Allow lazy python values

[evinism]

This should allow us to not recursively convert the whole of the data parameter to values, and should mostly be transparent from both the producer and internal parts of MistQL.

There are two main changes here:

1. Moving builtins to utilize dedicated/dunder methods for indexing, sizing, and a few other operations.
2. Creating a lazy value type that dynamically builds the underlying data based on access patterns

I believe there are some pathological cases in which this could technically add compute, but they're very rare and have a bounded performance hits.

I also need to consist-ify the type to return in the dunder/special methods.

[netlify]

✅ Deploy Preview for mistql canceled.

Name	Link
🔨 Latest commit	e843cf4
🔍 Latest deploy log	https://app.netlify.com/projects/mistql/deploys/68d4662a28eb2a00076fd5d6

[Copilot]

Pull Request Overview

This PR introduces lazy evaluation for Python values in MistQL to optimize performance by deferring conversion of nested data structures until accessed. The main goal is to avoid recursively converting entire data parameters to RuntimeValues upfront.

Key Changes

• Added lazy evaluation option to RuntimeValue creation with new LazyRuntimeValue class
• Modified builtins to use dunder methods (len, iter) and dedicated methods (access, index) for operations
• Updated MistQLInstance to support lazy evaluation configuration

Reviewed Changes

Copilot reviewed 6 out of 6 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
py/mistql/runtime_value.py	Core implementation of lazy evaluation with new LazyRuntimeValue class and dunder methods
py/mistql/builtins.py	Updated builtin functions to use new dunder methods instead of direct .value access
py/mistql/instance.py	Added lazy parameter to MistQLInstance constructor with default True
py/mistql/gardenwall.py	Modified input_garden_wall to accept lazy parameter
py/tests/test_shared.py	Updated tests to use MistQLInstance and test both lazy and non-lazy modes
py/tests/test_lazy_value.py	Comprehensive test suite for lazy value functionality

---

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

[Copilot]

Pull Request Overview

Copilot reviewed 6 out of 6 changed files in this pull request and generated 3 comments.

---

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

[Copilot]

This line creates a non-lazy RuntimeValue even when the parent is lazy, which defeats the purpose of lazy evaluation. It should pass the lazy flag: return RuntimeValue.of(self.value[index], lazy=True)

Suggested change

	return RuntimeValue.of(self.value[index])
	else:
	return RuntimeValue.of(self.value[index:index_two])
	return RuntimeValue.of(self.value[index], lazy=self.is_lazy())
	else:
	return RuntimeValue.of(self.value[index:index_two], lazy=self.is_lazy())

[Copilot]

This line creates a non-lazy RuntimeValue for slices even when the parent is lazy. It should pass the lazy flag: return RuntimeValue.of(self.value[index:index_two], lazy=True)

Suggested change

	return RuntimeValue.of(self.value[index:index_two])
	return RuntimeValue.of(self.value[index:index_two], lazy=self.is_lazy())

[Copilot]

This line in the non-lazy path is missing the lazy parameter. It should be RuntimeValue.of(value[key], lazy) to maintain consistency with the lazy parameter passed to the parent.

Suggested change

	{key: RuntimeValue.of(value[key]) for key in value},
	{key: RuntimeValue.of(value[key], lazy) for key in value},

[evinism]

Preliminary tests show substantial performance gains (~6x over a simple query on nobel prizes dataset)