Local lm revamp

README.md

@@ -27,10 +27,10 @@ Add DSRs to your `Cargo.toml`:
 ```toml
 [dependencies]
 # Option 1: Use the shorter alias (recommended)
-dsrs = { package = "dspy-rs", version = "0.7.0" }
+dsrs = { package = "dspy-rs", version = "0.7.1" }
 
 # Option 2: Use the full name
-dspy-rs = "0.7.0"
+dspy-rs = "0.7.1"
 ```
 
 Or use cargo:
@@ -64,17 +64,15 @@ struct SentimentAnalyzer {
 
 #[tokio::main]
 async fn main() -> Result<()> {
-    let lm = LM::builder()
-        .api_key(std::env::var("OPENAI_API_KEY")?.into())
-        .config(
-            LMConfig::builder()
-                .model("gpt-4.1-nano".to_string())
-                .temperature(0.5)
-                .build(),
-        )
-        .build();
-
-    configure(lm, ChatAdapter);
+    // API key automatically read from OPENAI_API_KEY env var
+    configure(
+        LM::builder()
+            .model("gpt-4o-mini".to_string())
+            .temperature(0.5)
+            .build()
+            .await?,
+        ChatAdapter,
+    );
 
     // Create a predictor
     let predictor = Predict::new(SentimentAnalyzer::new());
@@ -154,13 +152,20 @@ let predict = Predict::new(MySignature::new());
 
 #### 4. **Language Models** - Configurable LM Backends
 ```rust
-// Configure with OpenAI
+// Configure with OpenAI (API key read from OPENAI_API_KEY env var)
 let lm = LM::builder()
-    .api_key(secret_key)
-    .model("gpt-4")
+    .model("gpt-4o-mini".to_string())
     .temperature(0.7)
     .max_tokens(1000)
-    .build();
+    .build()
+    .await?;
+
+// For local models (e.g., vLLM, Ollama)
+let lm = LM::builder()
+    .base_url("http://localhost:11434".to_string())
+    .model("llama3".to_string())
+    .build()
+    .await?;
 ```
 
 #### 5. **Evaluation** - Evaluating your Modules

crates/dspy-rs/Cargo.toml

@@ -1,7 +1,7 @@
 [package]
 name = "dspy-rs"
 authors = ["Herumb Shandilya <herumb@stanford.edu>"]
-version = "0.7.0"
+version = "0.7.1"
 edition = "2024"
 description = "A DSPy rewrite(not port) to Rust."
 readme = "../../README.md"
@@ -25,7 +25,7 @@ tokio = { version = "1.46.1", features = ["full"] }
 async-trait = "0.1.83"
 anyhow = "1.0.99"
 bon = "3.7.0"
-dsrs_macros = { version = "0.7.0", path = "../dsrs-macros" }
+dsrs_macros = { version = "0.7.1", path = "../dsrs-macros" }
 csv = { version = "1.3.1" }
 hf-hub = { version = "0.4.3", features = ["tokio"] }
 parquet = { version = "56.1.0" }

crates/dspy-rs/examples/01-simple.rs

@@ -10,8 +10,8 @@ cargo run --example 01-simple
 use anyhow::Result;
 use bon::Builder;
 use dspy_rs::{
-    ChatAdapter, Example, LM, LMConfig, Module, Predict, Prediction, Predictor, Signature,
-    configure, example, prediction,
+    ChatAdapter, Example, LM, Module, Predict, Prediction, Predictor, Signature, configure,
+    example, prediction,
 };
 
 #[Signature(cot)]
@@ -70,11 +70,11 @@ impl Module for QARater {
 #[tokio::main]
 async fn main() -> Result<()> {
     configure(
-        LM::new(LMConfig {
-            model: "openai:gpt-4o-mini".to_string(),
-            ..LMConfig::default()
-        })
-        .await,
+        LM::builder()
+            .model("openai:gpt-4o-mini".to_string())
+            .build()
+            .await
+            .unwrap(),
         ChatAdapter,
     );
 

crates/dspy-rs/examples/03-evaluate-hotpotqa.rs

@@ -62,7 +62,14 @@ impl Evaluator for QARater {
 
 #[tokio::main]
 async fn main() -> anyhow::Result<()> {
-    configure(LM::default(), ChatAdapter {});
+    configure(
+        LM::builder()
+            .model("openai:gpt-4o-mini".to_string())
+            .build()
+            .await
+            .unwrap(),
+        ChatAdapter {},
+    );
 
     let examples = DataLoader::load_hf(
         "hotpotqa/hotpot_qa",

crates/dspy-rs/examples/04-optimize-hotpotqa.rs

@@ -58,7 +58,14 @@ impl Evaluator for QARater {
 
 #[tokio::main]
 async fn main() -> anyhow::Result<()> {
-    configure(LM::default(), ChatAdapter {});
+    configure(
+        LM::builder()
+            .model("openai:gpt-4o-mini".to_string())
+            .build()
+            .await
+            .unwrap(),
+        ChatAdapter {},
+    );
 
     let examples = DataLoader::load_hf(
         "hotpotqa/hotpot_qa",

crates/dspy-rs/examples/05-heterogenous-examples.rs

@@ -11,7 +11,14 @@ use dspy_rs::{ChatAdapter, LM, Predict, Predictor, configure, example, sign};
 
 #[tokio::main]
 async fn main() -> anyhow::Result<()> {
-    configure(LM::default(), ChatAdapter {});
+    configure(
+        LM::builder()
+            .model("openai:gpt-4o-mini".to_string())
+            .build()
+            .await
+            .unwrap(),
+        ChatAdapter {},
+    );
 
     let exp = example! {
         "number": "input" => 10,

crates/dspy-rs/examples/06-other-providers-batch.rs

@@ -10,8 +10,8 @@ cargo run --example 01-simple
 use anyhow::Result;
 use bon::Builder;
 use dspy_rs::{
-    ChatAdapter, Example, LM, LMConfig, Module, Predict, Prediction, Predictor, Signature,
-    configure, example, hashmap, prediction,
+    ChatAdapter, Example, LM, Module, Predict, Prediction, Predictor, Signature, configure,
+    example, hashmap, prediction,
 };
 
 #[Signature(cot)]
@@ -77,11 +77,11 @@ impl Module for QARater {
 async fn main() {
     // Anthropic
     configure(
-        LM::new(LMConfig {
-            model: "anthropic:claude-sonnet-4-5-20250929".to_string(),
-            ..LMConfig::default()
-        })
-        .await,
+        LM::builder()
+            .model("anthropic:claude-sonnet-4-5-20250929".to_string())
+            .build()
+            .await
+            .unwrap(),
         ChatAdapter,
     );
 
@@ -103,11 +103,11 @@ async fn main() {
 
     // Gemini
     configure(
-        LM::new(LMConfig {
-            model: "gemini:gemini-2.0-flash".to_string(),
-            ..LMConfig::default()
-        })
-        .await,
+        LM::builder()
+            .model("gemini:gemini-2.0-flash".to_string())
+            .build()
+            .await
+            .unwrap(),
         ChatAdapter,
     );
 

crates/dspy-rs/examples/07-inspect-history.rs

@@ -28,7 +28,11 @@ impl Module for QARater {
 
 #[tokio::main]
 async fn main() {
-    let lm = LM::default();
+    let lm = LM::builder()
+        .model("openai:gpt-4o-mini".to_string())
+        .build()
+        .await
+        .unwrap();
     configure(lm, ChatAdapter);
 
     let example = example! {

crates/dspy-rs/examples/09-gepa-sentiment.rs

@@ -117,11 +117,7 @@ async fn main() -> Result<()> {
     println!("GEPA Sentiment Analysis Optimization Example\n");
 
     // Setup LM
-    let lm = LM::new(LMConfig {
-        temperature: 0.7,
-        ..LMConfig::default()
-    })
-    .await;
+    let lm = LM::builder().temperature(0.7).build().await.unwrap();
 
     configure(lm.clone(), ChatAdapter);
 

crates/dspy-rs/examples/10-gepa-llm-judge.rs

@@ -223,18 +223,10 @@ async fn main() -> Result<()> {
 
     // Setup: Configure the LLM
     // Main LM for the task
-    let task_lm = LM::new(LMConfig {
-        temperature: 0.7,
-        ..LMConfig::default()
-    })
-    .await;
+    let task_lm = LM::builder().temperature(0.7).build().await.unwrap();
 
     // Judge LM (could use a different/cheaper model)
-    let judge_lm = LM::new(LMConfig {
-        temperature: 0.3,
-        ..LMConfig::default()
-    })
-    .await;
+    let judge_lm = LM::builder().temperature(0.3).build().await.unwrap();
 
     configure(task_lm, ChatAdapter);
 

crates/dspy-rs/src/adapter/chat.rs

@@ -283,7 +283,7 @@ impl Adapter for ChatAdapter {
         inputs: Example,
     ) -> Result<Prediction> {
         // Check cache first (release lock immediately after checking)
-        if lm.config.cache
+        if lm.cache
             && let Some(cache) = lm.cache_handler.as_ref()
         {
             let cache_key = inputs.clone();
@@ -304,7 +304,7 @@ impl Adapter for ChatAdapter {
         };
 
         // Store in cache if enabled
-        if lm.config.cache
+        if lm.cache
             && let Some(cache) = lm.cache_handler.as_ref()
         {
             let (tx, rx) = tokio::sync::mpsc::channel(1);