Add LLM parsing functionality

Introduce `Providers` struct to handle different language model providers. Implement `Parse` method to extract and validate provider/model from input string, then return a chat completion interface. Add error handling for invalid formats or unknown providers.

parse.go

@@ -0,0 +1,50 @@
+package go_llm
+
+import (
+	"strings"
+)
+
+// Providers are the allowed shortcuts in the providers, e.g.: if you set { "openai": OpenAI("key") } that'll allow
+// for the "openai" provider to be used when parsed.
+type Providers map[string]LLM
+
+// Parse will parse the provided input and attempt to return a LLM chat completion interface.
+// Input should be in the provided format:
+//   - provider/modelname
+//
+// where provider is a key inside Providers, and the modelname being passed to the LLM interface's GetModel
+func (providers Providers) Parse(input string) ChatCompletion {
+	sections := strings.Split(input, "/")
+
+	var provider LLM
+	var ok bool
+	var modelVersion string
+
+	if len(sections) < 2 {
+		// is there a default provider?
+		provider, ok = providers["default"]
+		if !ok {
+			panic("expected format: \"provider/model\" or provide a \"default\" provider to the Parse callback")
+		}
+
+		modelVersion = sections[0]
+	} else {
+		provider, ok = providers[sections[0]]
+		modelVersion = sections[1]
+	}
+
+	if !ok {
+		panic("expected format: \"provider/model\" or provide a \"default\" provider to the Parse callback")
+	}
+
+	if provider == nil {
+		panic("unknown provider: " + sections[0])
+	}
+
+	res, err := provider.ModelVersion(modelVersion)
+	if err != nil {
+		panic(err)
+	}
+
+	return res
+}