Use this open source tool to monitor variables in Python |

Use this open source tool to monitor variables in Python

Watchpoints is a simple but powerful tool to help you with monitoring variables while debugging Python.

Looking back with binoculars
Image by : 
Tyne & Wear Archives & Museums. Modified by CC BY-SA 4.0

Subscribe now

Get the highlights in your inbox every week.

When debugging code, you're often faced with figuring out when a variable changes. Without any advanced tools, you have the option of using print statements to announce the variables when you expect them to change. However, this is a very ineffective way because the variables could change in many places, and constantly printing them to a terminal is noisy, while printing them to a log file becomes unwieldy.

This is a common issue, but now there is a simple but powerful tool to help you with monitoring variables: watchpoints.

The watchpoint concept is common in C and C++ debuggers to monitor memories, but there's a lack of equivalent tools in Python. watchpoints fills in the gap.


To use it, you must first install it by using pip:

$ python3 -m pip install watchpoints

Using watchpoints in Python

For any variable you'd like to monitor, use the watch function on it.

from watchpoints import watch

a = 0
a = 1

As the variable changes, information about its value is printed to stdout:

====== Watchpoints Triggered ======

Call Stack (most recent call last):
  <module> (
> a = 1

The information includes:

  • The line where the variable was changed.
  • The call stack.
  • The previous/current value of the variable.

It not only works with the variable itself, but it also works with object changes:

from watchpoints import watch

a = []
a = {} # Trigger
a["a"] = 2 # Trigger

The callback is triggered when the variable a is reassigned, but also when the object assigned to a is changed.

What makes it even more interesting is that the monitor is not limited by the scope. You can watch the variable/object anywhere you want, and the callback is triggered no matter what function the program is executing.

from watchpoints import watch

def func(var):
    var["a"] = 1

a = {}

For example, this code prints:

====== Watchpoints Triggered ======

Call Stack (most recent call last):

  <module> (
> func(a)
  func (
> var["a"] = 1
{'a': 1}

The watch function can monitor more than a variable. It can also monitor the attributes and an element of a dictionary or list.

from watchpoints import watch

class MyObj:
    def __init__(self):
        self.a = 0

obj = MyObj()
d = {"a": 0}
watch(obj.a, d["a"]) # Yes you can do this
obj.a = 1 # Trigger
d["a"] = 1 # Trigger

This could help you narrow down to some specific objects that you are interested in.

If you are not happy about the format of the output, you can customize it. Just define your own callback function:

watch(a, callback=my_callback)

# Or set it globally


You can even bring up pdb when the trigger is hit:


This behaves similarly to breakpoint(), giving you a debugger-like experience.

If you don’t want to import the function in every single file, you can make it global by using install function:

watch.install() # or watch.install("func_name") and use it as func_name()

Personally, I think the coolest thing about watchpoints is its intuitive usage. Are you interested in some data? Just "watch" it, and you'll know when your variable changes.

Try watchpoints

I developed and maintain watchpoints on GitHub, and have released it under the licensed under Apache 2.0. Install it and use it, and of course contribution is always welcome.

magnifying glass on computer screen, finding a bug in the code

Troubleshoot your code with the GNU Debugger. Download our new cheat sheet.
Dump truck rounding a turn in the road

Check out Microsoft's open source tool for getting process information.
Colorful sound wave graph

VizTracer traces concurrent Python programs to help with logging, debugging, and profiling.


About the author

Tian Gao - Author of VizTracer