MadelineProto

Created by Daniil Gentili, licensed under AGPLv3.

Logo created by Matthew Hesketh (thanks again!).

PHP implementation of MTProto, based on telepy.

This project can run on PHP 7 and HHVM, both 32 bit and 64 bit systems are supported. You must also install the mbstring, curl extensions and the PHP Lua extension if you want to use the lua binding.

Also note that MadelineProto will perform better if python and a big math extension like gmp or bcmath are installed.

This project is in beta state.

The MadelineProto API documentation can be found here (layer 65). VERY IMPORTANT READ THIS.

The TD documentation can be found here.

The MadelineProto API documentation (mtproto tl scheme) can be found here.

The MadelineProto API documentations (old layers) can be found here.

Features:

• It allows you to do everything official clients can do, programmatically!

• It is very fast!

• It can be easily serialized!

• It featured update handling with callbacks or long polling!

• Easy to use wrappers to upload/download files and call mtproto methods

• Documentation for EVERY mtproto method!

• Internal peer management: you can provide a simple bot API chat id or a username to send a message or to call other mtproto methods!

• You can easily login as a user (2FA is supported) or as a bot!

• Simple error handling!

• It is highly customizable with a lot of different settings!

• Bot API file id/object support (even for users)!

• A Lua binding

• A lua wrapper for td-cli scripts

• Secret chats

• PFS

• PFS in secret chats

Acknowledgements

While writing this client, I looked at many projects for inspiration and help. Here's the full list:

• tgl

• Kotlogram

• Webogram

• Telethon

Thanks to the devs that contributed to these projects, MadelineProto is now an easy, well-written and complete MTProto client.

Usage

RTFM

If you have some questions about the usage of the methods of this library, you can join the support group or contact @danogentili.

But first, please read this WHOLE page very carefully, follow all links to external documentation, and read all examples in the repo.

If you don't understand something, read everything again.

I will NOT answer to questions that can be answered simply by reading this page; I will instead ask you to read it carefully twice.

Installation

git clone https://github.com/danog/MadelineProto
cd MadelineProto

Now copy .env.example to .env, edit the its values, read the docs and take a look at tests/testing.php, bot.php.

Dependencies

To install dependencies install composer and run:

composer update

In the cloned repo.

Instantiation

$MadelineProto = new \danog\MadelineProto\API();

Settings

The constructor accepts an optional parameter, which is the settings array. This array contains some other arrays, which are the settings for a specific MadelineProto function.
See here for the default values for the settings arrays and explanations for every setting.

You can provide part of any subsetting array, that way the remaining arrays will be automagically set to default and undefined values of specified subsetting arrays will be set to the default values.
Example:

$settings = [
    'authorization' => [ // Authorization settings
        'default_temp_auth_key_expires_in' => 86400, // a day
    ]
]

Becomes:

$settings = [
    'authorization' => [ // Authorization settings
        'default_temp_auth_key_expires_in' => 86400,
        'rsa_key'                          => '-----BEGIN RSA PUBLIC KEY-----
MIIBCgKCAQEAwVACPi9w23mF3tBkdZz+zwrzKOaaQdr01vAbU4E1pvkfj4sqDsm6
lyDONS789sVoD/xCS9Y0hkkC3gtL1tSfTlgCMOOul9lcixlEKzwKENj1Yz/s7daS
an9tqw3bfUV/nqgbhGX81v/+7RFAEd+RwFnK7a+XYl9sluzHRyVVaTTveB2GazTw
Efzk2DWgkBluml8OREmvfraX3bkHZJTKX4EQSjBbbdJ2ZXIsRrYOXfaA+xayEGB+
8hdlLmAjbCVfaigxX0CDqWeR1yFL9kwd9P0NsZRPsmoqVwMbMu7mStFai6aIhc3n
Slv8kg9qv1m6XHVQY3PnEw+QQtqSIXklHwIDAQAB
-----END RSA PUBLIC KEY-----',
    ]
    // The remaining subsetting arrays are the set to default
]

Note that only settings arrays or values of a settings array will be set to default.

The settings array can be accessed in the instantiated class like this:

$MadelineProto = new \danog\MadelineProto\API();
var_dump($MadelineProto->get_settings());

The settings array can be modified in the instantiated class like this:

$MadelineProto = new \danog\MadelineProto\API();
$settings = $MadelineProto->get_settings();
// Make changes to $settings
$MadelineProto->update_settings($settings);

Handling updates

When an update is received, the update callback function (see settings) is called. By default, the get_updates_update_handler MadelineProto method is called. This method stores all incoming updates into an array (its size limit is specified by the updates_array_limit parameter in the settings) and can be fetched by running the get_updates method.
IMPORTANT Note that you should turn off update handling if you don't plan to use it because the default get_updates update handling stores updates in an array inside the MadelineProto class, without deleting old ones unless they are read using get_updates. This will eventually fill up the RAM of your server if you don't disable updates or read them using get_updates.
This method accepts an array of options as the first parameter, and returns an array of updates (an array containing the update id and an object of type Update). Example:

$MadelineProto = new \danog\MadelineProto\API();
// Login or deserialize

$offset = 0;
while (true) {
    $updates = $MadelineProto->API->get_updates(['offset' => $offset, 'limit' => 50, 'timeout' => 1]); // Just like in the bot API, you can specify an offset, a limit and a timeout
    foreach ($updates as $update) {
        $offset = $update['update_id']; // Just like in the bot API, the offset must be set to the last update_id
        // Parse $update['update'], that is an object of type Update
    }
    var_dump($updates);
}

array(3) {
  [0]=>
  array(2) {
    ["update_id"]=>
    int(0)
    ["update"]=>
    array(5) {
      ["_"]=>
      string(22) "updateNewAuthorization"
      ["auth_key_id"]=>
      int(-8182897590766478746)
      ["date"]=>
      int(1483110797)
      ["device"]=>
      string(3) "Web"
      ["location"]=>
      string(25) "IT, 05 (IP = 79.2.51.203)"
    }
  }
  [1]=>
  array(2) {
    ["update_id"]=>
    int(1)
    ["update"]=>
    array(3) {
      ["_"]=>
      string(23) "updateReadChannelOutbox"
      ["channel_id"]=>
      int(1049295266)
      ["max_id"]=>
      int(8288)
    }
  }
  [2]=>
  array(2) {
    ["update_id"]=>
    int(2)
    ["update"]=>
    array(4) {
      ["_"]=>
      string(23) "updateNewChannelMessage"
      ["message"]=>
      array(12) {
        ["_"]=>
        string(7) "message"
        ["out"]=>
        bool(false)
        ["mentioned"]=>
        bool(false)
        ["media_unread"]=>
        bool(false)
        ["silent"]=>
        bool(false)
        ["post"]=>
        bool(false)
        ["id"]=>
        int(11521)
        ["from_id"]=>
        int(262946027)
        ["to_id"]=>
        array(2) {
          ["_"]=>
          string(11) "peerChannel"
          ["channel_id"]=>
          int(1066910227)
        }
        ["date"]=>
        int(1483110798)
        ["message"]=>
        string(3) "yay"
        ["entities"]=>
        array(1) {
          [0]=>
          array(4) {
            ["_"]=>
            string(24) "messageEntityMentionName"
            ["offset"]=>
            int(0)
            ["length"]=>
            int(3)
            ["user_id"]=>
            int(101374607)
          }
        }
      }
      ["pts"]=>
      int(13010)
      ["pts_count"]=>
      int(1)
    }
  }
}

To specify a custom callback change the correct value in the settings. The specified callable must accept one parameter for the update.

Uploading and downloading files

MadelineProto provides wrapper methods to upload and download files that support bot API file ids.

Every method described in this section accepts a last optional paramater with a callable function that will be called during the upload/download using the first parameter to pass a floating point number indicating the upload/download status in percentage.

The upload method returns an InputFile object that must be used to generate an InputMedia object, that can be later sent using the sendmedia method.

The upload_encrypted method returns an InputEncryptedFile object that must be used to generate an EncryptedMessage object, that can be later sent using the sendEncryptedFile method.

$inputFile = $MadelineProto->upload('file', 'optional new file name.ext');
// Generate an inputMedia object and store it in $inputMedia, see tests/testing.php
$MadelineProto->messages->sendMedia(['peer' => '@pwrtelegramgroup', 'media' => $inputMedia]);

$inputEncryptedFile = $MadelineProto->upload_encrypted('file', 'optional new file name.ext');

To convert the result of sendMedia to a bot API file id select the messageMedia object from the output of the method and pass it to $MadelineProto->API->MTProto_to_botAPI().

See tests/testing.php for more examples.

There are multiple download methods that allow you to download a file to a directory, to a file or to a stream.
The first parameter of these functions must always be either a messageMediaPhoto or a messageMediaDocument object, an EncryptedMessage or a bot API file id. These objects are usually received in updates, see bot.php for examples

$output_file_name = $MadelineProto->download_to_dir($message_media, '/tmp/dldir');
$custom_output_file_name = $MadelineProto->download_to_file($message_media, '/tmp/dldir/customname.ext');
$stream = fopen('php://output', 'w'); // Stream to browser like with echo
$MadelineProto->download_to_stream($message_media, $stream, $cb, $offset, $endoffset); // offset and endoffset are optional parameters that specify the byte from which to start downloading and the byte where to stop downloading (the latter non-inclusive), if not specified default to 0 and the size of the file

Secret chats

MadelineProto provides some wrappers to work with secret chats:

$secret_chat = $MadelineProto->request_secret_chat($InputUser);

request_secret_chat requests a secret secret chat to the InputUser specified, and returns a number that can be used instead of an InputEncryptedChat.

Secret chats are accepted or refused automatically, based on a value in the settings array (by default MadelineProto is set to accept all secret chats).

Before sending any message, you must check if the secret chat was accepted by the other client with the following method:

$status = $MadelineProto->secret_chat_status($chat);

Returns 0 if the chat cannot be found in the local database, 1 if the chat was requested but not yet accepted, and 2 if it is a valid accepted secret chat.

To send messages/files/service messages, simply use the sendEncrypted methods with objects that use the same layer used by the other client (specified by the number after the underscore in decryptedMessage object names, to obtain the layer that must be used for a secret chat use the following wrapper method).

$secret_chat = $MadelineProto->get_secret_chat($chat);
/*
[
    'key' => [ // The authorization key
        'auth_key' => 'string', // 256 bytes long
        'fingerprint' => 10387374747492, // a 64 bit signed integer
        'visualization_orig' => 'string', // 16 bytes long
        'visualization_46' => 'string', // 20 bytes long
         // The two visualization strings must be concatenated to generate a visual fingerprint
    ],
    'admin' => false, // Am I the creator of the chat?
    'user_id' => 101374607, // The user id of the other user
    'InputEncryptedChat' => [...], // An inputEncryptedChat object that represents the current chat
    'in_seq_no_x' => number, // in_seq_no must be multiplied by two and incremented by this before being sent over the network
    'out_seq_no_x' => number, // out_seq_no must be multiplied by two and incremeneted this begore being sent over the network
    'layer' => number, // The secret chat TL layer used by the other client
    'ttl' => number, // The default time to live of messages in this chat
    'ttr' => 100, // Time left before rekeying must be done, decremented by one every time a message as encrypted/decrypted with this key
    'updated' => time(), // Last time the key of the current chat was changed
    'incoming' => [], // Incoming messages, TL serialized strings
    'outgoing' => [], // Outgoing ~
    'created' => time(), // When was this chat created
    'rekeying' => [0] // Info for rekeying
];
*/

This method gets info about a certain chat.

Lua binding

The lua binding makes use of the Lua php extension.

When istantiating the \danog\MadelineProto\Lua class, the first parameter provided to the constructor must be the path to the lua script, and the second parameter a logged in instance of MadelineProto.

The class is basically a wrapper for the lua environment, so by setting an attribute you're setting a variable in the Lua environment, by reading an attribute you're reading a variable from the lua environment, and by calling a function you're actually calling a Lua function you declared in the script.

By assigning a callable to an attribute, you're actually assigning a new function in the lua environment that once called, will call the php callable.

Passing lua callables to a parameter of a PHP callable will throw an exception due to a bug in the PHP lua extension that I gotta fix (so passing the usual cb and cb_extra parameters to the td-cli wrappers isn't yet possible).

All MadelineProto wrapper methods (for example upload, download, upload_encrypted, get_self, and others) are imported in the Lua environment, as well as all MTProto wrappers (see the API docs for more info).

td-cli wrappers are also present: you can use the tdcli_function in lua and pass mtproto updates to the tdcli_update_callback via PHP, they will be automatically converted to/from td objects. Please note that the object conversion is not complete, feel free to contribute to the conversion module in src/danog/MadelineProto/Conversion/TD.php.

For examples, see lua/*.

Calling mtproto methods and available wrappers

The API documentation can be found here.
To call an MTProto method simply call it as if it is a method of the API class, substitute namespace sepators (.) with -> if needed. Also, an object of type User, InputUser, Chat, InputChannel, Peer or InputPeer must be provided as a parameter to a method, you can substitute it with the user/group/channel's username or bot API id.

$MadelineProto = new \danog\MadelineProto\API();
$checkedPhone = $MadelineProto->auth->checkPhone( // auth.checkPhone becomes auth->checkPhone
    [
        'phone_number'     => '3993838383', // Random invalid number, note that there should be no +
    ]
);
$ping = $MadelineProto->ping([3]); // parameter names can be omitted as long as the order specified by the TL scheme is respected
$message = "Hey! I'm sending this message with MadelineProto!";
$sentMessage = $MadelineProto->messages->sendMessage(['peer' => '@danogentili', 'message' => $message]);
var_dump($sentMessage);

$me = $MadelineProto->get_self(); // This gets info about the currently logged in user as a User object

The API class also provides some wrapper methods for logging in as a bot or as a normal user, and for getting inputPeer constructors to use in sendMessage and other methods:

$sentCode = $MadelineProto->phone_login($number); // Send code
var_dump($sentCode);
echo 'Enter the code you received: ';
$code = '';
for ($x = 0; $x < $sentCode['type']['length']; $x++) {
    $code .= fgetc(STDIN);
}
$authorization = $MadelineProto->complete_phone_login($code); // Complete authorization
var_dump($authorization);

$authorization = $MadelineProto->bot_login($token); // Note that every time you login as a bot or as a user MadelineProto will logout first, so now MadelineProto is logged in as the bot with token $token, not as the user with number $number
var_dump($authorization);

See tests/testing.php for more examples.

Use get_pwr_chat to get chat info, see here for the parameters and the result.

Methods that allow sending message entities (messages.sendMessage for example) also have an additional parse_mode parameter that enables or disables html/markdown parsing of the message to be sent. See the method-specific documentation for more info.

To convert the results of methods to bot API objects you must provide a second parameter to method wrappers, containing an array with the botAPI key set to true.

reply_markup accepts bot API reply markup objects as well as MTProto ones.

Note that when you login as a bot, MadelineProto also logins using the PWRTelegram API, to allow persistant storage of peers, even after a logout and another login.

Storing sessions

An istance of MadelineProto can be safely serialized or unserialized. To serialize MadelineProto to a file, usage of the \danog\MadelineProto\Serialization class is recommended:

$MadelineProto = \danog\MadelineProto\Serialization::deserialize('session.madeline');
// Do stuff
\danog\MadelineProto\Serialization::serialize('session.madeline', $MadelineProto);

That class serializes only if the $MadelineProto->API->should_serialize boolean is set to true. The same operation should be done when serializing to another destination manually, to avoid conflicts with other PHP scripts that are trying to serialize another instance of the class.

Exceptions

MadelineProto can throw lots of different exceptions:

• \danog\MadelineProto\Exception - Default exception, thrown when a php error occures and in a lot of other cases

• \danog\MadelineProto\RPCErrorException - Thrown when an RPC error occurres (an error received via the mtproto API)

• \danog\MadelineProto\TL\Exception - Thrown on TL serialization/deserialization errors

• \danog\MadelineProto\NothingInTheSocketException - Thrown if no data can be read from the TCP socket

• \danog\MadelineProto\SecurityException - Thrown on security problems (invalid params during generation of auth key or similar)

• \danog\MadelineProto\TL\Conversion\Exception - Thrown if some param/object can't be converted to/from bot API/TD/TD-CLI format (this includes markdown/html parsing)

Contributing

Here you can find this project's roadmap.

You can use this scheme of the structure of this project to help yourself:

build_docs.php - Builds API docs from TL scheme file
src/danog/MadelineProto/
    MTProtoTools/
        AckHandler - Handles acknowledgement of incoming and outgoing mtproto messages
        AuthKeyHandler - Handles generation of the temporary and permanent authorization keys
        CallHandler - Handles synchronous calls to mtproto methods or objects, also basic response management (waits until the socket receives a response)
        Crypt - Handles ige and aes encryption
        MessageHandler - Handles sending and receiving of mtproto messages (packs TL serialized data with message id, auth key id and encrypts it with Crypt if needed, adds them to the arrays of incoming and outgoing messages)
        MsgIdHandler - Handles message ids (checks if they are valid, adds them to the arrays of incoming and outgoing messages)
        ResponseHandler - Handles the content of responses received, service messages, rpc results, errors, and stores them into the response section of the outgoing messages array)
        SaltHandler - Handles server salts
        SeqNoHandler - Handles sequence numbers (checks validity)
        PeerHandler - Manages peers
        UpdateHandler - Handles updates
    TL/
        Exception - Handles exceptions in the TL namespace
        TL - Handles TL serialization and deserialization
        TLConstructor - Stores TL constructors
        TLMethod - Stores TL methods
        TLParams - Parses params
    Wrappers/
        Login - Handles logging in as a bot or a user, logging out
        PeerHandler - Eases getting of input peer objects using usernames or bot API chat ids
        SettingsManager - Eases updating settings
    API - Wrapper class that instantiates the MTProto class, sets the error handler, provides a wrapper for calling mtproto methods directly as class submethods, and uses the simplified wrappers from Wrappers/
    APIFactory - Provides a wrapper for calling namespaced mtproto methods directly as class submethods
    Connection - Handles tcp/udp/http connections and wrapping payloads generated by MTProtoTools/MessageHandler into the right message according to the protocol, stores authorization keys, session id and sequence number
    DataCenter - Handles mtproto datacenters (is a wrapper for Connection classes)
    DebugTools - Various debugging tools
    Exception - Handles exceptions and PHP errors
    RPCErrorException - Handles RPC errors
    MTProto - Handles initial connection, generation of authorization keys, instantiation of classes, writing of client info
    Logger - Static logging class
    prime.py and getpq.py - prime module (python) for p and q generation
    PrimeModule.php - prime module (php) for p and q generation by wrapping the python module, using wolfram alpha or a built in PHP engine
    RSA - Handles RSA public keys and signatures
    Tools - Various tools (positive modulus, string2bin, python-like range)

Check out the Contribution guide before contributing.