Handle edge cases when generating embeddings (#1215)

* Handle edge cases when generating embeddings
* Improve Python handling & add llmodel_c.h note
- In the Python bindings fail fast with a ValueError when text is empty
- Advice other bindings authors to do likewise in llmodel_c.h

gpt4all-backend/llmodel_c.cpp

@@ -168,10 +168,14 @@ void llmodel_prompt(llmodel_model model, const char *prompt,
 
 float *llmodel_embedding(llmodel_model model, const char *text, size_t *embedding_size)
 {
+    if (model == nullptr || text == nullptr || !strlen(text)) {
+        *embedding_size = 0;
+        return nullptr;
+    }
     LLModelWrapper *wrapper = reinterpret_cast<LLModelWrapper*>(model);
     std::vector<float> embeddingVector = wrapper->llModel->embedding(text);
     float *embedding = (float *)malloc(embeddingVector.size() * sizeof(float));
-    if(embedding == nullptr) {
+    if (embedding == nullptr) {
         *embedding_size = 0;
         return nullptr;
     }

gpt4all-backend/llmodel_c.h

@@ -173,6 +173,8 @@ void llmodel_prompt(llmodel_model model, const char *prompt,
 
 /**
  * Generate an embedding using the model.
+ * NOTE: If given NULL pointers for the model or text, or an empty text, a NULL pointer will be
+ * returned. Bindings should signal an error when NULL is the return value.
  * @param model A pointer to the llmodel_model instance.
  * @param text A string representing the text to generate an embedding for.
  * @param embedding_size A pointer to a size_t type that will be set by the call indicating the length