make libmf multithreading work

.cargo/config.toml

@@ -1,2 +1,17 @@
-[build]
-rustflags = ["-C", "link-arg=-fuse-ld=lld"]
+[target.x86_64-unknown-linux-gnu]
+rustflags = [
+    "-C",
+    "link-arg=-fuse-ld=lld",
+    "-C",
+    "link-arg=-L/usr/lib/x86_64-linux-gnu",
+    "-C",
+    "link-arg=-lgomp",
+    "-C",
+    "link-arg=-fopenmp",
+]
+
+[env]
+CXXFLAGS = "-fopenmp -pthread -DUSEOMP=1"
+LDFLAGS = "-fopenmp -pthread -DUSEOMP=1"
+CC = "gcc"
+CXX = "g++"

src/main.rs

@@ -28,6 +28,10 @@ struct Args {
     #[arg(long)]
     target_table: String,
 
+    /// Number of iterations for matrix factorization
+    #[arg(long, short = 'i', default_value = "100")]
+    iterations: i32,
+
     /// Batch size for loading data
     #[arg(long, default_value = "10000")]
     batch_size: i32,
@@ -51,6 +55,10 @@ struct Args {
     /// Number of threads for matrix factorization (defaults to number of CPU cores)
     #[arg(long, default_value_t = num_cpus::get() as i32)]
     threads: i32,
+
+    /// Number of bins to use for training
+    #[arg(long, default_value = "10")]
+    bins: i32,
 }
 
 async fn create_pool() -> Result<Pool> {
@@ -113,6 +121,7 @@ async fn load_data_batch(
     Ok(batch)
 }
 
+// TODO - don't load all item IDs at once
 async fn save_embeddings(pool: &Pool, args: &Args, model: &Model, item_ids: &[i32]) -> Result<()> {
     let client = pool.get().await?;
 
@@ -248,11 +257,12 @@ async fn main() -> Result<()> {
         .lambda_p2(args.lambda2)
         .lambda_q2(args.lambda2)
         .learning_rate(args.learning_rate)
-        .iterations(100)
+        .iterations(args.iterations)
         .loss(Loss::OneClassL2)
         .c(0.00001)
         .quiet(false)
         .threads(args.threads)
+        .bins(args.bins)
         .fit(&matrix)?;
 
     info!("Saving embeddings...");