Feature/azure openai templates (#104)

* Azure OpenAI LLM templates
* Bump version, fix package versions
* Add azure-openai to template generation

Makefile

@@ -1,6 +1,6 @@
 
 # VERSION=$(shell git describe | sed 's/^v//')
-VERSION=0.11.19
+VERSION=0.12.0
 
 DOCKER=podman
 
@@ -16,7 +16,7 @@ wheels:
 	pip3 wheel --no-deps --wheel-dir dist trustgraph-embeddings-hf/
 	pip3 wheel --no-deps --wheel-dir dist trustgraph-cli/
 
-packages:
+packages: update-package-versions
 	rm -rf dist/
 	cd trustgraph-base && python3 setup.py sdist --dist-dir ../dist/
 	cd trustgraph-flow && python3 setup.py sdist --dist-dir ../dist/

templates/all-patterns.jsonnet

@@ -7,6 +7,7 @@
     import "patterns/triple-store-neo4j.jsonnet",
     import "patterns/graph-rag.jsonnet",
     import "patterns/llm-azure.jsonnet",
+    import "patterns/llm-azure-openai.jsonnet",
     import "patterns/llm-bedrock.jsonnet",
     import "patterns/llm-claude.jsonnet",
     import "patterns/llm-cohere.jsonnet",

templates/components.jsonnet

@@ -1,5 +1,6 @@
 {
    "azure": import "components/azure.jsonnet",
+   "azure-openai": import "components/azure-openai.jsonnet",
    "bedrock": import "components/bedrock.jsonnet",
    "claude": import "components/claude.jsonnet",
    "cohere": import "components/cohere.jsonnet",

templates/components/azure-openai.jsonnet

@@ -0,0 +1,95 @@
+local base = import "base/base.jsonnet";
+local images = import "values/images.jsonnet";
+local url = import "values/url.jsonnet";
+local prompts = import "prompts/mixtral.jsonnet";
+
+{
+
+    "azure-openai-token":: "${AZURE_OPENAI_TOKEN}",
+    "azure-openai-model":: "GPT-3.5-Turbo",
+    "azure-openai-max-output-tokens":: 4192,
+    "azure-openai-temperature":: 0.0,
+
+    "text-completion" +: {
+    
+        create:: function(engine)
+
+            local container =
+                engine.container("text-completion")
+                    .with_image(images.trustgraph)
+                    .with_command([
+                        "text-completion-azure-openai",
+                        "-p",
+                        url.pulsar,
+                        "-k",
+                        $["azure-openai-token"],
+                        "-m",
+                        $["azure-openai-model"],
+                        "-x",
+                        std.toString($["azure-openai-max-output-tokens"]),
+                        "-t",
+                        std.toString($["azure-openai-temperature"]),
+                    ])
+                    .with_limits("0.5", "128M")
+                    .with_reservations("0.1", "128M");
+
+            local containerSet = engine.containers(
+                "text-completion", [ container ]
+            );
+
+            local service =
+                engine.internalService(containerSet)
+                .with_port(8000, 8000, "metrics");
+
+            engine.resources([
+                containerSet,
+                service,
+            ])
+
+    },
+
+    "text-completion-rag" +: {
+    
+        create:: function(engine)
+
+            local container =
+                engine.container("text-completion-rag")
+                    .with_image(images.trustgraph)
+                    .with_command([
+                        "text-completion-azure",
+                        "-p",
+                        url.pulsar,
+                        "-k",
+                        $["azure-openai-token"],
+                        "-e",
+                        $["azure-openai-model"],
+                        "-x",
+                        std.toString($["azure-openai-max-output-tokens"]),
+                        "-t",
+                        std.toString($["azure-openai-temperature"]),
+                        "-i",
+                        "non-persistent://tg/request/text-completion-rag",
+                        "-o",
+                        "non-persistent://tg/response/text-completion-rag-response",
+                    ])
+                    .with_limits("0.5", "128M")
+                    .with_reservations("0.1", "128M");
+
+            local containerSet = engine.containers(
+                "text-completion-rag", [ container ]
+            );
+
+            local service =
+                engine.internalService(containerSet)
+                .with_port(8000, 8000, "metrics");
+
+            engine.resources([
+                containerSet,
+                service,
+            ])
+
+
+    }
+
+} + prompts
+

templates/generate-all

@@ -122,8 +122,8 @@ def generate_all(output, version):
             "docker-compose", "minikube-k8s", "gcp-k8s"
     ]:
         for model in [
-                "azure", "bedrock", "claude", "cohere", "llamafile", "ollama",
-                "openai", "vertexai"
+                "azure", "azure-openai", "bedrock", "claude", "cohere",
+                "llamafile", "ollama", "openai", "vertexai"
         ]:
             for graph in [ "cassandra", "neo4j" ]:
 

templates/patterns/llm-azure-openai.jsonnet

@@ -0,0 +1,32 @@
+{
+    pattern: {
+	name: "azure-openai",
+        icon: "🤖💬",
+        title: "Add Azure OpenAI LLM endpoint for text completion",
+	description: "This pattern integrates an Azure OpenAI LLM endpoint hosted in the Azure cloud for text completion operations.  You need an Azure subscription to be able to use this service.",
+        requires: ["pulsar", "trustgraph"],
+        features: ["llm"],
+	args: [
+	    {
+		name: "azure-openai-max-output-tokens",
+                label: "Maximum output tokens",
+		type: "integer",
+		description: "Limit on number tokens to generate",
+                default: 4096,
+		required: true,
+            },
+	    {
+		name: "azure-openai-temperature",
+                label: "Temperature",
+		type: "slider",
+		description: "Controlling predictability / creativity balance",
+                min: 0,
+                max: 1,
+                step: 0.05,
+                default: 0.5,
+            },
+	],
+        category: [ "llm" ],
+    },
+    module: "components/azure.jsonnet",
+}