# PMC-OA Dataset

[中文文档](./README.zh.md)

- [PMC-OA Dataset](#pmc-oa-dataset)
  - [Daraset Structure](#daraset-structure)
  - [Sample](#sample)

## Daraset Structure

**PMC-OA** (seperated images, separated caption).
- `train.jsonl`: metafile of train set
- `valid.jsonl`: metafile of valid set
- `test.jsonl`: metafile of test set
- `images.zip`: images folder

**PMC-OA Beta** is dataset of (seperated images, whole caption). And it's structure is the same as PMC-OA.


## Sample

A row in `train.jsonl` is shown bellow,

```python
{
    "image": "PMC212319_Fig3_4.jpg",
    "caption": "A. Real time image of the translocation of ARF1-GFP to the plasma membrane ...",
    "pmcid": "PMC212319",
    "url_name": "1471-2121-4-13-3.jpg"
}
```

Explanation to each key

- image: path to the image
- caption: corresponding to the image
- pmcid: source paper's PMCID of the image
- url_name: image name in the web link