llmodel_c: improve quality of error messages (#1625)

gpt4all-bindings/golang/Makefile

@@ -139,7 +139,7 @@ $(info I CXX:      $(CXXV))
 $(info )
 
 llmodel.o:
-	mkdir buildllm
+	[ -e buildllm ] || mkdir buildllm
 	cd buildllm && cmake ../../../gpt4all-backend/ $(CMAKEFLAGS) && make
 	cd buildllm && cp -rf CMakeFiles/llmodel.dir/llmodel_c.cpp.o ../llmodel_c.o
 	cd buildllm && cp -rf CMakeFiles/llmodel.dir/llmodel.cpp.o ../llmodel.o
@@ -150,7 +150,7 @@ clean:
 	rm -rf buildllm
 	rm -rf example/main
 
-binding.o: 
+binding.o: binding.cpp binding.h
 	$(CXX) $(CXXFLAGS) binding.cpp -o binding.o -c $(LDFLAGS)
 
 libgpt4all.a: binding.o llmodel.o

gpt4all-bindings/golang/binding.cpp

@@ -17,11 +17,10 @@
 
 void* load_model(const char *fname, int n_threads) {
     // load the model
-    llmodel_error new_error{};
+    const char *new_error;
     auto model = llmodel_model_create2(fname, "auto", &new_error);
-    if (model == nullptr ){
-        fprintf(stderr, "%s: error '%s'\n",
-                __func__, new_error.message);
+    if (model == nullptr) {
+        fprintf(stderr, "%s: error '%s'\n", __func__, new_error);
         return nullptr;
     }
     if (!llmodel_loadModel(model, fname)) {

gpt4all-bindings/java/src/main/java/com/hexadevlabs/gpt4all/LLModel.java

@@ -1,6 +1,7 @@
 package com.hexadevlabs.gpt4all;
 
 import jnr.ffi.Pointer;
+import jnr.ffi.byref.PointerByReference;
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
 
@@ -176,7 +177,7 @@ public  class LLModel implements AutoCloseable {
         modelName = modelPath.getFileName().toString();
         String modelPathAbs = modelPath.toAbsolutePath().toString();
 
-        LLModelLibrary.LLModelError error = new LLModelLibrary.LLModelError(jnr.ffi.Runtime.getSystemRuntime());
+        PointerByReference error = new PointerByReference();
 
         // Check if model file exists
         if(!Files.exists(modelPath)){
@@ -192,7 +193,7 @@ public  class LLModel implements AutoCloseable {
         model = library.llmodel_model_create2(modelPathAbs, "auto", error);
 
         if(model == null) {
-            throw new IllegalStateException("Could not load, gpt4all backend returned error: " + error.message);
+            throw new IllegalStateException("Could not load, gpt4all backend returned error: " + error.getValue().getString(0));
         }
         library.llmodel_loadModel(model, modelPathAbs);
 
@@ -631,4 +632,4 @@ public  class LLModel implements AutoCloseable {
         library.llmodel_model_destroy(model);
     }
 
-}
+}

gpt4all-bindings/java/src/main/java/com/hexadevlabs/gpt4all/LLModelLibrary.java

@@ -1,6 +1,7 @@
 package com.hexadevlabs.gpt4all;
 
 import jnr.ffi.Pointer;
+import jnr.ffi.byref.PointerByReference;
 import jnr.ffi.Struct;
 import jnr.ffi.annotations.Delegate;
 import jnr.ffi.annotations.Encoding;
@@ -58,7 +59,7 @@ public interface LLModelLibrary {
         }
     }
 
-    Pointer llmodel_model_create2(String model_path, String build_variant, @Out LLModelError llmodel_error);
+    Pointer llmodel_model_create2(String model_path, String build_variant, PointerByReference error);
     void llmodel_model_destroy(Pointer model);
     boolean llmodel_loadModel(Pointer model, String model_path);
     boolean llmodel_isModelLoaded(Pointer model);

gpt4all-bindings/python/gpt4all/pyllmodel.py

@@ -42,10 +42,6 @@ def load_llmodel_library():
 llmodel = load_llmodel_library()
 
 
-class LLModelError(ctypes.Structure):
-    _fields_ = [("message", ctypes.c_char_p), ("code", ctypes.c_int32)]
-
-
 class LLModelPromptContext(ctypes.Structure):
     _fields_ = [
         ("logits", ctypes.POINTER(ctypes.c_float)),
@@ -77,7 +73,7 @@ class LLModelGPUDevice(ctypes.Structure):
 llmodel.llmodel_model_create.argtypes = [ctypes.c_char_p]
 llmodel.llmodel_model_create.restype = ctypes.c_void_p
 
-llmodel.llmodel_model_create2.argtypes = [ctypes.c_char_p, ctypes.c_char_p, ctypes.POINTER(LLModelError)]
+llmodel.llmodel_model_create2.argtypes = [ctypes.c_char_p, ctypes.c_char_p, ctypes.POINTER(ctypes.c_char_p)]
 llmodel.llmodel_model_create2.restype = ctypes.c_void_p
 
 llmodel.llmodel_model_destroy.argtypes = [ctypes.c_void_p]
@@ -150,6 +146,14 @@ def empty_response_callback(token_id: int, response: str) -> bool:
     return True
 
 
+def _create_model(model_path: bytes) -> ctypes.c_void_p:
+    err = ctypes.c_char_p()
+    model = llmodel.llmodel_model_create2(model_path, b"auto", ctypes.byref(err))
+    if model is None:
+        raise ValueError(f"Unable to instantiate model: {err.decode()}")
+    return model
+
+
 class LLModel:
     """
     Base class and universal wrapper for GPT4All language models
@@ -178,12 +182,8 @@ class LLModel:
 
     def memory_needed(self, model_path: str) -> int:
         model_path_enc = model_path.encode("utf-8")
-        self.model = llmodel.llmodel_model_create(model_path_enc)
-
-        if self.model is not None:
-            return llmodel.llmodel_required_mem(self.model, model_path_enc)
-        else:
-            raise ValueError("Unable to instantiate model")
+        self.model = _create_model(model_path_enc)
+        return llmodel.llmodel_required_mem(self.model, model_path_enc)
 
     def list_gpu(self, model_path: str) -> list:
         """
@@ -253,11 +253,7 @@ class LLModel:
         True if model loaded successfully, False otherwise
         """
         model_path_enc = model_path.encode("utf-8")
-        err = LLModelError()
-        self.model = llmodel.llmodel_model_create2(model_path_enc, b"auto", ctypes.byref(err))
-
-        if self.model is None:
-            raise ValueError(f"Unable to instantiate model: code={err.code}, {err.message.decode()}")
+        self.model = _create_model(model_path_enc)
 
         llmodel.llmodel_loadModel(self.model, model_path_enc)
 

gpt4all-bindings/typescript/index.cc

@@ -134,13 +134,10 @@ Napi::Value NodeModelWrapper::GetRequiredMemory(const Napi::CallbackInfo& info)
         device = config_object.Get("device").As<Napi::String>();
     }
     llmodel_set_implementation_search_path(library_path.c_str());
-    llmodel_error e = {
-        .message="looks good to me",
-        .code=0,
-    };
+    const char* e;
     inference_ = llmodel_model_create2(full_weight_path.c_str(), "auto", &e);
-    if(e.code != 0) {
-       Napi::Error::New(env, e.message).ThrowAsJavaScriptException(); 
+    if(!inference_) {
+       Napi::Error::New(env, e).ThrowAsJavaScriptException();
        return;
     }
     if(GetInference() == nullptr) {