Release 0.0.2

[eyurtsev]

List of issues to resolve for 0.0.2 release:

Backend:

• Update README.md to show new changes since release 0.0.1 (+ API changes) H #88
• Support different users (user id from cookie) H #43
• Add sharing endpoints H #35
• Swap to with_structured_extraction for extraction chain H #32
• Fix docker compose build issues H #33
• Add ability to limit file size + number of pages for extraction H #49
• Add ability to choose other models #37
• Add endpoint to serve UI assets H #57
• Limit by number of requests to LLM rather than by number of pages (a page has no limit on amount of content in a PDF) H #75
• Add model metadata to configuration endpoint #93
• Improve quality of suggest endpoints L
• Add patch/put endpoint to allow updating the system message for the extractor (used from playground page) L
• Convert relevant code to async (add asyncpg) L
• Bump to pydantic 2 and add langchain_mistralai L
• Make a release

UI:

• Deploy frontend to production #28 -- not needed instead building assets and serving by backend
• Add frontend build to docker compose for dev #29
• Support Different Users via a cookie H
• AJV should support format in json schema H #46
• Resolve with credentials with axios client H #60

Code cleanliness:

• Migrate frontend to next.js from vite #30
• Remove daisy ui dependency (used primarily for file upload in playground component) L
• Clean up dead code (lint, add missing type annotations etc) L
• Refactor code that looks like it was written by a python engineer to code that looks like it was written by a professional L
• Clean up dead dependencies L #79

Side bar:

• UI: Support sharing H #40
• Support renaming models L
• When deleting active extractor navigate to /new L
• Improve the currently selected extractor L
• Make the side bar hide / show L

Playground:

• UI: add ability to limit file by size H #39
• Ability to use another model H #73
• Add ability to use shared extractor H #55
• Add support for configuring system message M
• Add ability to administer examples for a given extractor M

Create extractor page:

• UI disable inputs if user hasn't typed any text M
• Add suggest cards for schemas -> clicking should populate the text input M
• (Maybe): Add button to fix the openapi schema M

Results page:

• Fix table rendering when we extracted entities with nested attributes H #61
• Add copy and paste into clipboard of CSV + JSON format L
• Add tab for viewing extraction as JSON L #65

[eyurtsev]

Released. closing for now