Possibility to set constraints of list of dict data? #90

SebastianDahlin · 2025-02-07T12:24:30Z

SebastianDahlin
Feb 7, 2025

Hello,
I am trying to solve a product csp-problem with python constraint.
For the sake of it. Let's say that my data looks as following.

front_frame = [
{"size": [40, 50], "color": ["green"]},
{"size": [60], "color": ["blue", "red"]]
]

back_frame = [
{"size": [50, 60], "color": ["green"]},
{"size": [30, 40], "color": ["blue", "red"]]

problem = Problem()
problem.addVariable("frame", frame)
problem.addVariable("wheel", wheel)

I am looking for a way to express that the size of the front_frame must be equal to the back_frame.
As in:
if size = 30, no solution
if size = 40, then front_frame = {"size": [40, 50], "color": ["green"]}, back_frame = {"size": [30, 40], "color": ["blue", "red"]]
if size = 50, then front_frame = {"size": [40, 50], "color": ["green"]}, back_frame = {"size": [50, 60], "color": ["green"]}
if size = 60, then front_frame = {"size": [60], "color": ["blue", "red"]], back_frame = {"size": [50, 60], "color": ["green"]}

Would this be possible when using python-constraint?

Answered by fjwillemsen

Feb 27, 2025

Hi Sebastian, that's definitely possible. If possible I would recommend not nesting variables you want to constraint in dictionaries, because it complicates things - I've given an alternative below. However, let's first solve it the way you provided it.
We start by defining the two variables, the front and back frames.

from constraint import Problem
problem = Problem()
problem.addVariable("front", [
    {"size": [40, 50], "color": ["green"]},
    {"size": [60], "color": ["blue", "red"]}
])
problem.addVariable("back", [
    {"size": [50, 60], "color": ["green"]},
    {"size": [30, 40], "color": ["blue", "red"]}
])

Next, we add the constraint and get the solutions:

def equal_size(front, ba…

View full answer

fjwillemsen · 2025-02-27T12:27:18Z

fjwillemsen
Feb 27, 2025
Maintainer

Hi Sebastian, that's definitely possible. If possible I would recommend not nesting variables you want to constraint in dictionaries, because it complicates things - I've given an alternative below. However, let's first solve it the way you provided it.
We start by defining the two variables, the front and back frames.

from constraint import Problem
problem = Problem()
problem.addVariable("front", [
    {"size": [40, 50], "color": ["green"]},
    {"size": [60], "color": ["blue", "red"]}
])
problem.addVariable("back", [
    {"size": [50, 60], "color": ["green"]},
    {"size": [30, 40], "color": ["blue", "red"]}
])

Next, we add the constraint and get the solutions:

def equal_size(front, back):
    return any(size in front["size"] for size in back["size"])

problem.addConstraint(equal_size, ("front", "back"))
print(problem.getSolutions())

Running this results in the following list, corresponding to the answer you provided:

[{'back': {'size': [30, 40], 'color': ['blue', 'red']}, 'front': {'size': [40, 50], 'color': ['green']}}, 
{'back': {'size': [50, 60], 'color': ['green']}, 'front': {'size': [60], 'color': ['blue', 'red']}}, 
{'back': {'size': [50, 60], 'color': ['green']}, 'front': {'size': [40, 50], 'color': ['green']}}]

While this solves your issue, it is a bit more cumbersome because you're already providing half of the solution yourself and doesn't get you the full benefit of the solver. The more natural approach is to think of constraint solvers as defining all possible combinations of values (a Cartesian product) and using the constraints to get the combinations that are actually possible. In your case that means a three-dimensional grid of frame, size, and color, something like this:

problem = Problem()
problem.addVariable("frame", ["front", "back"])
problem.addVariable("size", [30, 40, 50, 60])
problem.addVariable("color", ["green", "blue", "red"])

On which you define the following constraints:

def actual_values(frame, size, color):
    if frame == "front":
        if color == "green":
            return size == 40 or size == 50
        else:
            return size == 60
    else:
        if color == "green":
            return size == 50 or size == 60
        else:
            return size == 30 or size == 40

problem.addConstraint(actual_values, ("frame", "size", "color"))

Of course you can then repackage the solutions into your dictionary format.
I hope this provides an answer to your question, feel free to ask further questions or reopen if necessary.

4 replies

SebastianDahlin Feb 27, 2025
Author

Thanks you for your descriptive reply! This definitely solves it.
Just a follow up question. Is using the function actual_values really less cumbersome? Am I missing something.
Say that you for example add two more full rows of data in the front, and change some values in the back. Would not we have to rewrite the actual_values function in that case? And the .py file would need to be updated. Compared to loading json data.

fjwillemsen Jul 16, 2025
Maintainer

Hi Sebastian, apologies for the late reply! Yes, you are right that this is quite cumbersome. That is because of the nested way of handling data in your case, python-constraint isn't designed with that exact use case in mind. How would you ideally express this, and what do you mean by loading JSON data?

SebastianDahlin Jul 18, 2025
Author

Thank you very much for reaching out :)
I see now that I made a mistake when explaining the foundation problem.
But the gist is that it would be great to be able to set a constraint, that could automatically evaluate when the foundation data changes.
What I mean about "loading JSON data" is that I am trying to isolate the data from the logic. By putting the data in say a JSON file I can give all rows of possible frames and wheels. Then in the logic, as in the python file, we could put rules that could evaluate automatically.
So, in the best of worlds I would like to be able to say something in the lines of:
frame.size == wheel.size
Note that "size" is a list.
Happy to discuss this further. Hope my explanation is understandable.

fjwillemsen Jul 22, 2025
Maintainer

Thank you for clarifying! If I understand correctly, you would like python-constraint to support re-evaluating a problem when the input data (or file in this case) changes? This is currently not supported and is not part of the goals or roadmap, as it should be relatively easy to achieve yourself by watching the file for changes and just creating a new Problem for each new instance, and is not a core part of the CSP-solver that python-constraint is.

As for the problem specification, while I understand specifying it as frame.size == wheel.size would be ideal, the issue with this is that python-constraint does not natively support composite types (like lists, dicts, tuples, sets), just primitives (like strings, ints, floats, booleans). That is why your problem must be reformulated, for example like I showed above, to allow python-constraint to work optimally, and why something like size will not exist.
If you would like to have this functionality, feel free to submit a pull request or (likely more efficient) create a new package that wraps around python-constraint to achieve this.

I hope this answers your questions, feel free to ask more or suggest ideas.

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Possibility to set constraints of list of dict data? #90

Uh oh!

{{title}}

Uh oh!

Replies: 1 comment 4 replies

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{editor}}'s edit

{{editor}}'s edit

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{editor}}'s edit

{{editor}}'s edit

Uh oh!

Select a reply

Uh oh!

Possibility to set constraints of list of dict data? #90

Uh oh!

SebastianDahlin Feb 7, 2025

Replies: 1 comment · 4 replies

Uh oh!

fjwillemsen Feb 27, 2025 Maintainer

Uh oh!

SebastianDahlin Feb 27, 2025 Author

Uh oh!

Uh oh!

fjwillemsen Jul 16, 2025 Maintainer

Uh oh!

SebastianDahlin Jul 18, 2025 Author

Uh oh!

Uh oh!

fjwillemsen Jul 22, 2025 Maintainer

SebastianDahlin
Feb 7, 2025

Replies: 1 comment 4 replies

fjwillemsen
Feb 27, 2025
Maintainer

SebastianDahlin Feb 27, 2025
Author

fjwillemsen Jul 16, 2025
Maintainer

SebastianDahlin Jul 18, 2025
Author

fjwillemsen Jul 22, 2025
Maintainer