Add OPEA ChatQna example

Author: lucasmelogithub

examples/gen-ai-xeon-opea-chatqna/README.md

@@ -0,0 +1,66 @@
+<p align="center">
+  <img src="https://github.com/intel/terraform-intel-gcp-vm/blob/main/images/logo-classicblue-800px.png?raw=true" alt="Intel Logo" width="250"/>
+</p>
+
+# Intel® Optimized Cloud Modules for Terraform
+
+© Copyright 2024, Intel Corporation
+
+## GCP C4 Instance with 5th Generation Intel® Xeon® Scalable Processor (Emerald Rapids) & Open Platform for Enterprise AI (OPEA) ChatQnA Example
+
+This demo will showcase Retrieval Augmented Generation (RAG) CPU inference using 5th Gen Xeon Scalable Processors on GCP using the OPEA ChatQnA Example. For more information about OPEA, go [here](https://opea.dev/). For more information on this specific example, go [here](https://github.com/opea-project/GenAIExamples/tree/main/ChatQnA).
+
+## Usage
+
+## After cloning the repo, modify /examples/gen-ai-xeon-opea-chatqna/variables.tf to add your Huggingface Token
+
+Some Models required a Token. Modify the Huggingface Token variable to your specific Huggingface Token, for information on creating a Huggingface token go [here](https://huggingface.co/docs/hub/en/security-tokens)
+
+```hcl
+variable "huggingface_token" {
+  description = "Huggingface Token"
+  default     = " <YOUR HUGGINGFACE TOKEN> "
+  type        = string
+}
+```
+
+**If needed, also modify values on /examples/gen-ai-xeon-opea-chatqna/main.tf to match your needs**
+
+## GCP Cloud Shell Usage
+
+1. Log on to GCP Portal
+2. Enter the GCP Cloud Shell (terminal button on top right of page)
+3. Run the following commands in order: 
+
+
+```bash
+git clone https://github.com/intel/terraform-intel-gcp-vm.git
+cd terraform-intel-gcp-vm/examples/gen-ai-xeon-opea-chatqna
+# ADD TOKEN TO variables.tf (SEE ABOVE)
+terraform init 
+terraform plan
+terraform apply
+
+# (enter your GCP project ID and "yes" to confirm)
+
+```
+
+After the Terraform module successfully creates the GCP VM instance, **wait ~15 minutes** for the module to launch the containers and download the LLMs before continuing.
+
+## Accessing the Demo
+
+You can access the demos using the following:
+
+- OPEA ChatQnA: `http://yourpublicip:5174`
+
+
+## Deleting the Demo
+
+To delete the demo, run `terraform destroy` to delete all resources created.
+
+## Pre-requisites for running on a local Workstation (disregard if using GCP Cloud Shell)
+
+1. Google Cloud CLI: https://cloud.google.com/sdk/docs/install
+2. CGP account access configured: https://registry.terraform.io/providers/hashicorp/google/latest/docs/guides/provider_reference.html#running-terraform-on-your-workstation
+3. Terraform: https://learn.hashicorp.com/tutorials/terraform/install-cli
+4. Git: https://git-scm.com/book/en/v2/Getting-Started-Installing-Git

examples/gen-ai-xeon-opea-chatqna/cloud_init.yml

@@ -0,0 +1,18 @@
+#cloud-config
+package_update: true
+package_upgrade: true
+
+package:
+  - git
+
+runcmd:
+  - apt install ansible -y
+  - git clone https://github.com/intel/optimized-cloud-recipes.git /tmp/optimized-cloud-recipes
+  - cd /tmp/optimized-cloud-recipes/recipes/ai-opea-chatqna-xeon
+  - cp opea.sh /etc/profile.d/opea.sh
+  - echo 'export HUGGINGFACEHUB_API_TOKEN=${HUGGINGFACEHUB_API_TOKEN}' | sudo tee -a /etc/profile.d/opea.sh
+  - chmod +x /etc/profile.d/opea.sh
+  - source /etc/profile.d/opea.sh
+  - ansible-playbook recipe.yml
+
+    

examples/gen-ai-xeon-opea-chatqna/main.tf

@@ -0,0 +1,38 @@
+#Random ID to minimize the chances of name conflicts
+resource "random_id" "rid" {
+  byte_length = 3
+}
+
+#GCP Linux VM with Intel OPEA ChatQnA 
+module "linux_vm" {
+  source                    = "../.." #intel/gcp-vm/intel
+  project                   = var.project
+  boot_image_project        = "ubuntu-os-cloud"
+  boot_image_family         = "ubuntu-2204-lts"
+  name                      = "lmelo-ai-opea-chatqna-${random_id.rid.dec}"
+  zone                      = "us-east4-a" #"us-central1-a"
+  machine_type              = "c4-highcpu-48"
+  allow_stopping_for_update = true
+  tags                      = ["lmelo-ai-opea-chatqna-${random_id.rid.dec}"]
+  user_data = templatefile("./cloud_init.yml", { HUGGINGFACEHUB_API_TOKEN = var.huggingface_token })
+  access_config = [{
+    nat_ip                 = null
+    public_ptr_domain_name = null
+    network_tier           = "PREMIUM"
+  }, ]
+}
+
+#Required firewall rules
+resource "google_compute_firewall" "rules" {
+  project     = var.project
+  name        = "lmelo-ai-opea-chatqna-${random_id.rid.dec}"
+  network     = "default"
+  description = "Allows access to OPEA AI ChatQnA"
+
+  allow {
+    protocol = "tcp"
+    ports    = ["22","80", "443", "6379", "8001", "6006", "6007", "6000", "7000", "8808", "8000", "8888", "5173", "5174", "9009", "9000"]
+  }
+  source_ranges = ["0.0.0.0/0"]
+  target_tags   = ["lmelo-ai-opea-chatqna-${random_id.rid.dec}"]
+}

examples/gen-ai-xeon-opea-chatqna/variables.tf

@@ -0,0 +1,12 @@
+variable "project" {
+  type        = string
+  default     = "551221341017"
+  description = "Enter GCP Project ID"
+}
+
+# Variable for Huggingface Token
+variable "huggingface_token" {
+  description = "Enter a Huggingface Token to be used to download the models"
+  default     = "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"
+  type        = string
+}

main.tf

@@ -7,12 +7,14 @@
 # Intel CPU, we are not populating the min CPU platform. We are using the default CPU platform that GCP will provide for these older generation of instances
 
 locals {
-  machine_type_regex = "^([cemn][123u])"
+  machine_type_regex = "^([cemn][1234u])"
   machine_types = {
     "n2": "Intel Ice Lake",
     "c3": "Intel Sapphire Rapids",
     "m3": "Intel Ice Lake",
     "c2": "Intel Cascade Lake"
+    "c4": null
+    "n4": null
     "n1": null
     "m1": null
     "m2": null

variables.tf

@@ -103,13 +103,13 @@ variable "boot_image_family" {
 variable "boot_disk_size" {
   type        = number
   description = "Size of the OS disk"
-  default     = 100
+  default     = 500
 }
 
 variable "boot_disk_type" {
   type        = string
   description = "Disk type associated with the OS disk. Values can be either pd-ssd, local-ssd, or pd-standard"
-  default     = "pd-ssd"
+  default     = null
 }
 
 variable "boot_image_project" {

versions.tf

@@ -4,7 +4,7 @@ terraform {
   required_providers {
     google = {
       source  = "hashicorp/google"
-      version = "~> 5.11"
+      version = "~> 6.5"
     }
   }
 }