We Keep Coding

How to debug Django in VSCode with autoreload turned on

I set up debugging in VSCode (Django application) and it works great with the default setup. However, when debugging it seems automatic reloading is not possible. This is stated in the VSCode docs:
Note that automatic reloading of Django apps is not possible while debugging.
I was wondering if there's some way to make debugging (breakpoints etc.) work with reload enabled in Django.

It turns out you can make this work by using Microsoft's debugpy tool.
Django launches two processes when reloading is turned on (the default), one of them is the parent and the other is the subprocess which does the reloading magic.
Django differentiates these two processes by setting the environment variable RUN_MAIN to true in the subprocess (which does the reloading). Refer to: https://github.com/django/django/blob/8a902b7ee622ada258d15fb122092c1f02b82698/django/utils/autoreload.py#L241
By tweaking manage.py slightly, we can start a debug listener in the parent process and have it survive any reloads.
Add debugpy to however you manage your requirements (requirements.txt etc.)
Add the following function to initialize the debugger:
def initialize_debugger():
import debugpy
# optionally check to see what env you're running in, you probably only want this for
# local development, for example: if os.getenv("MY_ENV") == "dev":
# RUN_MAIN envvar is set by the reloader to indicate that this is the
# actual thread running Django. This code is in the parent process and
# initializes the debugger
if not os.getenv("RUN_MAIN"):
debugpy.listen(("0.0.0.0", 9999))
sys.stdout.write("Start the VS Code debugger now, waiting...\n")
debugpy.wait_for_client()
sys.stdout.write("Debugger attached, starting server...\n")
Alter the main function in manage.py as follows:
def main()
# <...>
initialize_debugger() # add this
execute_from_command_line(sys.argv)
Modify launch.json configuration in VSCode to attach to port 9999 (from above):
{
"name": "Python: Remote Attach (DebugPy)",
"type": "python",
"request": "attach",
"port": 9999,
"host": "localhost",
},
TIP: you can disable "Uncaught exceptions" because the reload itself causes a SystemExit

How to use the vscode python debugger with a project built with bazel?

I want to debug a python file which has a few dependencies that only appear in the runfiles from bazel. How can I debug a bazel build with the vscode debugger?

As someone famous said, "Yes, we can".
You would need to use the "ptvsd" python package.
One-time Setup
Add "ptvsd" as a Python dependency in Bazel
In VS Code, in your launch.json file, add the following configuration:
{
"name": "Python: Attach",
"type": "python",
"request": "attach",
"port": 5724,
"host": "localhost"
},
Debug
When you want to debug a specific file:
In the Python file you want to debug, add the following lines:
import ptvsd
ptvsd.enable_attach(address=('localhost', 5724), redirect_output=True)
print('Now is a good time to attach your debugger: Run: Python: Attach')
ptvsd.wait_for_attach()
Run Bazel on this file as you normally would (bazel run :server for instance)
Execution will stop at "Now is a good time to attach your debugger: Run: Python: Attach"
In VS Code, click on the "Python: Attach" debug option that we setup earlier:
That's it!
Feel free to change the port, 5724 in this example.

ptvsd is deprecated, replaced by debugpy.
Usage is the same as the accepted answer, except that the code snippet for the Python file has changed:
import debugpy
debugpy.listen(5678)
debugpy.wait_for_client() # blocks execution until client is attached

How to debug mpirun Python processes in Visual Studio Code on Ubuntu?

Create /home/bob/foobar.py in Visual Studio
Code with the VS Code Docker
Extension
import ptvsd
import time
ptvsd.enable_attach(address = ('0.0.0.0', 5678))
ptvsd.wait_for_attach()
time.sleep(2)
print("all righty then")
Set a breakpoint on the last line.
Debug|Add Configuration
In launch.json add to "configurations"
{
"name": "Python Attach (Remote Debug ptsvd default)",
"type": "python",
"request": "attach",
"pathMappings": [
{
"localRoot": "/home/bob", // You may also manually specify the directory containing your source code.
"remoteRoot": "/home/bob" // Linux example; adjust as necessary for your OS and situation.
}
],
"port": 5678, // Set to the remote port.
"host": "0.0.0.0" // Set to your remote host's public IP address.
},
$ python -m pip install --user --upgrade ptvsd
$ python foobar.py
Start the debugger with configuration "Python Attach (Remote Debug ptsvd default)". It stops at the breakpoint.
But if I run mpirun
$ mpirun --allow-run-as-root -np 2 -H localhost:2 -bind-to none -map-by slot -x PATH -mca pml ob1 -mca btl ^openib python ./foobar.py
I get the error socket.error: [Errno 98] Address already in use
Is there a way to assign multiple ports for an arbitrary number of processes in launch.json and in foobar.py?

This is not currently possible but has been added by the ptvsd team as a possible enhancement. Vote it up if this is important to you. Upvotes for this team seem to move features up the to-do list.

Python remote debuging with PyCharm and ptvsd

I have a virtual machine running a python process and I want to debug it from my host.
My virtual machine ip is 192.168.10.20.
As per the instructions I set up my launch.json as
{
"name": "Python: Attach",
"type": "python",
"request": "attach",
"localRoot": "/Users/wrkspace",
"remoteRoot": "/home/wrkspace",
"port": 3000,
"secret": "my_secret",
"host": "192.168.10.20"
},
I have two files,one on my host machine and on my vm called test.py which is a helloworld asking for a user input
My host machine has this at the top of the file
ptvsd.enable_attach("my_secret", address = ('192.168.10.20'3000))
# Enable the line of source code below only if you want the application
to wait until the debugger has attached to it
#ptvsd.wait_for_attach()
While on my vm, where I want to debug
ptvsd.enable_attach("my_secret", address = ('192.168.10.20'3000))
# Enable the line of source code below only if you want the application
to wait until the debugger has attached to it
ptvsd.wait_for_attach()
Once I run my attached debugger and then I kick of the test on my vm by going python test.py. It looks like it does something but then fails with
There was an error in starting the debug server.
Error="code":"ENETUNREACH","errno":"ENETUNREACH","syscall":"connect","address":"192.168.10.20","port":3000}
There was an error in starting the debug server. Error = {"code":"ENETUNREACH","errno":"ENETUNREACH","syscall":"connect","address":"192.168.10.20","port":3000}
Is there something I'm missing?

A few steps to trouble shoot these:
1) test the debug port to make sure it is open using "nc -zv test.com 30302"
2) make sure your webserver is not reloading the app after debugger was connected to it. this is the option for flask: " use_reloader=False"

Setting up Python remote debugging in VSCode

I am having my remote(ubuntu 16.04) drive mounted on my local system(ubuntu 16.04) so that I can edit the source files by opening them in vscode.
Also, in the integrated terminal I can ssh to remote system and run the program using remote python interpreter which is installed on virtual environment like:
$ssh username#remoteip
$workon remotevirtualenv
(remotevirtualenv)$python source.py
I want to enable remote debugging so that if I run the debug/run the remote files, the vscode which is installed on my local system uses my remote python interpreter.
I went through the docs suggesting the use of ptvsd extension(which asks to have 2 copies of source files one on local, another one server) but I am not sure how to configure it in this scenario.
Appreciate the help. Thanks.
Edit 1:
As, I said I have gone through the docs but I am not clear how to configure in this scenario. e.g. docs say
In the source code on both computers, add the following lines, replacing my_secret with the appropriate passphrase to authenticate
remote debugging, and replacing address with the appropriate IP
address (or localhost) and port number:
ptvsd.enable_attach("my_secret", address = ('0.0.0.0', 3000))
But I have only one copy of source file i.e. on remote system. I have just mounted it on my local file system. So, should I give my local ip address or it should be remote system ip and which port number should I use as I am having only one copy of source and hence ptvsd, how is it going to communicate.
Also, in the configuration, what should I use for localRoot location and remoteRoot location.
3. {
"name": "Attach (Remote Debug)",
"type": "python",
"request": "attach",
"localRoot": "${workspaceFolder}",
"remoteRoot": "${workspaceFolder}",
"port": 3000,
"secret": "my_secret",
"host": "localhost"
}
Edit: Thanks to remote development extension in VSCode, remote debugging is super easy now.

You need to put the remote ip address in both ptvsd.enable_attach("my_secret", address = ('remote_ip_address', 3000)) and in launch.json:
{
"name": "Attach (Remote Debug)",
"type": "python",
"request": "attach",
"localRoot": "${workspaceFolder}",
"remoteRoot": "/home/user1/scripts/",
"port": 3000,
"secret": "my_secret",
"host": "remote_ip_address"
}
You also need to change the remoteRoot value to the path of the directory where the script is located in the remote machine (e.g. /home/user1/scripts/).
Finally, open an ssh connection: ssh -L 3000:localhost:3000, run your script in the remote machine and attach the debugger in the local machine.

Please see the official docs on how to do remote debugging.