From 3f4fa33b9bc9c0aec7f92bd125634c26a2c938d4 Mon Sep 17 00:00:00 2001
From: jgrusewski <jgrusewski@localhost>
Date: Mon, 20 Apr 2026 22:45:51 +0200
Subject: [PATCH] =?UTF-8?q?cleanup(fflag):=20delete=207=20dead=20transform?=
 =?UTF-8?q?er=20flags=20=E2=80=94=20[FFLAG-016,=20FFLAG-017]?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

HFTTransformerConfig had seven write-only bool flags with zero conditional
readers anywhere in the workspace:
- FFLAG-017: use_market_microstructure, use_order_book_features,
  use_trade_flow_features
- FFLAG-016: use_flash_attention, use_sparse_attention, use_cuda_graphs,
  enable_profiling

Each was declared + defaulted + asserted in tests, but no code branched on
them. Deleted all 7 fields + Default init + production()/benchmark() setters
+ test assertions. Per feedback_no_feature_flags.md these were aspirational
knobs — the features they nominally gated have no implementation.
---
 crates/ml/src/transformers/mod.rs | 23 +----------------------
 1 file changed, 1 insertion(+), 22 deletions(-)

diff --git a/crates/ml/src/transformers/mod.rs b/crates/ml/src/transformers/mod.rs
index e1306e745..a99a6b155 100644
--- a/crates/ml/src/transformers/mod.rs
+++ b/crates/ml/src/transformers/mod.rs
@@ -117,13 +117,8 @@ pub struct HFTTransformerConfig {
 
     /// Feature configuration
     pub feature_dim: usize,
-    pub use_market_microstructure: bool,
-    pub use_order_book_features: bool,
-    pub use_trade_flow_features: bool,
 
     /// Optimization settings
-    pub use_flash_attention: bool,
-    pub use_sparse_attention: bool,
     pub attention_sparsity: f32,
 
     /// Memory optimization
@@ -132,8 +127,6 @@ pub struct HFTTransformerConfig {
 
     /// Hardware settings
     pub device_type: DeviceType,
-    pub use_cuda_graphs: bool,
-    pub enable_profiling: bool,
 }
 
 impl Default for HFTTransformerConfig {
@@ -147,17 +140,10 @@ impl Default for HFTTransformerConfig {
             ff_dim: 256,
             seq_len: 64,
             feature_dim: 32,
-            use_market_microstructure: true,
-            use_order_book_features: true,
-            use_trade_flow_features: true,
-            use_flash_attention: true,
-            use_sparse_attention: false,
             attention_sparsity: 0.1,
             pre_allocate_tensors: true,
             memory_pool_size: 1024 * 1024 * 64, // 64MB
             device_type: DeviceType::Cuda,
-            use_cuda_graphs: false, // Enable after validation
-            enable_profiling: false,
         }
     }
 }
@@ -209,19 +195,14 @@ impl HFTTransformerConfig {
     /// Enable all optimizations for production deployment
     pub fn production() -> Self {
         Self {
-            use_flash_attention: true,
             pre_allocate_tensors: true,
-            use_cuda_graphs: true,
             ..Self::micro()
         }
     }
 
     /// Configuration for benchmarking and validation
     pub fn benchmark() -> Self {
-        Self {
-            enable_profiling: true,
-            ..Self::micro()
-        }
+        Self::micro()
     }
 }
 
@@ -252,8 +233,6 @@ mod tests {
         assert_eq!(nano.hidden_dim, 32);
 
         let production = HFTTransformerConfig::production();
-        assert!(production.use_flash_attention);
         assert!(production.pre_allocate_tensors);
-        assert!(production.use_cuda_graphs);
     }
 }