-
Notifications
You must be signed in to change notification settings - Fork 40
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
- Loading branch information
Showing
1 changed file
with
2 additions
and
200 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
|
|
@@ -35,204 +35,6 @@ The following are the original authors of BaseAI: | |
| - Saad Irfan ([@MrSaadIrfan](https://twitter.com/MrSaadIrfan)) | ||
| - Ahmad Bilal ([@AhmadBilalDev](https://twitter.com/ahmadbilaldev)) | ||
|
|
||
| ## Local Development | ||
| ## Security | ||
|
|
||
| ```bash | ||
| # To install a new node module cd into pkg and pnpm install XYZ in it | ||
| # Terminal 1 | ||
| # Root | ||
| pnpm install | ||
|
|
||
| # For our dev. Compiling and watching for changes. Run all packages in dev mode. | ||
| pnpm dev:pkgs | ||
|
|
||
| # Nextjs dev | ||
| cd examples/nextjs | ||
| pnpm dev | ||
|
|
||
| # BaseAI dev inside nextjs | ||
| cd examples/nextjs | ||
| npx baseai dev | ||
|
|
||
| # Add pipe | ||
| cd examples/nextjs | ||
| npx baseai pipe | ||
|
|
||
|
|
||
| # Global link. | ||
| # pnpm link --global | ||
|
|
||
| # Terminal 2 | ||
| # Root — runs `baseai dev` live from the local package so all updates are auto loaded. | ||
| # pnpm baseai:dev | ||
|
|
||
| # Testing | ||
| # npx tsx ./packages/testing/index.ts | ||
|
|
||
| # Issues? Reset. | ||
| # if errors then go to root | ||
| pnpm -g rm baseai | ||
| pnpm clean-all | ||
| pnpm install | ||
| pnpm build:pkgs | ||
| pnpm install | ||
|
|
||
| # now back to step 1 | ||
| ``` | ||
|
|
||
| ## Creating Memory and Using it in a Pipe | ||
|
|
||
| 0. Set `OPENAI_API_KEY` for your testing script. | ||
|
|
||
| 1. Create a new memory | ||
| `baseai memory` | ||
|
|
||
| → It will give you a url where you can paste your documents. | ||
|
|
||
| 2. Paste documents to the directory it gives. | ||
|
|
||
| 3. Embed it using: | ||
|
|
||
| `baseai embed --memory="<MEMORY_NAME_FROM_STEP_1>"` | ||
|
|
||
| 4. Attach to a pipe | ||
|
|
||
| Use `baseai pipe` to create a pipe, and when it prompts you to select memory, select the memory that you just created. | ||
|
|
||
| 5. Run your testing script | ||
| Ask a question from your files. | ||
|
|
||
| ## Memory Retrieval | ||
|
|
||
| 0. Make sure you have a memory created and embedded. | ||
|
|
||
| 1. Use `baseai memory --list` to see all the memory. | ||
|
|
||
| 2. Use `baseai retrieve --memory="<MEMORY_NAME>" --query="<YOUR_QUERY>"` to retrieve similar chunks. | ||
|
|
||
| Up-to maximum of 20 chunks will be returned. | ||
|
|
||
| ## List Memory | ||
|
|
||
| `baseai memory --list` to list all the available memory. | ||
|
|
||
| ## Deploy | ||
|
|
||
| `baseai deploy` to deploy everything to Langbase. | ||
|
|
||
| > [!NOTE] | ||
| > Memory deploy is in beta currently. It support following features: | ||
| > 1. If memory is not present in Langbase, it will create a new memory. | ||
| > 2. If memory is present in Langbase, it will compare the documents. | ||
| > - If documents are same, it will skip the memory. | ||
| > - If documents in prod are subset of local, it will upload the new documents. | ||
| > - If documents in prod are superset of local, it will ask for override permission. | ||
| If you want to use memory deployment uncomment out L53-58 in `baseai/src/deploy/index.ts`. | ||
|
|
||
| ## Use a local model using Ollama | ||
|
|
||
| Follow these steps to use a local model using Ollama in BaseAI: | ||
|
|
||
| 1. Install Ollama | ||
|
|
||
| Follow the instructions in the [Ollama repository](https://github.com/ollama/ollama/blob/main/README.md) to install Ollama. | ||
|
|
||
| 2. Run a local model using Ollama. For example, let's run the small `tinyllama` model: | ||
|
|
||
| ```bash | ||
| ollama run tinyllama | ||
|
|
||
| # or even small 6MB model | ||
| ollama run ben1t0/tiny-llm | ||
| ``` | ||
|
|
||
| 1. Add the model to your pipe configuration file. | ||
|
|
||
| ```bash | ||
| ... | ||
| model: 'ollama:tinyllama', | ||
|
|
||
| # Or | ||
| model: 'ollama:ben1t0/tiny-llm', | ||
| ... | ||
| ``` | ||
|
|
||
| Run the pipe and you should see the output from the local Ollama model. | ||
|
|
||
| > [!NOTE] | ||
| > BaseAI uses the default local url `http://localhost:11434` to connect to the Ollama model. Make sure the Ollama model is running on this url. | ||
| ## Logger | ||
|
|
||
| The logging system in BaseAI allows for flexible and hierarchical control over log output. To log a message, use the `log` function from `loggerUtils`. | ||
|
|
||
| ### Usage | ||
|
|
||
| The `log` function takes three arguments: | ||
|
|
||
| ```ts | ||
| logger(category: LogCategories, value: unknown, logHeader?: string) | ||
| ``` | ||
|
|
||
| - `category`: The category of the log message. It uses a dot-notation to represent hierarchy. | ||
| - `value`: The value to be logged. | ||
| - `logHeader` (optional): A header for the log message. | ||
|
|
||
| Example usage: | ||
|
|
||
| ```ts | ||
| import { log } from './loggerUtils'; | ||
|
|
||
| logger('pipe.request.body', requestBody, 'Request Body'); | ||
| logger('pipe.response.status', responseStatus, 'Response Status'); | ||
| ``` | ||
|
|
||
| ### Log Config Options | ||
|
|
||
| Find default configuration in `packages/baseai/src/config/config-handler.ts`. Here is what it looks like: | ||
|
|
||
| ```ts | ||
| export const defaultConfig: BaseAIConfig = { | ||
| log: { | ||
| // Enable or disable logging | ||
| isEnabled: true, | ||
| // Log sensitive data | ||
| logSensitiveData: false, | ||
| // Toggle specific log categories | ||
| pipe: true, | ||
| 'pipe.completion': true, | ||
| 'pipe.request': true, | ||
| 'pipe.response': true, | ||
| tool: false, | ||
| memory: false | ||
| } | ||
| // Other default configuration options can be added here | ||
| }; | ||
| ``` | ||
|
|
||
| - `isEnabled`: Master switch to enable/disable all logging. | ||
| - `logSensitiveData`: Controls whether sensitive data should be logged. | ||
| - Category toggles: Use dot-notation to enable/disable specific categories and subcategories. | ||
|
|
||
| ### Hierarchical Categories | ||
|
|
||
| The logging system supports hierarchical categories. This means: | ||
|
|
||
| 1. If a parent category is set to `false`, all its subcategories will be disabled, regardless of their individual settings. | ||
| 2. If a parent category is not specified or set to `true`, subcategories can be individually controlled. | ||
|
|
||
| For example: | ||
|
|
||
| - If `pipe: false`, all `pipe.*` logs will be disabled. | ||
| - If `pipe: true` and `pipe.request: false`, all `pipe.request.*` logs will be disabled, but `pipe.response.*` logs will still be active (if not explicitly disabled). | ||
|
|
||
| You can add more categories as needed in your configuration. To add a new category, add it in the `config.ts` interface in both `baseai` and `core` packages. Update the default config in `config-handler.ts` if required. | ||
|
|
||
| ## Config | ||
|
|
||
| The `baseai.config.ts` file contains the configuration for BaseAI for the logger. It resides in the `baseai` folder of your project. | ||
|
|
||
| To add new config options you can modify the `defaultConfig` object in the `config-handler.ts` file in the `baseai` package. For config types, modify the `types/config.ts` interfaces in baseai and core packages. | ||
|
|
||
| Config updates are not hot-reloaded currently. You need to restart the server to apply changes. | ||
| If you've found a security vulnerability in BaseAI, please report it privately by emailing [[email protected]](mailto:[email protected]). Please do not open a public issue. For more details on Langbase [security](https://langbase.com/security) and how to report, visit this link. | ||