Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/aaronpowell/generateandchill
A demo app of how to use Azure OpenAI Service to generate verify prompts and generate images.
https://github.com/aaronpowell/generateandchill
Last synced: about 1 month ago
JSON representation
A demo app of how to use Azure OpenAI Service to generate verify prompts and generate images.
- Host: GitHub
- URL: https://github.com/aaronpowell/generateandchill
- Owner: aaronpowell
- License: mit
- Created: 2023-08-18T03:23:54.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2023-12-04T00:51:15.000Z (about 1 year ago)
- Last Synced: 2024-11-17T19:45:46.457Z (about 2 months ago)
- Language: Bicep
- Size: 3.94 MB
- Stars: 6
- Watchers: 4
- Forks: 6
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Image Generator Sample
This sample application uses [Azure OpenAI Service](https://learn.microsoft.com/azure/ai-services/openai/) to generate images from text.
## How it works
There are two parts to the AI component of this application, initial text processing of the user prompt and image generation.
### Text Processing
To aim for consistent generation of images the text prompt provided needs to be in English and needs to have a provided style for the image to be generated. To ensure this, an initial [Chat Completion](https://learn.microsoft.com/azure/ai-services/openai/chatgpt-quickstart?tabs=command-line&pivots=programming-language-csharp) is performed using a [System Prompt](https://learn.microsoft.com/azure/ai-services/openai/concepts/system-message) to guide the LLM into how it should generate the prompt.
The System Prompt can be found [here](./blob/main/src/GenerateAndChill.Server/Routes.cs#L13).
### Image Generation
Once the prompt has been returned from the Chat Completion it is provided to the [image model](https://learn.microsoft.com/azure/ai-services/openai/dall-e-quickstart?tabs=command-line&pivots=programming-language-csharp) to generate an image. The returned image is then stored in Azure Storage before being returned to the user.
## Running the sample
### Local development
[](https://codespaces.new/aaronpowell/GenerateAndChill)
This sample uses a [devcontainer](https://containers.dev) to prepare a development environment, meaning it can be opened in GitHub Codespaces or locally using the [VSCode Remote Containers extension](https://code.visualstudio.com/docs/remote/containers).
#### Manual setup
If you don't want to use the devcontainer you need to ensure you have the following:
- [.NET 7 SDK](https://dotnet.microsoft.com/download/dotnet/7.0)
- [Node.js 20.x](https://nodejs.org/en/download/)
- [Azurite](https://github.com/azure/azurite)
- Alternatively, you can use a deployed Azure Storage account
- [Azure OpenAI Service](https://learn.microsoft.com/azure/ai-services/openai/)
- You'll need to create a new resource and note the API key and endpoint
- [Azure Developer CLI](https://aka.ms/azd)
### Provisioning Azure resources
To provision the Azure resources you can use the [Azure Developer CLI](https://aka.ms/azd) and run the following commands:
```bash
azd auth login
azd env new -n MyDemoEnv
azd provision
```
Once the deployment is complete, navgiate to the resource group in Azure and copy the OpenAI + Storage connection string into the `appsettings.Development.json` file like so:
```json
{
"Azure": {
"OpenAIEndpoint": "",
"OpenAIKey": "",
"BlobStorageConnectionString": "",
"TableStorageConnectionString": ""
}
}
```
### Running the application
From VS Code launch the application with the `Full Stack` launch profile, which will start the .NET server, the React frontend and the [Azure Static Web Apps CLI](https://learn.microsoft.com/azure/static-web-apps/static-web-apps-cli).
Navigate to http://localhost:4280 to see the application running.
### Deploying to Azure
To deploy the application to Azure you can use the [Azure Developer CLI](https://aka.ms/azd) and run the following commands:
```bash
azd up
```
Once the deployment is complete the URL for the frontend application will be displayed in the terminal.
## License
[MIT](./LICENSE)