Merge pull request #238 from eljamm/llama-cpp

add: script llama-cpp-integration

Author: pbek

llama-cpp-integration/info.json

@@ -0,0 +1,9 @@
+{
+  "name": "llama-cpp integration",
+  "identifier": "llama-cpp-integration",
+  "script": "llama-cpp-integration.qml",
+  "version": "0.0.1",
+  "minAppVersion": "24.6.2",
+  "authors": ["@pbek"],
+  "description" : "This script provides integration for a local <a href=\"https://github.com/ggerganov/llama.cpp\">llama-cpp</a> AI backend."
+}

llama-cpp-integration/llama-cpp-integration.qml

@@ -0,0 +1,47 @@
+import QtQml 2.0
+import QOwnNotesTypes 1.0
+
+/**
+ * This script provides integration for a local llama-cpp backend
+ * See: https://github.com/ggerganov/llama.cpp
+ * List of models: https://github.com/ggerganov/llama.cpp#description
+ * OpenAPI endpoint: https://github.com/ggerganov/llama.cpp/blob/master/examples/server/README.md#result-json-1
+ */
+Script {
+    property string baseUrl;
+    property string models;
+
+    // register your settings variables so the user can set them in the script settings
+    property variant settingsVariables: [
+        {
+            "identifier": "baseUrl",
+            "name": "API base URL",
+            "description": "The chat base URL of the llama-cpp API.",
+            "type": "string",
+            "default": "http://127.0.0.1:8080/v1/chat/completions",
+        },
+        {
+            "identifier": "models",
+            "name": "Models",
+            "description": "Comma separated list of models to use.",
+            "type": "string",
+            "default": "llama3,gemma:2b",
+        },
+    ];
+
+    /**
+     * This function is called when the OpenAI service config is reloaded
+     * It returns a list of objects with config parameters for new OpenAI backends
+     */
+    function openAiBackendsHook() {
+        return [
+            {
+                "id": "llama-cpp",
+                "name": "llama-cpp",
+                "baseUrl": baseUrl,
+                "apiKey": "llama-cpp",
+                "models": models.split(",")
+            },
+        ];
+    }
+}