README.md•2.65 kB
# Context Caching Test Application
This is a sample application demonstrating the use of **context caching** with Google's Gemini models via the Google Generative AI plugin in Genkit. The application highlights how to efficiently reuse context in large language model interactions.
## Features
- **Context Caching**: Efficiently reuse cached content to improve performance and reduce costs.
- **Supports Large Texts**: Works seamlessly with large texts, such as _War and Peace_ and _Lord of the Rings_.
- **Flexible Model Configuration**: Easily switch between supported Gemini models.
## Installation
To run this application, ensure you have the necessary dependencies installed.
### Prerequisites
- **Node.js** version 20 or higher
- **npm** or **pnpm** for package management
### Install Dependencies
```bash
npm install
```
or
```bash
pnpm install
```
## Scripts
The following scripts are available:
- **`build`**: Compile TypeScript files.
- **`build:watch`**: Watch for changes and recompile TypeScript files.
- **`start`**: Run the compiled application.
- **`dev`**: Start the application in development mode with environment variable setup.
- **`genkit:dev`**: Run the application with `GENKIT_ENV=dev`.
## Configuration
This application uses the Google Generative AI plugin for Genkit. Ensure you have a valid API key for the Gemini API.
### API Key Setup
You can configure the API key in two ways:
1. **Environment Variable**:
```bash
export GOOGLE_GENAI_API_KEY=your_api_key_here
```
2. **Inline Configuration** (for testing only, not recommended for production):
```ts
googleAI({ apiKey: yourKey });
```
## Usage
### Running the Application
To start the application:
```bash
npm run dev
```
or
```bash
pnpm dev
```
### Sample Flow
The application defines a `lotrFlow` for analyzing _Lord of the Rings_ text. Modify the input file path in the `lotrFlow` definition to test different texts:
```ts
const defaultQuery =
"Describe Gandalf's relationship with Frodo, referencing Gandalf quotes from the text.";
```
### Output
The output will be generated by the model based on the provided context and query.
## Dependencies
- **@genkit-ai/googleai**: Provides the interface for Google Gemini models.
- **genkit**: Framework for integrating generative models.
- **typescript**: TypeScript support for the application.
- **cross-env**: Cross-platform environment variable setup.
## License
This project is licensed under the ISC License.
---
**Note**: Ensure your API key is managed securely in production environments. Avoid embedding sensitive information directly in your codebase.