Alex Kuo

[Improvement] Short disclaimer about what kind of data people can query for, so they don't think the app is broken

1

There's a lot of data we want to integrate. But for now, it's mostly limited to queries about race, age, gender, population, and crime of zip codes/cities/counties/states. We should add...

Add middle step to prompt LM to return relevant tables to answer a given natural language query.

Create a dev database + frontend + backend for testing new features/datasets.

Prompt LM to use markdown to indicate code text. Currently, we take text before ';' and after ':' from the LM response to extract the SQL query.

Starting a discussion around datasets people want to see. We currently only have the following tables/columns: crime_by_city - city (TEXT) - violent_crime (DOUBLE_PRECISION) - murder_and_nonnegligent_manslaughter (DOUBLE_PRECISION) - rape (DOUBLE_PRECISION) -...

[Feature Request] Chat interface to progressively build on queries

2

We envision users progressively building on their queries, instead of asking 1 time questions. For example: 1. User asks "What are the 5 wealthiest cities in CA?" 2. CensusGPT responds...

[Improvement] Advanced SQL Generation (RESDSQL)

1

https://github.com/RUCKBReasoning/RESDSQL ^ This repo has an implementation of this paper: https://arxiv.org/pdf/2302.05965.pdf It details an advanced strategy for generating accurate SQL to answer a natural language query. As far as I...

[BYOD] support for different SQL dialects (MySQL)

1

Currently, the SQL statements for setting up table and enum metadata, for the language model, are written in PostgreSQL. We want to add support for MySQL as well. The functions...