gomyway1216 / rag

0 stars 2 forks source link

Feature: Enhance our RAG to learn personality #5

Closed gomyway1216 closed 3 weeks ago

gomyway1216 commented 1 month ago

This is one of the end goal of this RAG. We want to make it as a chatbot that provides feedback based on your personality. We want to get the personal feedback instead of getting general suggestion by AI.

However, it is very cumbersome to type the same information about you every time you ask question. Additionally, that can get super long and might be everywhere in your note, blog, etc.. Ideally, we want to store it in one place efficiently so that the AI can understand it and give the best feedback to you.

Currently, I only created a single endpoint that provide takes the text. But we don't know if storing a long paragraphs into one text. We want to know what would be the best way to store it so that AI can access the important information. be aware of the context.

Here are some searches to start with.

https://www.google.com/search?q=rag+%E6%96%87%E5%AD%97%E6%95%B0+%E3%83%96%E3%83%AD%E3%82%B0&rlz=1C5CHFA_enUS1089US1089&oq=rag+%E6%96%87%E5%AD%97%E6%95%B0+%E3%83%96%E3%83%AD%E3%82%B0&gs_lcrp=EgZjaHJvbWUyBggAEEUYOTIGCAEQRRg7MgoIAhAAGIAEGKIE0gEHMTcwajBqN6gCALACAA&sourceid=chrome&ie=UTF-8

https://www.google.com/search?q=rag+%E6%96%87%E5%AD%97%E6%95%B0+%E3%83%96%E3%83%AD%E3%82%B0%E3%80%80%E3%83%81%E3%83%A3%E3%83%B3%E3%82%AF&sca_esv=c8ddd2c726ef673e&sca_upv=1&rlz=1C5CHFA_enUS1089US1089&ei=PDDjZs_VC6yr0PEPhLrB8Qc&ved=0ahUKEwjP44n2gb6IAxWsFTQIHQRdMH4Q4dUDCBA&uact=5&oq=rag+%E6%96%87%E5%AD%97%E6%95%B0+%E3%83%96%E3%83%AD%E3%82%B0%E3%80%80%E3%83%81%E3%83%A3%E3%83%B3%E3%82%AF&gs_lp=Egxnd3Mtd2l6LXNlcnAiJnJhZyDmloflrZfmlbAg44OW44Ot44Kw44CA44OB44Oj44Oz44KvMgoQABiwAxjWBBhHMgoQABiwAxjWBBhHMgoQABiwAxjWBBhHMgoQABiwAxjWBBhHMgoQABiwAxjWBBhHMgoQABiwAxjWBBhHMgoQABiwAxjWBBhHMgoQABiwAxjWBBhHMgoQABiwAxjWBBhHMgoQABiwAxjWBBhHSIBWUABYAHADeAGQAQCYAQCgAQCqAQC4AQPIAQCYAgOgAgeYAwCIBgGQBgqSBwEzoAcA&sclient=gws-wiz-serp

https://fintan.jp/page/10301/ https://zenn.dev/knowledgesense/articles/9303b94ea2c4eb https://zenn.dev/khisa/articles/9c48af5a1f417b https://note.com/izai/n/n0594d48470f1 https://zenn.dev/knowledgesense/articles/71e3d2d2ff3858 https://tech.explaza.jp/entry/2023/12/04/162659

Please provide the analysis and documented.

huyfififi commented 3 weeks ago

Can we break down this ticket into smaller/actionable items? For example, we can create an issue to improve the chunking algorithm in the "learn" API. Issues for PoC are also fine