housekeeping

Cargo.lock

@@ -2351,6 +2351,57 @@ version = "2.0.6"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "f4c7245a08504955605670dbf141fceab975f15ca21570696aebe9d2e71576bd"
 
+[[package]]
+name = "inference-engine"
+version = "0.1.0"
+dependencies = [
+ "ab_glyph",
+ "accelerate-src",
+ "anyhow",
+ "axum 0.7.9",
+ "bindgen_cuda",
+ "byteorder",
+ "candle-core",
+ "candle-datasets",
+ "candle-flash-attn",
+ "candle-nn",
+ "candle-onnx",
+ "candle-transformers",
+ "clap",
+ "cpal",
+ "csv",
+ "cudarc",
+ "either",
+ "enterpolation",
+ "half",
+ "hf-hub",
+ "image",
+ "imageproc",
+ "intel-mkl-src",
+ "memmap2",
+ "num-traits",
+ "palette",
+ "pdf2image",
+ "pyo3",
+ "rand 0.9.1",
+ "rayon",
+ "reborrow",
+ "rubato",
+ "safetensors",
+ "serde",
+ "serde_json",
+ "symphonia",
+ "tokenizers",
+ "tokio",
+ "tower 0.4.13",
+ "tower-http 0.5.2",
+ "tracing",
+ "tracing-chrome",
+ "tracing-subscriber",
+ "utoipa",
+ "uuid",
+]
+
 [[package]]
 name = "instant"
 version = "0.1.13"
@@ -2620,57 +2671,6 @@ version = "0.8.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "241eaef5fd12c88705a01fc1066c48c4b36e0dd4377dcdc7ec3942cea7a69956"
 
-[[package]]
-name = "local_inference_engine"
-version = "0.1.0"
-dependencies = [
- "ab_glyph",
- "accelerate-src",
- "anyhow",
- "axum 0.7.9",
- "bindgen_cuda",
- "byteorder",
- "candle-core",
- "candle-datasets",
- "candle-flash-attn",
- "candle-nn",
- "candle-onnx",
- "candle-transformers",
- "clap",
- "cpal",
- "csv",
- "cudarc",
- "either",
- "enterpolation",
- "half",
- "hf-hub",
- "image",
- "imageproc",
- "intel-mkl-src",
- "memmap2",
- "num-traits",
- "palette",
- "pdf2image",
- "pyo3",
- "rand 0.9.1",
- "rayon",
- "reborrow",
- "rubato",
- "safetensors",
- "serde",
- "serde_json",
- "symphonia",
- "tokenizers",
- "tokio",
- "tower 0.4.13",
- "tower-http 0.5.2",
- "tracing",
- "tracing-chrome",
- "tracing-subscriber",
- "utoipa",
- "uuid",
-]
-
 [[package]]
 name = "lock_api"
 version = "0.4.12"

Cargo.toml

@@ -1,7 +1,7 @@
 [workspace]
 members = [
     "crates/agent-server",
-    "crates/local_inference_engine",
+    "crates/inference-engine",
 ]
 
 resolver = "2"

README.md

@@ -1,13 +1,17 @@
 # open-web-agent-rs
 
-A Rust-based web agent with an embedded OpenAI-compatible inference server (supports Gemma models only).
+A Rust-based web agent with an embedded OpenAI-compatible inference server (supports Gemma models only). It is packaged and deployed as a container.
 
 ## Project Structure
 
 This project is organized as a Cargo workspace with the following crates:
 
-- `agent-server`: The main web agent server
-- `local_inference_engine`: An embedded OpenAI-compatible inference server for Gemma models
+- `crates`
+  - [agent-server](crates/agent-server): The main web agent server
+  - [inference-engine](crates/inference-engine): An embedded OpenAI-compatible inference server for Gemma models
+- `packages`
+  - [genaiscript](packages/genaiscript): GenaiScript scripts  
+  - [genaiscript-rust-shim](packages/genaiscript-rust-shim): An embedded OpenAI-compatible inference server for Gemma models
 
 ## Acknowledgements
 
@@ -61,9 +65,9 @@ flowchart LR
     end
 
     %% ─────────────── Local inference ───────────────
-    subgraph INFERENCE["Local Inference Engine"]
+    subgraph INFERENCE["Inference Engine"]
         direction TB
-        LIE["Local Inference Engine"]:::core -->|loads| MODELS["Gemma Models"]:::model
+        LIE["Inference Engine"]:::core -->|loads| MODELS["Gemma Models"]:::model
         LIE -->|exposes| OPENAI_API["OpenAI‑compatible API"]:::api
         MODELS -->|runs on| ACCEL
 
@@ -121,12 +125,12 @@ flowchart LR
 
 ## Running the Project
 
-### Local Inference Engine
+### Inference Engine
 
 To run the local inference engine:
 
 ```bash
-cd crates/local_inference_engine
+cd crates/inference-engine
 cargo run --release -- --server
 ```
 
@@ -159,5 +163,5 @@ To build a specific crate:
 ```bash
 cargo build -p agent-server
 # or
-cargo build -p local_inference_engine
+cargo build -p inference-engine
 ```

crates/agent-server/src/agents/deep_research.rs

@@ -13,6 +13,7 @@ mod tests {
     use std::fmt::Debug;
 
     #[tokio::test]
+    #[ignore]
     async fn test_deepresearch() {
         // a really provocative question for research that generally yields infinite complexity with each run
         let input = "What is a life of meaning?";

crates/agent-server/src/agents/search.rs

@@ -12,7 +12,8 @@ mod tests {
     use std::fmt::Debug;
     use crate::agents::search::agent;
 
-    #[tokio::test] 
+    #[tokio::test]
+    #[ignore]
     async fn test_search_execution() {
         let input = "Who won the 2024 presidential election?";
 

crates/inference-engine/Cargo.toml

@@ -1,5 +1,5 @@
 [package]
-name = "local_inference_engine"
+name = "inference-engine"
 version = "0.1.0"
 edition = "2021"
 

crates/inference-engine/README.md

@@ -1,4 +1,4 @@
-# Local Inference Engine
+# @open-web-agent-rs/inference-engine
 
 A Rust-based inference engine for running large language models locally. This tool supports both CLI mode for direct text generation and server mode with an OpenAI-compatible API.
 
@@ -30,7 +30,7 @@ A Rust-based inference engine for running large language models locally. This to
 
 2. Build the local inference engine:
    ```bash
-   cargo build -p local_inference_engine --release
+   cargo build -p inference-engine --release
    ```
 
 ## Usage
@@ -40,7 +40,7 @@ A Rust-based inference engine for running large language models locally. This to
 Run the inference engine in CLI mode to generate text directly:
 
 ```bash
-cargo run -p local_inference_engine --release -- --prompt "Your prompt text here" --which 3-1b-it
+cargo run -p inference-engine --release -- --prompt "Your prompt text here" --which 3-1b-it
 ```
 
 #### CLI Options
@@ -62,7 +62,7 @@ cargo run -p local_inference_engine --release -- --prompt "Your prompt text here
 Run the inference engine in server mode to expose an OpenAI-compatible API:
 
 ```bash
-cargo run -p local_inference_engine --release -- --server --port 3777 --which 3-1b-it
+cargo run -p inference-engine --release -- --server --port 3777 --which 3-1b-it
 ```
 
 This starts a web server on the specified port (default: 3777) with an OpenAI-compatible chat completions endpoint.

crates/inference-engine/openai-api-test.js

@@ -173,4 +173,4 @@
 })();
 
 console.log("\nAll test requests have been sent. Check the server logs for more details.");
-console.log("To run the server, use: cargo run --bin local_inference_engine -- --server");
+console.log("To run the server, use: cargo run --bin inference-engine -- --server");

crates/inference-engine/src/main.rs

@@ -246,7 +246,7 @@ async fn chat_completions(
         StatusCode::BAD_REQUEST,
         Json(serde_json::json!({
             "error": {
-                "message": "The OpenAI API is currently not supported due to compatibility issues with the tensor operations. Please use the CLI mode instead with: cargo run --bin local_inference_engine -- --prompt \"Your prompt here\"",
+                "message": "The OpenAI API is currently not supported due to compatibility issues with the tensor operations. Please use the CLI mode instead with: cargo run --bin inference-engine -- --prompt \"Your prompt here\"",
                 "type": "unsupported_api"
             }
         })),
@@ -292,6 +292,7 @@ use candle_core::{DType, Device, MetalDevice, Tensor};
 use candle_nn::VarBuilder;
 use candle_transformers::generation::LogitsProcessor;
 use hf_hub::{Repo, RepoType, api::sync::Api};
+use serde_json::json;
 use tokenizers::Tokenizer;
 use crate::token_output_stream::TokenOutputStream;
 use crate::utilities_lib::device;
@@ -596,7 +597,7 @@ impl TextGeneration {
                 logits
             } else {
                 let start_at = tokens.len().saturating_sub(self.repeat_last_n);
-
+ 
                 // Manual implementation of repeat penalty to avoid type conflicts
                 let mut logits_vec = logits.to_vec1::<f32>()?;
 

crates/inference-engine/src/server.rs

@@ -67,7 +67,7 @@ pub async fn chat_completions(
                 StatusCode::BAD_REQUEST,
                 Json(serde_json::json!({
                     "error": {
-                        "message": "The OpenAI API is currently not supported due to compatibility issues with the tensor operations. Please use the CLI mode instead with: cargo run --bin local_inference_engine -- --prompt \"Your prompt here\"",
+                        "message": "The OpenAI API is currently not supported due to compatibility issues with the tensor operations. Please use the CLI mode instead with: cargo run --bin inference-engine -- --prompt \"Your prompt here\"",
                         "type": "unsupported_api"
                     }
                 })),

crates/inference-engine/tests/model_tests.rs

@@ -1,4 +1,4 @@
-use local_inference_engine::model::{Model, Which};
+use inference_engine::model::{Model, Which};
 
 #[cfg(test)]
 mod tests {

crates/inference-engine/tests/text_generation_tests.rs

@@ -1,11 +1,8 @@
-use local_inference_engine::text_generation::TextGeneration;
-use local_inference_engine::model::{Model, Which};
-use local_inference_engine::token_output_stream::TokenOutputStream;
-use tokenizers::Tokenizer;
-use candle_core::{DType, Device, Tensor};
-use candle_transformers::generation::LogitsProcessor;
 use anyhow::Result;
-use std::sync::Arc;
+use candle_transformers::generation::LogitsProcessor;
+use inference_engine::model::Which;
+use inference_engine::token_output_stream::TokenOutputStream;
+use tokenizers::Tokenizer;
 
 #[cfg(test)]
 mod tests {

crates/inference-engine/tests/token_output_stream_tests.rs

@@ -1,4 +1,4 @@
-use local_inference_engine::token_output_stream::TokenOutputStream;
+use inference_engine::token_output_stream::TokenOutputStream;
 use tokenizers::Tokenizer;
 use std::path::PathBuf;
 use anyhow::Result;

package.json

@@ -7,6 +7,8 @@
   "private": true,
   "scripts": {
     "clean": "rm -rf .genaiscript && rm -rf dist && rm -rf node_modules && rm -rf open-web-agent-rs && rm -rf target && rm -rf packages/genaiscript-rust-shim/dist",
+    "build": "(cd packages/genaiscript-rust-shim && bun run buildShim && bun run setupDev && cargo build -p agent-server)",
+    "mcp-inspector": "bunx @modelcontextprotocol/inspector",
     "dev": "bun i && ./killport.js 3006 && bun run build && cargo watch -x 'run -p agent-server'",
     "start": "docker compose up --build",
     "ai:search": "genaiscript run packages/genaiscript/genaisrc/web-search.genai.mts --vars USER_INPUT='who won the 2024 election?'",
@@ -16,8 +18,6 @@
     "ai:url:scrape": "npx genaiscript run packages/genaiscript/genaisrc/web-scrape.genai.mts --vars USER_INPUT='{\"url\":\"https://www.time4learning.com/homeschool-curriculum/high-school/eleventh-grade/math.html\",\"query\":\"What is on this page?\", \"action\": \"scrape\"}'",
     "prod:logs": "fly logs",
     "test-http": "test/test-search.ts",
-    "mcp-inspector": "bunx @modelcontextprotocol/inspector",
-    "build": "(cd packages/genaiscript-rust-shim && bun run buildShim && bun run setupDev && cargo build -p agent-server)"
   },
   "dependencies": {
     "@modelcontextprotocol/inspector": "^0.14.0"