run-llama / create-llama

The easiest way to get started with LlamaIndex
MIT License
805 stars 98 forks source link

Generate log with parameters when calling create-llama #177

Open vaibhavkumbhar4 opened 1 month ago

vaibhavkumbhar4 commented 1 month ago

In the web chat application, we are not observing citations in the responses, which were present in the past but are missing in the most recent version.

marcusschiesser commented 1 month ago

@vaibhavkumbhar4 what's the configuration used? Can you send a screenshot like this:

image
balasubramaniam-ramasamy commented 1 month ago

Hi @marcusschiesser When the npx create-llamais executed, do we log the user input somewhere? If not, is it possible to log in a file, so this combination can be retrieved, later whenever needed. Sometimes, when we try to reproduce, we may miss some steps and won't be reproduced. Thanks!

marcusschiesser commented 1 month ago

@balasubramaniam-ramasamy

log the user input somewhere which user input are you talking about?

balasubramaniam-ramasamy commented 1 month ago

@balasubramaniam-ramasamy

log the user input somewhere which user input are you talking about?

@marcusschiesser I meant ... the series of questions and answers during the generation of new app using create-llama question "what is your project name?" answer "my-app"

If we log these questions and users answers (inputs), it will be easy to traceback, in which combinations the run-llama gives error or issues - for both the end-users like me and resolvers like you.