kim-sha/text-to-socrata-sql

---
dataset_info:
features:
- name: question
dtype: string
- name: context
dtype: string
- name: answer
struct:
- name: plot
dtype: string
- name: query
dtype: string
splits:
- name: train
num_bytes: 21311
num_examples: 18
download_size: 12039
dataset_size: 21311
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
---

# Text to Socrata SQL Training Data (WIP)

Dataset repository for collecting training data composed of:
- **Input**: Natural language questions (`question`) about a specific table schema (`context`)
- **Output**: Corresponding SoQL queries (`answer > query`) and Python Plotly code snippets (`answer > plot`)

This will serve as training data for a future iteration of [`sql-sodabot-v1.0`](https://huggingface.co/kim-sha/sql-sodabot-v1.0#sql-sodabot-v10).

Dataset for collecting natural language questions and their corresponding SoQL queries and Python Plotly code snippets, containing three main features: question, context, and answer. The question and context are both string types, while the answer is a structure containing two parts: query and plot, both of which are string types. The dataset is divided into a training set (train) with 18 samples, totaling 21311 bytes.

原始信息汇总

数据集概述

数据集信息

• 特征:

  • question: 自然语言问题，数据类型为字符串。
  • context: 特定表模式的上下文，数据类型为字符串。
  • answer: 结构化数据，包含以下字段：
    • plot: Python Plotly代码片段，数据类型为字符串。
    • query: 对应的SoQL查询，数据类型为字符串。

• 数据分割:

  • train: 训练集，包含21311字节的数据和18个样本。

• 数据大小:

  • 下载大小: 12039字节
  • 数据集大小: 21311字节

配置

• 默认配置:
  • 数据文件路径: data/train-*