geoffsee/predict-otron-9001
1
0
Fork 0
mirror of https://github.com/geoffsee/predict-otron-9001.git synced 2025-09-08 22:46:44 +00:00
Code Issues Packages Projects Releases Wiki Activity

- Change default server host to localhost for improved security.

Browse Source 
- Increase default maximum tokens in CLI configuration to 256.
- Refactor and reorganize CLI
This commit is contained in:
geoffsee
2025-08-27 21:47:24 -04:00
parent 766d41af78
commit 719beb3791
20 changed files with 1703 additions and 490 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
crates/embeddings-engine/Cargo.toml
View File

@@ -24,3 +24,12 @@ tracing-subscriber = { version = "0.3", features = ["env-filter"] }
rand = "0.8.5"
async-openai = "0.28.3"
once_cell = "1.19.0"
[package.metadata.kube]
image = "ghcr.io/geoffsee/embeddings-service:latest"
replicas = 1
port = 8080
resources.cpu = "500m"
resources.memory = "256Mi"
#ingress.host = "my-service.example.com"
#env = { RUST_LOG = "info", DATABASE_URL = "postgres://..." }

2
crates/embeddings-engine/src/main.rs
View File

@@ -10,7 +10,7 @@ use std::env;
use tower_http::trace::TraceLayer;
use tracing;
const DEFAULT_SERVER_HOST: &str = "0.0.0.0";
const DEFAULT_SERVER_HOST: &str = "127.0.0.1";
const DEFAULT_SERVER_PORT: &str = "8080";
async fn root() -> &'static str {