mirror of
https://github.com/run-llama/create-llama.git
synced 2026-07-16 11:04:26 -04:00
Compare commits
10 Commits
| Author | SHA1 | Date | |
|---|---|---|---|
| fa28cb5d0d | |||
| 8c1087f5f1 | |||
| 27333973f1 | |||
| cf3ec97a4c | |||
| 505b8e944a | |||
| 578f7f9e50 | |||
| adc40cf770 | |||
| 7bce7386d5 | |||
| c011455dc4 | |||
| 38a8be8d12 |
@@ -1,5 +1,25 @@
|
||||
# create-llama
|
||||
|
||||
## 0.2.7
|
||||
|
||||
### Patch Changes
|
||||
|
||||
- 505b8e9: bump: use latest ai package version
|
||||
- cf3ec97: Dynamically select model for Groq
|
||||
- 8c1087f: feat: enhance style for markdown
|
||||
|
||||
## 0.2.6
|
||||
|
||||
### Patch Changes
|
||||
|
||||
- adc40cf: fix: vercel ai update crash sending annotations
|
||||
|
||||
## 0.2.5
|
||||
|
||||
### Patch Changes
|
||||
|
||||
- 38a8be8: fix: filter in mongo vector store
|
||||
|
||||
## 0.2.4
|
||||
|
||||
### Patch Changes
|
||||
|
||||
@@ -3,8 +3,55 @@ import prompts from "prompts";
|
||||
import { ModelConfigParams } from ".";
|
||||
import { questionHandlers, toChoice } from "../../questions";
|
||||
|
||||
const MODELS = ["llama3-8b", "llama3-70b", "mixtral-8x7b"];
|
||||
const DEFAULT_MODEL = MODELS[0];
|
||||
import got from "got";
|
||||
import ora from "ora";
|
||||
import { red } from "picocolors";
|
||||
|
||||
const GROQ_API_URL = "https://api.groq.com/openai/v1";
|
||||
|
||||
async function getAvailableModelChoicesGroq(apiKey: string) {
|
||||
if (!apiKey) {
|
||||
throw new Error("Need Groq API key to retrieve model choices");
|
||||
}
|
||||
|
||||
const spinner = ora("Fetching available models from Groq").start();
|
||||
try {
|
||||
const response = await got(`${GROQ_API_URL}/models`, {
|
||||
headers: {
|
||||
Authorization: `Bearer ${apiKey}`,
|
||||
},
|
||||
timeout: 5000,
|
||||
responseType: "json",
|
||||
});
|
||||
const data: any = await response.body;
|
||||
spinner.stop();
|
||||
|
||||
// Filter out the Whisper models
|
||||
return data.data
|
||||
.filter((model: any) => !model.id.toLowerCase().includes("whisper"))
|
||||
.map((el: any) => {
|
||||
return {
|
||||
title: el.id,
|
||||
value: el.id,
|
||||
};
|
||||
});
|
||||
} catch (error: unknown) {
|
||||
spinner.stop();
|
||||
console.log(error);
|
||||
if ((error as any).response?.statusCode === 401) {
|
||||
console.log(
|
||||
red(
|
||||
"Invalid Groq API key provided! Please provide a valid key and try again!",
|
||||
),
|
||||
);
|
||||
} else {
|
||||
console.log(red("Request failed: " + error));
|
||||
}
|
||||
process.exit(1);
|
||||
}
|
||||
}
|
||||
|
||||
const DEFAULT_MODEL = "llama3-70b-8192";
|
||||
|
||||
// Use huggingface embedding models for now as Groq doesn't support embedding models
|
||||
enum HuggingFaceEmbeddingModelType {
|
||||
@@ -66,12 +113,14 @@ export async function askGroqQuestions({
|
||||
// use default model values in CI or if user should not be asked
|
||||
const useDefaults = ciInfo.isCI || !askModels;
|
||||
if (!useDefaults) {
|
||||
const modelChoices = await getAvailableModelChoicesGroq(config.apiKey!);
|
||||
|
||||
const { model } = await prompts(
|
||||
{
|
||||
type: "select",
|
||||
name: "model",
|
||||
message: "Which LLM model would you like to use?",
|
||||
choices: MODELS.map(toChoice),
|
||||
choices: modelChoices,
|
||||
initial: 0,
|
||||
},
|
||||
questionHandlers,
|
||||
|
||||
+1
-1
@@ -1,6 +1,6 @@
|
||||
{
|
||||
"name": "create-llama",
|
||||
"version": "0.2.4",
|
||||
"version": "0.2.7",
|
||||
"description": "Create LlamaIndex-powered apps with one command",
|
||||
"keywords": [
|
||||
"rag",
|
||||
|
||||
@@ -3,7 +3,7 @@ import mimetypes
|
||||
import os
|
||||
from io import BytesIO
|
||||
from pathlib import Path
|
||||
from typing import Any, List, Tuple
|
||||
from typing import List, Optional, Tuple
|
||||
|
||||
from app.engine.index import IndexConfig, get_index
|
||||
from llama_index.core import VectorStoreIndex
|
||||
@@ -72,7 +72,12 @@ class PrivateFileService:
|
||||
return documents
|
||||
|
||||
@staticmethod
|
||||
def process_file(file_name: str, base64_content: str, params: Any) -> List[str]:
|
||||
def process_file(
|
||||
file_name: str, base64_content: str, params: Optional[dict] = None
|
||||
) -> List[str]:
|
||||
if params is None:
|
||||
params = {}
|
||||
|
||||
file_data, extension = PrivateFileService.preprocess_base64_file(base64_content)
|
||||
|
||||
# Add the nodes to the index and persist it
|
||||
|
||||
@@ -126,13 +126,7 @@ def init_fastembed():
|
||||
def init_groq():
|
||||
from llama_index.llms.groq import Groq
|
||||
|
||||
model_map: Dict[str, str] = {
|
||||
"llama3-8b": "llama3-8b-8192",
|
||||
"llama3-70b": "llama3-70b-8192",
|
||||
"mixtral-8x7b": "mixtral-8x7b-32768",
|
||||
}
|
||||
|
||||
Settings.llm = Groq(model=model_map[os.getenv("MODEL")])
|
||||
Settings.llm = Groq(model=os.getenv("MODEL"))
|
||||
# Groq does not provide embeddings, so we use FastEmbed instead
|
||||
init_fastembed()
|
||||
|
||||
|
||||
@@ -1,14 +1,11 @@
|
||||
/* eslint-disable turbo/no-undeclared-env-vars */
|
||||
import * as dotenv from "dotenv";
|
||||
import {
|
||||
MongoDBAtlasVectorSearch,
|
||||
VectorStoreIndex,
|
||||
storageContextFromDefaults,
|
||||
} from "llamaindex";
|
||||
import { storageContextFromDefaults, VectorStoreIndex } from "llamaindex";
|
||||
import { MongoDBAtlasVectorSearch } from "llamaindex/storage/vectorStore/MongoDBAtlasVectorStore";
|
||||
import { MongoClient } from "mongodb";
|
||||
import { getDocuments } from "./loader";
|
||||
import { initSettings } from "./settings";
|
||||
import { checkRequiredEnvVars } from "./shared";
|
||||
import { checkRequiredEnvVars, POPULATED_METADATA_FIELDS } from "./shared";
|
||||
|
||||
dotenv.config();
|
||||
|
||||
@@ -30,6 +27,12 @@ async function loadAndIndex() {
|
||||
dbName: databaseName,
|
||||
collectionName: vectorCollectionName, // this is where your embeddings will be stored
|
||||
indexName: indexName, // this is the name of the index you will need to create
|
||||
indexedMetadataFields: POPULATED_METADATA_FIELDS,
|
||||
embeddingDefinition: {
|
||||
dimensions: process.env.EMBEDDING_DIM
|
||||
? parseInt(process.env.EMBEDDING_DIM)
|
||||
: 1536,
|
||||
},
|
||||
});
|
||||
|
||||
// now create an index from all the Documents and store them in Atlas
|
||||
|
||||
@@ -1,16 +1,23 @@
|
||||
/* eslint-disable turbo/no-undeclared-env-vars */
|
||||
import { MongoDBAtlasVectorSearch, VectorStoreIndex } from "llamaindex";
|
||||
import { VectorStoreIndex } from "llamaindex";
|
||||
import { MongoDBAtlasVectorSearch } from "llamaindex/storage/vectorStore/MongoDBAtlasVectorStore";
|
||||
import { MongoClient } from "mongodb";
|
||||
import { checkRequiredEnvVars } from "./shared";
|
||||
import { checkRequiredEnvVars, POPULATED_METADATA_FIELDS } from "./shared";
|
||||
|
||||
export async function getDataSource(params?: any) {
|
||||
checkRequiredEnvVars();
|
||||
const client = new MongoClient(process.env.MONGO_URI!);
|
||||
const client = new MongoClient(process.env.MONGODB_URI!);
|
||||
const store = new MongoDBAtlasVectorSearch({
|
||||
mongodbClient: client,
|
||||
dbName: process.env.MONGODB_DATABASE!,
|
||||
collectionName: process.env.MONGODB_VECTORS!,
|
||||
indexName: process.env.MONGODB_VECTOR_INDEX,
|
||||
indexedMetadataFields: POPULATED_METADATA_FIELDS,
|
||||
embeddingDefinition: {
|
||||
dimensions: process.env.EMBEDDING_DIM
|
||||
? parseInt(process.env.EMBEDDING_DIM)
|
||||
: 1536,
|
||||
},
|
||||
});
|
||||
|
||||
return await VectorStoreIndex.fromVectorStore(store);
|
||||
|
||||
@@ -5,6 +5,8 @@ const REQUIRED_ENV_VARS = [
|
||||
"MONGODB_VECTOR_INDEX",
|
||||
];
|
||||
|
||||
export const POPULATED_METADATA_FIELDS = ["private", "doc_id"]; // for filtering in MongoDB VectorSearchIndex
|
||||
|
||||
export function checkRequiredEnvVars() {
|
||||
const missingEnvVars = REQUIRED_ENV_VARS.filter((envVar) => {
|
||||
return !process.env[envVar];
|
||||
|
||||
@@ -12,7 +12,8 @@ generate = "app.engine.generate:generate_datasource"
|
||||
[tool.poetry.dependencies]
|
||||
python = "^3.11"
|
||||
llama-index-agent-openai = ">=0.3.0,<0.4.0"
|
||||
llama-index = "^0.11.4"
|
||||
llama-index = "0.11.9"
|
||||
llama-index-core = "0.11.9"
|
||||
fastapi = "^0.112.2"
|
||||
python-dotenv = "^1.0.0"
|
||||
uvicorn = { extras = ["standard"], version = "^0.23.2" }
|
||||
|
||||
@@ -15,12 +15,12 @@
|
||||
"dev": "concurrently \"tsup index.ts --format esm --dts --watch\" \"nodemon --watch dist/index.js\""
|
||||
},
|
||||
"dependencies": {
|
||||
"ai": "^3.0.21",
|
||||
"ai": "3.3.38",
|
||||
"cors": "^2.8.5",
|
||||
"dotenv": "^16.3.1",
|
||||
"duck-duck-scrape": "^2.2.5",
|
||||
"express": "^4.18.2",
|
||||
"llamaindex": "0.5.20",
|
||||
"llamaindex": "0.6.2",
|
||||
"pdf2json": "3.0.5",
|
||||
"ajv": "^8.12.0",
|
||||
"@e2b/code-interpreter": "^0.0.5",
|
||||
|
||||
@@ -138,14 +138,8 @@ function initGroq() {
|
||||
"all-mpnet-base-v2": "Xenova/all-mpnet-base-v2",
|
||||
};
|
||||
|
||||
const modelMap: Record<string, string> = {
|
||||
"llama3-8b": "llama3-8b-8192",
|
||||
"llama3-70b": "llama3-70b-8192",
|
||||
"mixtral-8x7b": "mixtral-8x7b-32768",
|
||||
};
|
||||
|
||||
Settings.llm = new Groq({
|
||||
model: modelMap[process.env.MODEL!],
|
||||
model: process.env.MODEL!,
|
||||
});
|
||||
|
||||
Settings.embedModel = new HuggingFaceEmbedding({
|
||||
|
||||
@@ -138,14 +138,8 @@ function initGroq() {
|
||||
"all-mpnet-base-v2": "Xenova/all-mpnet-base-v2",
|
||||
};
|
||||
|
||||
const modelMap: Record<string, string> = {
|
||||
"llama3-8b": "llama3-8b-8192",
|
||||
"llama3-70b": "llama3-70b-8192",
|
||||
"mixtral-8x7b": "mixtral-8x7b-32768",
|
||||
};
|
||||
|
||||
Settings.llm = new Groq({
|
||||
model: modelMap[process.env.MODEL!],
|
||||
model: process.env.MODEL!,
|
||||
});
|
||||
|
||||
Settings.embedModel = new HuggingFaceEmbedding({
|
||||
|
||||
@@ -29,6 +29,7 @@ export default function ChatSection() {
|
||||
const message = JSON.parse(error.message);
|
||||
alert(message.detail);
|
||||
},
|
||||
sendExtraMessageFields: true,
|
||||
});
|
||||
|
||||
return (
|
||||
|
||||
@@ -133,7 +133,11 @@ export default function Markdown({
|
||||
return <></>;
|
||||
}
|
||||
}
|
||||
return <a href={href}>{children}</a>;
|
||||
return (
|
||||
<a href={href} target="_blank">
|
||||
{children}
|
||||
</a>
|
||||
);
|
||||
},
|
||||
}}
|
||||
>
|
||||
|
||||
@@ -2,11 +2,15 @@
|
||||
.custom-markdown ul {
|
||||
list-style-type: disc;
|
||||
margin-left: 20px;
|
||||
margin-top: 20px;
|
||||
margin-bottom: 20px;
|
||||
}
|
||||
|
||||
.custom-markdown ol {
|
||||
list-style-type: decimal;
|
||||
margin-left: 20px;
|
||||
margin-top: 20px;
|
||||
margin-bottom: 20px;
|
||||
}
|
||||
|
||||
.custom-markdown li {
|
||||
@@ -21,3 +25,55 @@
|
||||
.custom-markdown ol ol {
|
||||
margin-left: 20px;
|
||||
}
|
||||
|
||||
.custom-markdown img {
|
||||
border-radius: 8px;
|
||||
box-shadow: 0 0 10px rgba(0, 0, 0, 0.1);
|
||||
margin: 10px 0;
|
||||
}
|
||||
|
||||
.custom-markdown a {
|
||||
text-decoration: underline;
|
||||
color: #007bff;
|
||||
}
|
||||
|
||||
.custom-markdown h1,
|
||||
h2,
|
||||
h3,
|
||||
h4,
|
||||
h5,
|
||||
h6 {
|
||||
font-weight: bold;
|
||||
margin-bottom: 20px;
|
||||
margin-top: 20px;
|
||||
}
|
||||
|
||||
.custom-markdown h6 {
|
||||
font-size: 16px;
|
||||
}
|
||||
|
||||
.custom-markdown h5 {
|
||||
font-size: 18px;
|
||||
}
|
||||
|
||||
.custom-markdown h4 {
|
||||
font-size: 20px;
|
||||
}
|
||||
|
||||
.custom-markdown h3 {
|
||||
font-size: 22px;
|
||||
}
|
||||
|
||||
.custom-markdown h2 {
|
||||
font-size: 24px;
|
||||
}
|
||||
|
||||
.custom-markdown h1 {
|
||||
font-size: 26px;
|
||||
}
|
||||
|
||||
.custom-markdown hr {
|
||||
border: 0;
|
||||
border-top: 1px solid #e1e4e8;
|
||||
margin: 20px 0;
|
||||
}
|
||||
|
||||
@@ -17,7 +17,7 @@
|
||||
"@radix-ui/react-hover-card": "^1.0.7",
|
||||
"@radix-ui/react-select": "^2.1.1",
|
||||
"@radix-ui/react-slot": "^1.0.2",
|
||||
"ai": "^3.0.21",
|
||||
"ai": "3.3.38",
|
||||
"ajv": "^8.12.0",
|
||||
"class-variance-authority": "^0.7.0",
|
||||
"clsx": "^2.1.1",
|
||||
@@ -25,7 +25,7 @@
|
||||
"duck-duck-scrape": "^2.2.5",
|
||||
"formdata-node": "^6.0.3",
|
||||
"got": "^14.4.1",
|
||||
"llamaindex": "0.5.20",
|
||||
"llamaindex": "0.6.2",
|
||||
"lucide-react": "^0.294.0",
|
||||
"next": "^14.2.4",
|
||||
"react": "^18.2.0",
|
||||
|
||||
Reference in New Issue
Block a user