Streamlined SCM Variable Access Methods & Mapped Out Counterfactual Queries

[ryansherby]

Proposed location for recursive querying of counterfactual worlds.

Modified do() method to check for valid counterfactual operations (same domain).

Streamlined variable access methods by allowing for index or value retrieval; e.g., scm.v.x, scm.v['x'], scm.v[0], scm.v

[kl2792]

It may also be good to add code examples (effectively, tests) to a separate file, perhaps a chapter5.ipynb notebook.

[kl2792]

Can we make this more general, to allow for arbitrary interventional expressions, which may use counterfactual symbols?

m = SymbolicSCM(f=dict(z=..., x=..., w=..., y=...), pu={...})
z_1, x_1, w_1, y_1 = m.do({x: 1}).v

# all of these could be valid
m.do({x: 1})
m.do({x: z ^ 1})
m.do({w: x_1})
m.do({w: x_1 | z})

[ryansherby]

How should I approach z ^ 1? It gives me an error about performing operations on sp.symbols vs int. I would normally implement a wrapper, but it happens before it is passed as a param.

[kl2792]

Oops. Let me correct that second do line to one or both of:

m.do({x: ~z})
m.do({w: x ^ z})

[ryansherby]

All expressions are now ingestible, but I'm not sure if I can guarantee meaningful results

[Copilot]

Pull Request Overview

This PR introduces a streamlined variable access system for Structural Causal Models (SCMs) and implements foundational components for counterfactual queries. The changes enhance the user interface for accessing SCM variables and establish infrastructure for handling counterfactual worlds.

• Implements enhanced variable access patterns allowing index, value, and attribute-based retrieval
• Adds core symbolic classes for probabilistic expressions and counterfactual variables
• Establishes framework for multi-world networks and counterfactual graph operations

Reviewed Changes

Copilot reviewed 36 out of 142 changed files in this pull request and generated 5 comments.

Show a summary per file

File	Description
src/citk/sympy_classes/variable.py	Defines Variable class with intervention tracking for counterfactual analysis
src/citk/return_classes/symbol_container.py	Implements SymbolContainer with multiple access patterns (index, attribute, value)
src/citk/scm/init.py	Updates SymbolicSCM with streamlined variable access and counterfactual support
src/citk/causal_graph/init.py	Adds CausalGraph class with d-separation and adjustment methods
src/citk/sympy_classes/pr.py	Implements Pr class for probabilistic expressions with LaTeX rendering

Comments suppressed due to low confidence (2)

src/citk/causal_graph/accessors.py:1

• The method accesses self._ctf_graphs but this attribute is not defined in the abstract base class. This will cause an AttributeError when called on classes that don't define this attribute.

from abc import ABC, abstractmethod

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[Copilot]

Missing import statement for sympy. The class inherits from sp.Expr but sympy is not imported as sp.

Suggested change

	import sympy as sp

[Copilot]

The function splits on spaces but the docstring indicates comma-separated names. Either change the split to use commas or update the docstring to reflect space separation.

Suggested change

	tup = [Variable(name.strip()) for name in names.split(" ")]
	tup = [Variable(name.strip()) for name in names.split(",")]

[Copilot]

[nitpick] Import statements should be grouped and sorted. Group standard library imports first, then third-party imports, then local imports, with blank lines between groups.

Suggested change

	import itertools
	import math
	from typing import Dict, List, Optional, Union
	import pandas as pd
	import sympy as sp
	from IPython.display import Latex
	from src.sympy_classes.pr import Pr
	from src.return_classes import SymbolContainer
	from src import CausalGraph
	from src.sympy_classes.summation import Summation
	from src.sympy_classes.variable import Variable
	import itertools
	import math
	from typing import Dict, List, Optional, Union
	import pandas as pd
	import sympy as sp
	from IPython.display import Latex
	from src import CausalGraph
	from src.return_classes import SymbolContainer
	from src.sympy_classes.pr import Pr
	from src.sympy_classes.summation import Summation
	from src.sympy_classes.variable import Variable

[Copilot]

Returning hash(output) as an integer is misleading since the method name suggests it evaluates a symbolic expression. Consider returning the actual evaluated value or renaming the method to reflect its hash-based behavior.

[Copilot]

The method calls k.update_interventions(x) for every variable in self.v, but interventions in x only apply to specific variables. Consider filtering to only update variables that are actually being intervened upon.