feat(text-splitters): add R programming language support

[gorkachea]

Description:

This PR adds support for the R programming language to LangChain's text splitters, enabling users to intelligently split R code while preserving semantic structure.

Changes:

• Added Language.R enum value to the Language enum
• Implemented R-specific separators in RecursiveCharacterTextSplitter.get_separators_for_language()
• Splits R code along:
  • Function definitions (<- function, = function)
  • Package loading (library(), require())
  • Control flow statements (if(), for(), while(), switch())
  • Data structure creation (data.frame(), list())
  • Standard separators (double newlines, single newlines, spaces)

Why This Matters:

R is widely used in data science, statistical analysis, machine learning, and bioinformatics. This change enables users working with R codebases to properly chunk R scripts for RAG applications and build AI assistants that understand R code structure.

Example Usage:

from langchain_text_splitters import Language, RecursiveCharacterTextSplitter

r_code = """
library(dplyr)

calculate_mean <- function(x) {
  mean(x, na.rm = TRUE)
}
"""

splitter = RecursiveCharacterTextSplitter.from_language(
    language=Language.R,
    chunk_size=100,
    chunk_overlap=0
)

chunks = splitter.split_text(r_code)
# Splits intelligently at library calls and function definitions

Testing:
Tested locally with various R code samples including package imports, function definitions (both <- and = syntax), control flow statements, and data structure creation. All separators work correctly with proper regex escaping.

Issue: Fixes #33824

Dependencies: None

[makkruo]

@gorkachea
Hello, there! I noticed that you submitted this PR last week, but so far, there have been no reviews. A few days ago, I also submitted a PR and worked on something similar to yours—contributing MySQL language text splitter rules for Langchain. I noticed that the scope of your changes in the project closely aligns with mine. Currently, my PR is stuck at the CodeQL scanning stage, and that workflow hasn’t even started yet. This has prevented me from requesting a review. I’m not sure if your PR is facing a similar problem or if this might be a bug in the text-splitters workflow of the project. All I can see is that you’ve done similar work and also haven’t requested a review. If you’re experiencing a similar problem or have any workarounds or ideas, I’d be very glad to discuss and exchange ideas with you. Wishing you a smooth merge of your contribution!

[gorkachea]

Hey @makkruo , thanks for the note! I’m seeing the same thing: no option to request a review and CodeQL hasn’t kicked off. From what I understand, external contributors can’t request reviewers directly, that’s something maintainers do when they triage PRs. I’ll keep an eye on this one and see if it gets picked up on their review board.

[gorkachea]

Hi @mdrxy ! Since you labeled the related issue earlier, just wanted to give you a quick heads-up that the PR implementing R support is ready and includes tests. No rush at all, if you have a moment to take a look or let me know if something should be adjusted, I’d be happy to update it. Thanks!