Add LODR support to online and offline recognizers (#2026)

This PR integrates LODR (Level-Ordered Deterministic Rescoring) support from Icefall into both online and offline recognizers, enabling LODR for LM shallow fusion and LM rescore.

- Extended OnlineLMConfig and OfflineLMConfig to include lodr_fst, lodr_scale, and lodr_backoff_id.
- Implemented LodrFst and LodrStateCost classes and wired them into RNN LM scoring in both online and offline code paths.
- Updated Python bindings, CLI entry points, examples, and CI test scripts to accept and exercise the new LODR options.

sherpa-onnx/csrc/hypothesis.h

@@ -12,9 +12,11 @@
 #include <unordered_map>
 #include <utility>
 #include <vector>
+#include <memory>
 
 #include "onnxruntime_cxx_api.h"  // NOLINT
 #include "sherpa-onnx/csrc/context-graph.h"
+#include "sherpa-onnx/csrc/lodr-fst.h"
 #include "sherpa-onnx/csrc/math.h"
 #include "sherpa-onnx/csrc/onnx-utils.h"
 
@@ -61,6 +63,9 @@ struct Hypothesis {
   // the nn lm states
   std::vector<CopyableOrtValue> nn_lm_states;
 
+  // the LODR states
+  std::shared_ptr<LodrStateCost> lodr_state;
+
   const ContextState *context_state;
 
   // TODO(fangjun): Make it configurable