lupyuen/lupyuen.org
1
0
Fork 0
mirror of https://github.com/lupyuen/lupyuen.github.io.git synced 2025-01-13 03:18:31 +08:00
Code Issues Projects Releases Packages Wiki Activity Actions 1

Update article

Browse source
This commit is contained in:
Lup Yuen Lee 2024-09-28 09:51:08 +08:00
parent e37c544a03
commit 05c4a1a890
2 changed files with 7 additions and 5 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

BIN
images/llm-quota.png Normal file
View file

Binary file not shown.
Side by side

After

Width:  |  Height:  |  Size: 187 KiB

12
src/llm.md
View file

@ -71,7 +71,7 @@ How did this happen? We find out...
_Our Bot gets really chatty and naggy. Why?_
That's because we programmed it with a __Long List of Requirements__ for PR Review, in Markdown Format: [main.rs](https://github.com/lupyuen/nuttx-pr-bot/blob/main/src/main.rs#L24-L66)
That's because we programmed it with a __Long List of Requirements__ for PR Review (pic above), in Markdown Format: [main.rs](https://github.com/lupyuen/nuttx-pr-bot/blob/main/src/main.rs#L24-L66)
<div style="margin-left: 5%; width: 90%; font-size:90%">
@ -228,7 +228,7 @@ let response_text =
.text.clone().unwrap();
```
__Gemini Response__ will be in Markdown Format...
__Gemini Response__ will be in Markdown Format (pic above)...
<div style="margin-left: 5%; width: 90%; font-size:90%">
@ -273,9 +273,11 @@ _But we didn't tell Gemini to respond in Markdown?_
That's the law of __"LLM See, LLM Do"__! We fed the input in Markdown Format, so it produced the output in Markdown Format too.
![TODO](https://lupyuen.github.io/images/llm-quota.png)
_Are we paying Google for the LLM?_
Right now we call the __Free Tier__ of Gemini Pro 1.5. Which limits us to [__50 LLM Requests__](TODO) per day.
Right now we call the __Free Tier__ of Gemini Pro 1.5. Which limits us to [__50 LLM Requests__](TODO) per day. (Pic above)
That's sufficient to review __50 Pull Requests__ for NuttX every day. (For both the [__NuttX Kernel Repo__](TODO) and the [__NuttX Apps Repo__](TODO))
@ -291,7 +293,7 @@ Thus we retry __up to 3 times__, in case the Pull Request has some problematic t
_How does our Bot know when to stop the retries?_
Our Bot uses __Emoji Reactions__ to tag each attempt. We'll see the Pull Request tagged with...
Our Bot uses __Emoji Reactions__ (pic above) to tag each attempt. We'll see the Pull Request tagged with...
|||
|----|---------------|
@ -326,7 +328,7 @@ async fn bump_reactions(issues: &IssueHandler<'_>, pr_id: u64, reactions: (Optio
_How do we fetch the Pull Request? And post the Review Comment?_
We call the GitHub API with the [__Octocrab Crate__](https://github.com/XAMPPRocky/octocrab).
We call the GitHub API with the [__Octocrab Crate__](https://github.com/XAMPPRocky/octocrab). (Pic above)
Here we fetch the __20 Newest Pull Requests__ that are Open: [main.rs](https://github.com/lupyuen/nuttx-pr-bot/blob/main/src/main.rs#L97-L115)