Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/malipetek/semantic-router
https://github.com/malipetek/semantic-router
Last synced: about 1 month ago
JSON representation
- Host: GitHub
- URL: https://github.com/malipetek/semantic-router
- Owner: malipetek
- Created: 2024-02-10T18:58:26.000Z (12 months ago)
- Default Branch: main
- Last Pushed: 2024-02-10T21:33:16.000Z (12 months ago)
- Last Synced: 2024-12-12T05:04:24.363Z (about 2 months ago)
- Language: JavaScript
- Size: 111 KB
- Stars: 2
- Watchers: 1
- Forks: 0
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Semantic Router JS
A semantic router using fastembed. WIP but looks like working fine.
## Installation
You can install the package using npm:
```bash
npm install @malipetek/semantic-router
```
```bash
yarn add @malipetek/semantic-router
```
```bash
pnpm add @malipetek/semantic-router
```
## Usage
```javascript
import semantic from'@malipetek/semantic-router';
// mimicing express syntax not very useful tho
const app = semantic();
app.on('toolcall', [
'use the tool',
'retrieve data from database',
'look up from files',
'read the file'
], () => {
console.log('A tool was called');
});
await app.route('we need to query database for this information, we might need to use a tool'); // callback gets called
```
Alternate usage:
```javascript
import semantic from'@malipetek/semantic-router';
// mimicing express syntax not very useful tho
const app = semantic();
app.on('toolcall', [
'use the tool',
'retrieve data from database',
'look up from files',
'read the file'
]);
const result = await app.route('we need to query database for this information, we might need to use a tool');
console.log(result); // result will be Route onject with name, do whatever you want with it
/**
* {
* name: 'toolcall',
* data: [
* 'use the tool',
* 'retrieve data from database',
* 'look up from files',
* 'read the file'
* ]
* }
*/
```
## FastEmbed unzip issue
When you run router for the first time fastembed will download and unzip the model for creating vector embeddings, sometimes `tar` command fails to unzip all files in the archive. This is a known issue.
If you encounter this error, look up `local_cache` folder in your project directory. If the folder next to the `tar.gz` file only has 2 files, delete the folder and unzip the tar.gz manually.
```
throw new Error(`Tokenizer file not found at ${tokenizerPath}`);
```
Faulty folder:
Correct folder:
### Warning
Do not forget to add `local_cache` folder to .gitignore, it aint small:
