README.md

metadata

dataset_info:
  features:
    - name: problem_id
      dtype: string
    - name: source
      dtype: string
    - name: task_type
      dtype: string
    - name: in_source_id
      dtype: string
    - name: prompt
      dtype: string
    - name: golden_standard_solution
      dtype: string
    - name: verification_info
      dtype: string
    - name: metadata
      dtype: string
  splits:
    - name: train
      num_bytes: 6988481341
      num_examples: 69752
  download_size: 2821986433
  dataset_size: 6988481341
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*

ds_up = ds_debug.map(lambda x, idx: {"problem_id": f"swe_fixer_{idx}"}, with_indices=True)
ds_up = ds_up.map(lambda x: {"source": "internlm/SWE-Fixer-Train-Editing-CoT-70K", "task_type": "swe_fixer"})

num_proc = 16

# Function to format code files for display
def format_files(files):
    formatted = ""
    for file_info in files:
        formatted += f"## `{file_info['file']}`\n```\n{file_info['file content']}\n```\n\n"
    return formatted

ds_up = ds_up.map(lambda x: {"in_source_id": x["instance_id"]}, num_proc=num_proc)

# Format the prompt using the template
ds_up = ds_up.map(lambda x: {
    "prompt": prompt_template.format(
        issue_description=x['input']['input']['issue'],
        files=format_files(x['input']['input']['files to be modified'])
    )
}, num_proc=num_proc)

# Format the golden_standard_solution properly - use repr() to ensure it's a valid Python literal
ds_up = ds_up.map(lambda x: {"golden_standard_solution": repr({
    "edited code": x["output"]["edited code"]
})}, num_proc=num_proc)

ds_up = ds_up.map(lambda x: {"verification_info": repr({
    "input": x["input"]["input"],
    "output": x["output"]
})}, num_proc=num_proc)

# Format the metadata as a string representation of a dictionary
ds_up = ds_up.map(lambda x: {"metadata": repr({
    # "input": x["input"]["input"]
})}, num_proc=num_proc)

ds_up = ds_up.select_columns(["problem_id", "source", "task_type", "in_source_id", "prompt", "golden_standard_solution", "verification_info", "metadata"])