README.md

---
dataset_info:
  features:
  - name: repo
    dtype: string
  - name: commit
    dtype: string
  - name: path
    dtype: string
  - name: lang
    dtype: string
  - name: license
    dtype: string
  - name: message
    dtype: string
  - name: old_code
    dtype: string
  - name: new_code
    dtype: string
  - name: diff_string
    dtype: string
  - name: n_added
    dtype: int64
  - name: n_removed
    dtype: int64
  - name: n_hunks
    dtype: int64
  - name: change_kind
    dtype: string
  splits:
  - name: test
    num_bytes: 3425951
    num_examples: 1000
  download_size: 1664793
  dataset_size: 3425951
configs:
- config_name: default
  data_files:
  - split: test
    path: data/test-*
---

# diff-xyz-noisy

A noisy version of [diff-xyz](https://huggingface.co/datasets/JetBrains-Research/diff-xyz/) dataset. 
This dataset is obtained from the original one by replacing alpha symbols with another alpha symbol using a random character mapping.
Components: the original file contents (`old_code`), the modified contents (`new_code`), and the unified diff between
them (`diff_string`).

This format enables evaluation of LLM capabilities on three code editing tasks:

- **Apply**: Given old code and diff, generate new code
- **Anti-Apply**: Given new code and diff, generate old code
- **Diff-Generation**: Given old and new code, generate the unified diff