pruning upstreamed and unused python packages

mods/python/ai/bindings.nix

@@ -39,82 +39,6 @@ rec {
 
   };
 
-  llama-cpp-python =
-    let
-      osSpecific =
-        if isM1 then with darwin.apple_sdk_11_0.frameworks; [ Accelerate MetalKit MetalPerformanceShaders MetalPerformanceShadersGraph ]
-        else if isDarwin then with darwin.apple_sdk.frameworks; [ Accelerate CoreGraphics CoreVideo ]
-        else [ ];
-    in
-    buildPythonPackage rec {
-      pname = "llama-cpp-python";
-      version = "0.3.4";
-      pyproject = true;
-      src = fetchFromGitHub {
-        owner = "abetlen";
-        repo = pname;
-        rev = "refs/tags/v${version}";
-        hash = "sha256-D8k+vdZPQGUTyAcq+IYFn52peABMg1ZgPuef90Ya9Kg=";
-      };
-
-      cuda = false;
-
-      _CMAKE_ARGS = (optionals isM1 [ "-DLLAMA_METAL=on" ]) ++ (optionals cuda [ "-DLLAMA_CUBLAS=on" ]);
-      CMAKE_ARGS = builtins.concatStringsSep " " _CMAKE_ARGS;
-      FORCE_CMAKE = if (isM1 || cuda) then "1" else null;
-
-      preConfigure = ''
-        cp -r ${llama-cpp-pin}/. ./vendor/llama.cpp
-        chmod -R +w ./vendor/llama.cpp
-      '';
-      preBuild = ''
-        cd ..
-        sed -E -i \
-          -e '/"ninja/d' \
-          -e '/"cmake/d' \
-          -e 's/(requires =).*/\1 ["scikit-build-core[pyproject]>=0.8.2"]/g' \
-          ./pyproject.toml
-      '';
-      buildInputs = osSpecific;
-
-      nativeBuildInputs = (with prev.pkgs; [
-        cmake
-        ninja
-      ]) ++ (with prev; [
-        pythonRelaxDepsHook
-        scikit-build-core
-        pathspec
-        pyproject-metadata
-      ]) ++ (optionals cuda [ cudatoolkit ]);
-      pythonRelaxDeps = [ "diskcache" ];
-      propagatedBuildInputs = with prev; [
-        diskcache
-        jinja2
-        numpy
-        typing-extensions
-
-        # server mode
-        fastapi
-        sse-starlette
-        uvicorn
-      ];
-
-      pythonImportsCheck = [ "llama_cpp" ];
-
-      passthru.cuda = llama-cpp-python.overridePythonAttrs (old: {
-        CMAKE_ARGS = "-DLLAMA_CUBLAS=on";
-        FORCE_CMAKE = 1;
-        nativeBuildInputs = old.nativeBuildInputs ++ [ cudatoolkit ];
-      });
-
-      meta = {
-        description = "A Python wrapper for llama.cpp";
-        homepage = "https://github.com/abetlen/llama-cpp-python";
-        license = licenses.mit;
-        maintainers = with maintainers; [ jpetrucciani ];
-      };
-    };
-
   ggml-python =
     let
       osSpecific =