,52 @@ class TL:
 
         # Read constructors
 
-        f = open("TL_schema", 'r')
-        for line in f:
-            if line.startswith("---functions---"):
-                break
-            try:
-                z = TlElement(line)
-                self.obj_dict_id[z.id] = z
-                self.obj_dict_name[z.result] = z
-            except SyntaxError:
-                pass
+        for elem in self.constructors:
+            z = TlElement(elem)
+            self.func_dict_id = {}
 
-        # Read methods
+    def serialize(self, type_, data):
+        # 1. Type constructor ID
+        # 2. type costructor params
 
-        for line in f:
-            if line.startswith("---functions---"):
-                break
-            try:
-                z = TlElement(line)
-                self.func_dict_id[z.id] = z
-                self.func_dict_name[z.name] = z
-            except SyntaxError:
-                pass
+        # Bare types
+        if type_ == 'string':
+            assert isinstance(data, str)
+            struct.pack("<L", len(data))
+        if type_ == 'int':
+            assert isinstance(data, str)
+            struct.pack("<L", len(data))
 
+    def deserialize(self, byte_string, type_=None, subtype=None):
 
-    def tl_serialize(self, elem, kwargs):
-        assert isinstance(elem, TlElement)
-        for arg in elem.args:
-            pass
-        return struct.pack("<L", )
-
-    def tl_function_generator(elem, session, tl):
-        def dummy(**kwargs):
-            message = tl_serialize(elem, kwargs)
-            session.send_message(message)
-            answer = session.recv_message()
-            return deserialize(answer, TL)
-
-    def class_generator(self, tl_element):
-        class Dummy:
-            def __init__(self, bstring):
-                assert isinstance(bstring, bytes)
-                f = io.BytesIO(bstring)
-                dict = self.deserialize(f, type=tl_element.type)
-
-    def serialize(self, type=None, subtype=None):
-        pass
-
-    def deserialize(self, string, type=None, subtype=None):
-
-        if isinstance(string, io.BytesIO):
-            bytes_io = string
-        elif isinstance(string, bytes):
-            bytes_io = io.BytesIO(string)
+        if isinstance(byte_string, io.BytesIO):
+            bytes_io = byte_string
+        elif isinstance(byte_string, bytes):
+            bytes_io = io.BytesIO(byte_string)
         else:
-            raise Exception("Bad input type, use bytes string or BytesIO object")
+            raise TypeError("Bad input type, use bytes string or BytesIO object")
 
         # Built-in bare types
 
-        if type == 'int':
+        if type_ == 'int':
             x = struct.unpack('<i', bytes_io.read(4))[0]
-        elif type == '#':
+        elif type_ == '#':
             x = struct.unpack('<I', bytes_io.read(4))[0]
-        elif type == 'long':
+        elif type_ == 'long':
             x = struct.unpack('<q', bytes_io.read(8))[0]
-        elif type == 'double':
+        elif type_ == 'double':
             x = struct.unpack('<d', bytes_io.read(8))[0]
-        elif type == 'int128':
+        elif type_ == 'int128':
             t = struct.unpack('<16s', bytes_io.read(16))[0]
             x = int.from_bytes(t, 'little')
-        elif type == 'int256':
+        elif type_ == 'int256':
             t = struct.unpack('<32s', bytes_io.read(32))[0]
             x = int.from_bytes(t, 'little')
-        elif type == 'bytes':
+        elif type_ == 'bytes':
             l = int.from_bytes(bytes_io.read(1), 'little')
             x = bytes_io.read(l)
             bytes_io.read(-(l+1) % 4)  # skip padding bytes
-        elif type == 'string':
+        elif type_ == 'string':
             l = int.from_bytes(bytes_io.read(1), 'little')
             assert l <=254
             if l == 254:
@@ -152,10 +97,10 @@ class TL:
                 x = bytes_io.read(l)
                 bytes_io.read(-(l+1) % 4)  # skip padding bytes
             assert isinstance(x, bytes)
-        elif type == 'vector':
+        elif type_ == 'vector':
             assert subtype is not None
             count = int.from_bytes(bytes_io.read(4), 'little')
-            x = [self.deserialize(bytes_io, type=subtype) for i in range(count)]
+            x = [self.deserialize(bytes_io, type_=subtype) for i in range(count)]
         else:
             # Boxed types
 
@@ -163,57 +108,85 @@ class TL:
             try:
                 tl_elem = self.obj_dict_id[i]
             except:
-                raise Exception("Could not extract type: %s" % type)
+                raise Exception("Could not extract type: %s" % type_)
             base_boxed_types = ["Vector", "Int", "Long", "Double", "String", "Int128", "Int256"]
             if tl_elem.result in base_boxed_types:
-                x = self.deserialize(bytes_io, type=tl_elem.name, subtype=subtype)
+                x = self.deserialize(bytes_io, type_=tl_elem.name, subtype=subtype)
 
             else:  # other types
                 x = {}
                 for arg in tl_elem.args:
-                    x[arg['name']] = self.deserialize(bytes_io, type=arg['type'], subtype=arg['subtype'])
+                    x[arg['name']] = self.deserialize(bytes_io, type_=arg['type'], subtype=arg['subtype'])
         return x
 
 
-
 class Session:
-    def __init__(self, credentials):
-        config = configparser.ConfigParser()
-        config.read(credentials)
+    def __init__(self, ip, port):
+        # creating socket
         self.sock = socket.socket()
-        self.sock.connect((config['App data']['ip_adress'], config['App data'].getint('port')))
-        self.api_id = config['App data'].getint('api_id')
-        self.api_hash = config['App data']['api_hash']
-        self.auth_key_id = b'\x00\x00\x00\x00\x00\x00\x00\x00'
+        self.sock.connect((ip, port))
+        self.auth_key_id = None
         self.number = 0
-        self.TL = TL()
 
-    def header(self, message):
-        return (self.auth_key_id +
-                struct.pack('<Q', int(datetime.utcnow().timestamp()*2**32)) +
+    def __del__(self):
+        # closing socket when session object is deleted
+        self.sock.close()
+
+    @staticmethod
+    def header_unencrypted(message):
+        """
+        Creating header for the unencrypted message:
+        :param message: byte string to send
+        """
+        # Basic instructions: https://core.telegram.org/mtproto/description#unencrypted-message
+
+        # Message id: https://core.telegram.org/mtproto/description#message-identifier-msg-id
+        msg_id = int(datetime.utcnow().timestamp()*2**30)*4
+
+        return (b'\x00\x00\x00\x00\x00\x00\x00\x00' +
+                struct.pack('<Q', msg_id) +
                 struct.pack('<L', len(message)))
 
+    # TCP Transport
+
+    # Instructions may be found here: https://core.telegram.org/mtproto#tcp-transport
+    # If a payload (packet) needs to be transmitted from server to client or from client to server,
+    # it is encapsulated as follows: 4 length bytes are added at the front (to include the length,
+    # the sequence number, and CRC32; always divisible by 4) and 4 bytes with the packet sequence number
+    # within this TCP connection (the first packet sent is numbered 0, the next one 1, etc.),
+    # and 4 CRC32 bytes at the end (length, sequence number, and payload together).
+
     def send_message(self, message):
+        """
+        Forming the message frame and sending message to server
+        :param message: byte string to send
+        """
+
         print('>>')
-        vis(message)
-        #self.number += 1
-        data = self.header(message) + message
+        vis(message)  # Sending message visualisation to console
+        data = self.header_unencrypted(message) + message
         step1 = struct.pack('<LL', len(data)+12, self.number) + data
         step2 = step1 + struct.pack('<L', crc32(step1))
         self.sock.send(step2)
+        self.number += 1
 
     def recv_message(self):
+        """
+        Reading socket and receiving message from server. Check the CRC32 and
+        """
         packet_length_data = self.sock.recv(4)
         if len(packet_length_data) > 0:  # if we have smth. in the socket
             packet_length = struct.unpack("<L", packet_length_data)[0]
             packet = self.sock.recv(packet_length - 4)
             self.number = struct.unpack("<L", packet[0:4])[0]
             auth_key_id = struct.unpack("<8s", packet[4:12])[0]
-            message_id =  struct.unpack("<8s", packet[12:20])[0]
+            message_id = struct.unpack("<8s", packet[12:20])[0]
             message_length = struct.unpack("<I", packet[20:24])[0]
             data = packet[24:24+message_length]
             crc = packet[-4:]
-            if crc32(packet_length_data + packet[0:-4]).to_bytes(4,'little') == crc:
+
+            # Checking the CRC32 correctness of received data
+            if crc32(packet_length_data + packet[0:-4]).to_bytes(4, 'little') == crc:
                 print('<<')
-                vis(data)
+                vis(data)  # Received message visualisation to console
                 return data

From b2bcf483d40970d423bb448b9965104eef319d68 Mon Sep 17 00:00:00 2001
From: Anton Grigoryev <grig.anton@gmail.com>
Date: Wed, 11 Mar 2015 19:58:26 +0300
Subject: [PATCH 2/2] TL Schema moved to JSON mode.

---
 TL_schema.JSON | 1 +
 1 file changed, 1 insertion(+)
 create mode 100644 TL_schema.JSON

diff --git a/TL_schema.JSON b/TL_schema.JSON
new file mode 100644
index 00000000..3bd91940
--- /dev/null
+++ b/TL_schema.JSON
@@ -0,0 +1 @@
+{"constructors":[{"id":"481674261","predicate":"vector","params":[],"type":"Vector t"},{"id":"85337187","predicate":"resPQ","params":[{"name":"nonce","type":"int128"},{"name":"server_nonce","type":"int128"},{"name":"pq","type":"bytes"},{"name":"server_public_key_fingerprints","type":"Vector<long>"}],"type":"ResPQ"},{"id":"-2083955988","predicate":"p_q_inner_data","params":[{"name":"pq","type":"bytes"},{"name":"p","type":"bytes"},{"name":"q","type":"bytes"},{"name":"nonce","type":"int128"},{"name":"server_nonce","type":"int128"},{"name":"new_nonce","type":"int256"}],"type":"P_Q_inner_data"},{"id":"1013613780","predicate":"p_q_inner_data_temp","params":[{"name":"pq","type":"bytes"},{"name":"p","type":"bytes"},{"name":"q","type":"bytes"},{"name":"nonce","type":"int128"},{"name":"server_nonce","type":"int128"},{"name":"new_nonce","type":"int256"},{"name":"expires_in","type":"int"}],"type":"P_Q_inner_data"},{"id":"2043348061","predicate":"server_DH_params_fail","params":[{"name":"nonce","type":"int128"},{"name":"server_nonce","type":"int128"},{"name":"new_nonce_hash","type":"int128"}],"type":"Server_DH_Params"},{"id":"-790100132","predicate":"server_DH_params_ok","params":[{"name":"nonce","type":"int128"},{"name":"server_nonce","type":"int128"},{"name":"encrypted_answer","type":"bytes"}],"type":"Server_DH_Params"},{"id":"-1249309254","predicate":"server_DH_inner_data","params":[{"name":"nonce","type":"int128"},{"name":"server_nonce","type":"int128"},{"name":"g","type":"int"},{"name":"dh_prime","type":"bytes"},{"name":"g_a","type":"bytes"},{"name":"server_time","type":"int"}],"type":"Server_DH_inner_data"},{"id":"1715713620","predicate":"client_DH_inner_data","params":[{"name":"nonce","type":"int128"},{"name":"server_nonce","type":"int128"},{"name":"retry_id","type":"long"},{"name":"g_b","type":"bytes"}],"type":"Client_DH_Inner_Data"},{"id":"1003222836","predicate":"dh_gen_ok","params":[{"name":"nonce","type":"int128"},{"name":"server_nonce","type":"int128"},{"name":"new_nonce_hash1","type":"int128"}],"type":"Set_client_DH_params_answer"},{"id":"1188831161","predicate":"dh_gen_retry","params":[{"name":"nonce","type":"int128"},{"name":"server_nonce","type":"int128"},{"name":"new_nonce_hash2","type":"int128"}],"type":"Set_client_DH_params_answer"},{"id":"-1499615742","predicate":"dh_gen_fail","params":[{"name":"nonce","type":"int128"},{"name":"server_nonce","type":"int128"},{"name":"new_nonce_hash3","type":"int128"}],"type":"Set_client_DH_params_answer"},{"id":"-212046591","predicate":"rpc_result","params":[{"name":"req_msg_id","type":"long"},{"name":"result","type":"Object"}],"type":"RpcResult"},{"id":"558156313","predicate":"rpc_error","params":[{"name":"error_code","type":"int"},{"name":"error_message","type":"string"}],"type":"RpcError"},{"id":"1579864942","predicate":"rpc_answer_unknown","params":[],"type":"RpcDropAnswer"},{"id":"-847714938","predicate":"rpc_answer_dropped_running","params":[],"type":"RpcDropAnswer"},{"id":"-1539647305","predicate":"rpc_answer_dropped","params":[{"name":"msg_id","type":"long"},{"name":"seq_no","type":"int"},{"name":"bytes","type":"int"}],"type":"RpcDropAnswer"},{"id":"155834844","predicate":"future_salt","params":[{"name":"valid_since","type":"int"},{"name":"valid_until","type":"int"},{"name":"salt","type":"long"}],"type":"FutureSalt"},{"id":"-1370486635","predicate":"future_salts","params":[{"name":"req_msg_id","type":"long"},{"name":"now","type":"int"},{"name":"salts","type":"vector<future_salt>"}],"type":"FutureSalts"},{"id":"880243653","predicate":"pong","params":[{"name":"msg_id","type":"long"},{"name":"ping_id","type":"long"}],"type":"Pong"},{"id":"-501201412","predicate":"destroy_session_ok","params":[{"name":"session_id","type":"long"}],"type":"DestroySessionRes"},{"id":"1658015945","predicate":"destroy_session_none","params":[{"name":"session_id","type":"long"}],"type":"DestroySessionRes"},{"id":"-1631450872","predicate":"new_session_created","params":[{"name":"first_msg_id","type":"long"},{"name":"unique_id","type":"long"},{"name":"server_salt","type":"long"}],"type":"NewSession"},{"id":"1945237724","predicate":"msg_container","params":[{"name":"messages","type":"vector<%Message>"}],"type":"MessageContainer"},{"id":"1538843921","predicate":"message","params":[{"name":"msg_id","type":"long"},{"name":"seqno","type":"int"},{"name":"bytes","type":"int"},{"name":"body","type":"Object"}],"type":"Message"},{"id":"-530561358","predicate":"msg_copy","params":[{"name":"orig_message","type":"Message"}],"type":"MessageCopy"},{"id":"812830625","predicate":"gzip_packed","params":[{"name":"packed_data","type":"bytes"}],"type":"Object"},{"id":"1658238041","predicate":"msgs_ack","params":[{"name":"msg_ids","type":"Vector<long>"}],"type":"MsgsAck"},{"id":"-1477445615","predicate":"bad_msg_notification","params":[{"name":"bad_msg_id","type":"long"},{"name":"bad_msg_seqno","type":"int"},{"name":"error_code","type":"int"}],"type":"BadMsgNotification"},{"id":"-307542917","predicate":"bad_server_salt","params":[{"name":"bad_msg_id","type":"long"},{"name":"bad_msg_seqno","type":"int"},{"name":"error_code","type":"int"},{"name":"new_server_salt","type":"long"}],"type":"BadMsgNotification"},{"id":"2105940488","predicate":"msg_resend_req","params":[{"name":"msg_ids","type":"Vector<long>"}],"type":"MsgResendReq"},{"id":"-630588590","predicate":"msgs_state_req","params":[{"name":"msg_ids","type":"Vector<long>"}],"type":"MsgsStateReq"},{"id":"81704317","predicate":"msgs_state_info","params":[{"name":"req_msg_id","type":"long"},{"name":"info","type":"bytes"}],"type":"MsgsStateInfo"},{"id":"-1933520591","predicate":"msgs_all_info","params":[{"name":"msg_ids","type":"Vector<long>"},{"name":"info","type":"bytes"}],"type":"MsgsAllInfo"},{"id":"661470918","predicate":"msg_detailed_info","params":[{"name":"msg_id","type":"long"},{"name":"answer_msg_id","type":"long"},{"name":"bytes","type":"int"},{"name":"status","type":"int"}],"type":"MsgDetailedInfo"},{"id":"-2137147681","predicate":"msg_new_detailed_info","params":[{"name":"answer_msg_id","type":"long"},{"name":"bytes","type":"int"},{"name":"status","type":"int"}],"type":"MsgDetailedInfo"},{"id":"1973679973","predicate":"bind_auth_key_inner","params":[{"name":"nonce","type":"long"},{"name":"temp_auth_key_id","type":"long"},{"name":"perm_auth_key_id","type":"long"},{"name":"temp_session_id","type":"long"},{"name":"expires_at","type":"int"}],"type":"BindAuthKeyInner"}],"methods":[{"id":"1615239032","method":"req_pq","params":[{"name":"nonce","type":"int128"}],"type":"ResPQ"},{"id":"-686627650","method":"req_DH_params","params":[{"name":"nonce","type":"int128"},{"name":"server_nonce","type":"int128"},{"name":"p","type":"bytes"},{"name":"q","type":"bytes"},{"name":"public_key_fingerprint","type":"long"},{"name":"encrypted_data","type":"bytes"}],"type":"Server_DH_Params"},{"id":"-184262881","method":"set_client_DH_params","params":[{"name":"nonce","type":"int128"},{"name":"server_nonce","type":"int128"},{"name":"encrypted_data","type":"bytes"}],"type":"Set_client_DH_params_answer"},{"id":"1491380032","method":"rpc_drop_answer","params":[{"name":"req_msg_id","type":"long"}],"type":"RpcDropAnswer"},{"id":"-1188971260","method":"get_future_salts","params":[{"name":"num","type":"int"}],"type":"FutureSalts"},{"id":"2059302892","method":"ping","params":[{"name":"ping_id","type":"long"}],"type":"Pong"},{"id":"-213746804","method":"ping_delay_disconnect","params":[{"name":"ping_id","type":"long"},{"name":"disconnect_delay","type":"int"}],"type":"Pong"},{"id":"-414113498","method":"destroy_session","params":[{"name":"session_id","type":"long"}],"type":"DestroySessionRes"},{"id":"-1835453025","method":"http_wait","params":[{"name":"max_delay","type":"int"},{"name":"wait_after","type":"int"},{"name":"max_wait","type":"int"}],"type":"HttpWait"}]}
\ No newline at end of file