Inspiration
We work in an industry that handles a lot of text file and string parsing, making the isolation of infomation important. The basic pattern is that the information needs to be identified, and acquired in a flexible but deterministic way. We already have a lot of great tools for this purpose, but what if we could make a schema that makes this information easy to identify in contexts like large files or semi-cryptic formats like banking or health data files?
What it does
This project takes a schema in JSON that defines a bunch of named regex capture groups without having to know how to put together an entire regex line. You'll be able to upload a schema.json, a data file and see the lines that the parser isolated.
How we built it
Boris built the front-end in React and Vite. Shaun developed a simple schema parser library. Anup is working on an interface to suggest schemas for files based on what users are looking for.
Challenges we ran into
As a team, we're fairly new to each other and spent the first half of the project brainstorming. We work with a lot of different tools. Shaun may have dismissed some fairly decent ideas too soon.
Accomplishments that we're proud of
We got a user interface and parser library together in short order despite our busy lives.
What we learned
Working styles and that we might like to work together in the future.
What's next for Reverse-Schema for Text Documents
This project is an extreme simplification of an overall parser language for extracting information from documents. The language will handle repeated blocks and hierarchies much like handlebars, but the plan is to use the language as a declarative parser and not a populator for the files.
Log in or sign up for Devpost to join the conversation.