Helenlei/ai gpt costing

[leihelen]

Script that can be used to get the numbers needed to calculated the total costs for creating "cornellian says" summaries, depending on which model we use. For example, below are the results for if we wanted to create summaries for all courses that have at least 3 reviews.

[dti-github-bot]

[diff-counting] Significant lines: 151.

[qiandrewj]

This looks great, Helen! Code is streamlined and I like that we have different options to only look at courses with a minimum number of reviews - it'll be a huge help as we look into OpenAI costing. The one thing I can point out is in the '/costing' route it would be helpful to get the minimum from the request body and pass it in to the function. Great work overall :)

[jacquelinecai]

Great work Helen! The structure of the code is very easy to follow. I think something that might also help with the OpenAI costing would be to look at the average number of distinct words in each review.

[CLAassistant]

All committers have signed the CLA.

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you all sign our Contributor License Agreement before we can accept your contribution.
3 out of 4 committers have signed the CLA.

✅ wizhaaa
✅ leihelen
✅ qiandrewj
❌ jacquelinecai
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[wizhaaa]

This is amazing Helen - great to see you getting the pieces connected together and running the functions on our staging database.

Here are some more reference points which you can use to provide & calculate more concrete prices from https://openai.com/pricing

(4 characters = 1 token) and GPT-4 and GPT-3.5 turbo have some pretty drastic price points.

Currently discussing production-level database access with Sarah right now - looking like we may be able to get temporary read access first.