Utilize llama.cpp for both completion and chat models (#141)

libs/kotaemon/pyproject.toml

@@ -11,7 +11,7 @@ packages.find.exclude = ["tests*", "env*"]
 # metadata and dependencies
 [project]
 name = "kotaemon"
-version = "0.3.7"
+version = "0.3.8"
 requires-python = ">= 3.10"
 description = "Kotaemon core library for AI development."
 dependencies = [
@@ -64,6 +64,7 @@ dev = [
     "elasticsearch",
     "pypdf",
     "html2text",
+    "llama-cpp-python",
 ]
 
 [project.scripts]