n+1-1 in sum/prod

[odashi]

Currently the analyzer of sum and prod does not recognize the reducable addition and subtraction in the stop parameter of range:

It would be good if we could get $n$ rather than $n+1-1$ in most cases. This reduction would be safe because the operands of range must be integers, whose objects must be always associative and commutative in terms of operations $+$ and $-$.

[odashi]

Note that in general arithmetic is not associative even if all operands are numbers. For example:

So this treatment is available only when the function always takes integers.

[Casper-Guo]

Implemented here:

latexify_py/src/latexify/codegen/function_codegen.py

Lines 335 to 354 in c6c291d

	elt_str = self.visit(node.elt)
	target_str = self.visit(comp.target)
	args_str = [self.visit(arg) for arg in comp.iter.args]
	if len(args_str) == 1:
	lower_str = "0"
	upper_plus_1 = args_str[0]
	else:
	lower_str = args_str[0]
	upper_plus_1 = args_str[1]
	# Upper bound of range is exclusive. Try to numerically subtract it by 1.
	try:
	upper_plus_1_unwrapped = upper_plus_1[1:-1] # Remove { and }
	upper_str = str(int(upper_plus_1_unwrapped) - 1)
	except ValueError:
	upper_str = "{" + upper_plus_1 + " - 1}"
	# Used for: \sum_{target_str = lower_str}^{upper_str} elt_str
	return elt_str, target_str, lower_str, upper_str

Current behavior is caused by not being able to convert n+1 to a numeric type.

The n+1 case can perhaps be treated as a special case since it is the most common but a more general solution might be preferred. Since reduction is only needed for n+x or n-x maybe we can explicitly check for these two cases.

[odashi]

@Casper-Guo I think the code above is an obsolete permalink. In the current implementation, FunctionCodegen introduced an additional analyzer that obtains subtree of range:

latexify_py/src/latexify/codegen/function_codegen.py

Line 334 in 071a12d

range_info = analyzers.analyze_range(node.iter)

We can implement a specific process if we detected range_info.stop is BinOp and the operator is Add or Sub.

[Casper-Guo]

We would like to work on this. Can you please assign me and @LakeBlair.

[odashi]

@Casper-Guo Thanks for taking your effort! Assigned to you anyway since you are the only person who attended this discussion.

[odashi]

Resolved by #115