metadata

dataset_info:
  features:
    - name: text
      dtype: string
    - name: source
      dtype: string
  splits:
    - name: train
      num_bytes: 8452105
      num_examples: 100000
  download_size: 2718505
  dataset_size: 8452105
license: mit
task_categories:
  - text-classification
tags:
  - code
  - applescript
size_categories:
  - 100K<n<1M

Dataset Card for "applescript-lines-100k-non-annotated"

Description

Dataset of 100,000 unique lines of AppleScript code scraped from GitHub and GitHub Gists. The dataset has been de-duplicated, comments have been removed (both single and multi-line), and effort has been made to merge multi-line structures such as records into one (however, expect some variability in this regard).

The dataset is constructed as an intermediate step to a fully-annotated AppleScript dataset.

Each row has fields for text and source, with text being the raw text of the line and source being the file name and extension from which the line was obtained. Full source links have been omitted for anonymity.