Dataset Card for liaozhai-zhiyi

Description

Summary

This dataset contains 493 stories from the book Liaozhai Zhiyi (also known as Strange Tales from a Chinese Studio), a collection of supernatural tales written by Pu Songling during the Qing dynasty. The stories were exported from Wikisource and processed into a structured format.

Languages

Traditional Chinese

Structure

Data Instances

Each instance in the dataset represents a single story from Liaozhai Zhiyi, with the following fields:

• id: A unique identifier for the story (integer)
• volume: The volume number in which the story appears (integer)
• order_in_volume: The order of the story within its volume (integer)
• title: The title of the story (string)
• content: The full text content of the story (string)

Dataset Creation

Curation Rationale

This dataset was created to make the stories from Liaozhai Zhiyi more accessible for computational analysis and natural language processing tasks. The book is an important work in Chinese literature and provides insights into Qing dynasty culture and folklore.

Source Data

Data Collection and Normalization

Source: 聊齋志異

The source data was obtained from Wikisource using the "Download" / "Export" function to download the book in plaintext format. A Python script was then used to extract the stories.

Annotations

This dataset does not contain any additional annotations beyond the structure provided (id, volume, order, title, and content).

Considerations

Social Impact

This dataset provides easier access to an important work of Chinese literature, which can facilitate research in literature, history, and cultural studies. However, users should be aware that the stories reflect historical attitudes and may contain outdated or offensive content.

Biases

The stories in Liaozhai Zhiyi reflect the cultural norms and beliefs of Qing dynasty China. They may contain biases related to gender, social class, ethnicity, and supernatural beliefs that were prevalent at the time of writing.

Other Known Limitations

The dataset is based on a single edition from Wikisource and does not reflect variations or alternate versions of the stories.

License

This dataset is licensed under Creative Commons Attribution-ShareAlike 4.0 International (CC BY-SA 4.0), in accordance with the Wikisource license for the source material.