Forums

Hi everyone, to advice all trying to install telegram bot api server here on PythonAnywhere, I will post here instructinos how to do it. Hope to get some feedback, also from more experienced developers or admins, what here can cause a trouble, or just your comments.

To begin with I am using telebot to run my telegram bot. And I wanted to install server here as I don't have separate laptop/pc that can be run 24/7, which is in my opinion better and more flexible. So I tried to start it here and thats how it worked.

                                      Running BOT on local server

So, the bot is used more effectively while running on local server build, from tdlib library. It allows increased limit of file size up to 2000mb and other bells & whistles.

As this server is built in PythonAnywhere.com service, It does not allow sudo and apt commands obviously, so you have to install it locally, as well as build hash library 'gperf' manually.

Use following steps to build the server:
Refer to build setup from website https://tdlib.github.io/telegram-bot-api/build.html?os=Linux.
Main docs - https://github.com/tdlib/telegram-bot-api.
Use script for ubuntu 20, g++, local/bin.

To begin with - enter bash terminal.
I. Clone repository of Tdlib.

$ git clone --recursive https://github.com/tdlib/telegram-bot-api.git
$ cd telegram-bot-api

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
II. Build the server

$ rm -rf build
$ mkdir build
$ cd build
$ cmake -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX:PATH=$HOME ..
$ cmake --build . --target install
$ cd ../..

So on this step with first cmake command next error will be received, so the build won't be complete.
Error:

Could NOT find gperf.  Add path to gperf executable to PATH environment
  variable or specify it manually using GPERF_EXECUTABLE option, i.e.  'cmake
  -DGPERF_EXECUTABLE:FILEPATH="<path to gperf executable>"

The error message indicates that the gperf executable (GNU gperf) is not found in your system's PATH, and the build process is unable to locate it. PythonAnywhere doesn't allow installing and compile C packages via SUDO and apt-pkg, so we have to install them manually in .local folder. Next steps are provided by PA FAQ as instructions to compile C programs. https://help.pythonanywhere.com/pages/CompilingCPrograms/

First visit the gperf download page: https://www.gnu.org/software/gperf/ and download source code (gperf-3.1.tar.gz) file. Upload it to PythoAnywhere files. Extract downloaded file using next command:

$ tar -xvzf gperf-3.1.tar.gz

Move to folder with extracted file:

 $ cd gperf-3.1

Use prefix to indicate that build will be installed in our local home directory:

$ ./configure --prefix=$HOME/.local

(Note, you have to cd in gperf-3.1 folder)

Here's what each part of the command means:

./configure: This command runs configure script that comes with the source code of the software. Configure script is responsible for preparing the build process and checking dependencies.

--prefix=/home/$HOME/.local: This flag specifies the installation prefix directory for the software. In this case, it's set to $HOME/.local, which means the software will be installed under the .local directory within your home directory ($HOME refers to your home directory).

Run 'make' command.

$ make

After success, you will see something like this:

make[1]: Entering directory '/home/User/gperf/gperf-3.1/lib'
make[1]: Nothing to be done for 'all'.
make[1]: Leaving directory...

If you see this output, it suggests that the software has already been built or processed successfully.

Run make install command

$ make install

or

make install prefix=$HOME/.local

(if first doesn't work.)

You can check if it's installed by running

$ gperf --version

or

$ which gperf

Add next lines to .bashrc configuration file of bash console to set PATH,
that will be ready each time bash console is activated:

export PATH=$PATH:$HOME/.local/bin
source ~/.bashrc
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
III. Retry building with CMake:

$ cd telegram-bot-api
$ rm -rf build
$ mkdir build
$ cd build
$ cmake -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX:PATH=$HOME -DGPERF_EXECUTABLE:FILEPATH=$HOME/.local/bin/gperf ..
$ cmake --build . --target install
$ cd ../..

This should start process of building the server.
It takes around 30 minutes and around 400 mbs of space.

IV. To finally run server use next steps: Install environmental variables constantly for bash console:
For Bash: .bashrc file.
In this file write:
export TELEGRAM_API_ID="12345"
export TELEGRAM_API_HASH="12345"

Then use command to start server.

$HOME/bin/telegram-bot-api --api-id $TELEGRAM_API_ID --api-hash $TELEGRAM_API_HASH --local &

I used & marker in the end to run process in the background, otherwise it blocks the console input.

--local flag will create local server which will increase the size limit to 2000mb for files.

To connect locally you will be able to use IP 0.0.0.0:8081, port 8081 is default lsitening port of telegram api bot server.

To terminate server:
Find the process ID (PID) of the running telegram-bot-api server. You can use the ps command to list the running processes:

$ ps

or

$ ps aux | grep telegram-bot-api

This will display a list of processes containing the string "telegram-bot-api" along with their corresponding PIDs.

Identify the process you want to stop based on its PID.

Use the kill command to send a termination signal to the process. Replace <PID> with the actual PID you identified:

$ kill 123

Adjusting code of pytelegrambotapi to run commands through new local server.
So, by default pytelegrambotapi is calling to api.telegram.org, to change it use next code block:

from telebot import apihelper
apihelper.API_URL = 'http://0.0.0.0:8081/bot{0}/{1}'
apihelper.FILE_URL = 'http://0.0.0.0:8081'

Incldue next function to log out from bot, before switching between servers.
For example if you ran your bot from IDE without local server, to start it on local server, frist you will have to logout your bot from current session.

def log_out_function():
    return bot.log_out()

And thats it. I followed these steps and was able to run server locally on PythonAnywhere. Hope it will help!

Thanks (a lot)) for instruction!!!

Thanks for mentioning! Please tell if it worked for you in the end.

Interesting. Thanks for sharing!

Interesting. Thanks for sharing!

Yes! It's working )))) Thanks 100times ) But line "source ~/.bashrc" stops downloading Bash console, after hide(commenting by #) this line => working properly