Add API container + static client

This commit is contained in:
Oskar Hane
2023-10-02 13:36:29 +02:00
parent d0eb8a25ed
commit 2b1bfae162
26 changed files with 2503 additions and 2 deletions

View File

@@ -2,3 +2,5 @@
!*.py
!requirements.txt
!images/*
!front-end/*
front-end/node_modules/*

21
api.Dockerfile Normal file
View File

@@ -0,0 +1,21 @@
FROM langchain/langchain
WORKDIR /app
RUN apt-get update && apt-get install -y \
build-essential \
curl \
software-properties-common \
&& rm -rf /var/lib/apt/lists/*
COPY requirements.txt .
RUN pip install --upgrade -r requirements.txt
COPY api.py .
COPY utils.py .
COPY chains.py .
HEALTHCHECK CMD curl --fail http://localhost:8504
ENTRYPOINT [ "uvicorn", "api:app", "--host", "0.0.0.0", "--port", "8504" ]

146
api.py Normal file
View File

@@ -0,0 +1,146 @@
import os
from langchain.graphs import Neo4jGraph
from dotenv import load_dotenv
from utils import (
create_vector_index,
BaseLogger,
)
from chains import (
load_embedding_model,
load_llm,
configure_llm_only_chain,
configure_qa_rag_chain,
)
from fastapi import FastAPI, Depends
from pydantic import BaseModel
from langchain.callbacks.base import BaseCallbackHandler
from threading import Thread
from queue import Queue, Empty
from collections.abc import Generator
from sse_starlette.sse import EventSourceResponse
from fastapi.middleware.cors import CORSMiddleware
import json
load_dotenv(".env")
url = os.getenv("NEO4J_URI")
username = os.getenv("NEO4J_USERNAME")
password = os.getenv("NEO4J_PASSWORD")
ollama_base_url = os.getenv("OLLAMA_BASE_URL")
embedding_model_name = os.getenv("EMBEDDING_MODEL")
llm_name = os.getenv("LLM")
# Remapping for Langchain Neo4j integration
os.environ["NEO4J_URL"] = url
embeddings, dimension = load_embedding_model(
embedding_model_name,
config={ollama_base_url: ollama_base_url},
logger=BaseLogger(),
)
# if Neo4j is local, you can go to http://localhost:7474/ to browse the database
neo4j_graph = Neo4jGraph(url=url, username=username, password=password)
create_vector_index(neo4j_graph, dimension)
llm = load_llm(
llm_name, logger=BaseLogger(), config={"ollama_base_url": ollama_base_url}
)
llm_chain = configure_llm_only_chain(llm)
rag_chain = configure_qa_rag_chain(
llm, embeddings, embeddings_store_url=url, username=username, password=password
)
class QueueCallback(BaseCallbackHandler):
"""Callback handler for streaming LLM responses to a queue."""
def __init__(self, q):
self.q = q
def on_llm_new_token(self, token: str, **kwargs) -> None:
self.q.put(token)
def on_llm_end(self, *args, **kwargs) -> None:
return self.q.empty()
def stream(cb, q) -> Generator:
job_done = object()
def task():
x = cb()
q.put(job_done)
t = Thread(target=task)
t.start()
content = ""
# Get each new token from the queue and yield for our generator
while True:
try:
next_token = q.get(True, timeout=1)
if next_token is job_done:
break
content += next_token
yield next_token, content
except Empty:
continue
app = FastAPI()
origins = ["*"]
app.add_middleware(
CORSMiddleware,
allow_origins=origins,
allow_credentials=True,
allow_methods=["*"],
allow_headers=["*"],
)
@app.get("/")
async def root():
return {"message": "Hello World"}
class Question(BaseModel):
text: str
rag: bool = False
@app.get("/query-stream")
def qstream(question: Question = Depends()):
output_function = llm_chain
if question.rag:
output_function = rag_chain
q = Queue()
def cb():
output_function(
{"question": question.text, "chat_history": []},
callbacks=[QueueCallback(q)],
)
def generate():
yield json.dumps({"init": True, "model": llm_name})
for token, _ in stream(cb, q):
yield json.dumps({"token": token})
return EventSourceResponse(generate(), media_type="text/event-stream")
@app.get("/query")
async def ask(question: Question = Depends()):
output_function = llm_chain
if question.rag:
output_function = rag_chain
result = output_function(
{"question": question.text, "chat_history": []}, callbacks=[]
)
return json.dumps({"result": result["answer"], "model": llm_name})

View File

@@ -71,6 +71,8 @@ services:
ignore:
- bot.py
- pdf_bot.py
- api.py
- front-end/
bot:
@@ -104,6 +106,8 @@ services:
ignore:
- loader.py
- pdf_bot.py
- api.py
- front-end/
ports:
- 8501:8501
@@ -138,9 +142,64 @@ services:
ignore:
- loader.py
- bot.py
- api.py
- front-end/
ports:
- 8503:8503
api:
build:
dockerfile: api.Dockerfile
volumes:
- $PWD/embedding_model:/embedding_model
environment:
- NEO4J_URI=${NEO4J_URI-neo4j://database:7687}
- NEO4J_PASSWORD=${NEO4J_PASSWORD-password}
- NEO4J_USERNAME=${NEO4J_USERNAME-neo4j}
- OPENAI_API_KEY=${OPENAI_API_KEY}
- OLLAMA_BASE_URL=${OLLAMA_BASE_URL-http://host.docker.internal:11434}
- LLM=${LLM-llama2}
- EMBEDDING_MODEL=${EMBEDDING_MODEL-sentence_transformer}
- LANGCHAIN_ENDPOINT=${LANGCHAIN_ENDPOINT-"https://api.smith.langchain.com"}
- LANGCHAIN_TRACING_V2=${LANGCHAIN_TRACING_V2-false}
- LANGCHAIN_PROJECT=${LANGCHAIN_PROJECT}
- LANGCHAIN_API_KEY=${LANGCHAIN_API_KEY}
networks:
- net
depends_on:
database:
condition: service_healthy
x-develop:
watch:
- action: rebuild
path: .
ignore:
- loader.py
- bot.py
- pdf_bot.py
- front-end/
ports:
- 8504:8504
front-end:
build:
dockerfile: front-end.Dockerfile
x-develop:
watch:
- action: sync
path: ./front-end
target: /app
ignore:
- ./front-end/node_modules/
- action: rebuild
path: ./front-end/package.json
depends_on:
api:
condition: service_healthy
networks:
- net
ports:
- 8505:8505
networks:
net:

11
front-end.Dockerfile Normal file
View File

@@ -0,0 +1,11 @@
FROM node:alpine
WORKDIR /app
COPY front-end/ .
RUN npm install
EXPOSE 8505
ENTRYPOINT [ "npm", "run", "dev" ]

24
front-end/.gitignore vendored Normal file
View File

@@ -0,0 +1,24 @@
# Logs
logs
*.log
npm-debug.log*
yarn-debug.log*
yarn-error.log*
pnpm-debug.log*
lerna-debug.log*
node_modules
dist
dist-ssr
*.local
# Editor directories and files
.vscode/*
!.vscode/extensions.json
.idea
.DS_Store
*.suo
*.ntvs*
*.njsproj
*.sln
*.sw?

3
front-end/.vscode/extensions.json vendored Normal file
View File

@@ -0,0 +1,3 @@
{
"recommendations": ["svelte.svelte-vscode"]
}

47
front-end/README.md Normal file
View File

@@ -0,0 +1,47 @@
# Svelte + Vite
This template should help get you started developing with Svelte in Vite.
## Recommended IDE Setup
[VS Code](https://code.visualstudio.com/) + [Svelte](https://marketplace.visualstudio.com/items?itemName=svelte.svelte-vscode).
## Need an official Svelte framework?
Check out [SvelteKit](https://github.com/sveltejs/kit#readme), which is also powered by Vite. Deploy anywhere with its serverless-first approach and adapt to various platforms, with out of the box support for TypeScript, SCSS, and Less, and easily-added support for mdsvex, GraphQL, PostCSS, Tailwind CSS, and more.
## Technical considerations
**Why use this over SvelteKit?**
- It brings its own routing solution which might not be preferable for some users.
- It is first and foremost a framework that just happens to use Vite under the hood, not a Vite app.
This template contains as little as possible to get started with Vite + Svelte, while taking into account the developer experience with regards to HMR and intellisense. It demonstrates capabilities on par with the other `create-vite` templates and is a good starting point for beginners dipping their toes into a Vite + Svelte project.
Should you later need the extended capabilities and extensibility provided by SvelteKit, the template has been structured similarly to SvelteKit so that it is easy to migrate.
**Why `global.d.ts` instead of `compilerOptions.types` inside `jsconfig.json` or `tsconfig.json`?**
Setting `compilerOptions.types` shuts out all other types not explicitly listed in the configuration. Using triple-slash references keeps the default TypeScript setting of accepting type information from the entire workspace, while also adding `svelte` and `vite/client` type information.
**Why include `.vscode/extensions.json`?**
Other templates indirectly recommend extensions via the README, but this file allows VS Code to prompt the user to install the recommended extension upon opening the project.
**Why enable `checkJs` in the JS template?**
It is likely that most cases of changing variable types in runtime are likely to be accidental, rather than deliberate. This provides advanced typechecking out of the box. Should you like to take advantage of the dynamically-typed nature of JavaScript, it is trivial to change the configuration.
**Why is HMR not preserving my local component state?**
HMR state preservation comes with a number of gotchas! It has been disabled by default in both `svelte-hmr` and `@sveltejs/vite-plugin-svelte` due to its often surprising behavior. You can read the details [here](https://github.com/sveltejs/svelte-hmr/tree/master/packages/svelte-hmr#preservation-of-local-state).
If you have state that's important to retain within a component, consider creating an external store which would not be replaced by HMR.
```js
// store.js
// An extremely simple external store
import { writable } from 'svelte/store'
export default writable(0)
```

13
front-end/index.html Normal file
View File

@@ -0,0 +1,13 @@
<!doctype html>
<html lang="en">
<head>
<meta charset="UTF-8" />
<link rel="icon" type="image/svg+xml" href="/vite.svg" />
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
<title>Vite + Svelte</title>
</head>
<body>
<div id="app"></div>
<script type="module" src="/src/main.js"></script>
</body>
</html>

32
front-end/jsconfig.json Normal file
View File

@@ -0,0 +1,32 @@
{
"compilerOptions": {
"moduleResolution": "bundler",
"target": "ESNext",
"module": "ESNext",
/**
* svelte-preprocess cannot figure out whether you have
* a value or a type, so tell TypeScript to enforce using
* `import type` instead of `import` for Types.
*/
"verbatimModuleSyntax": true,
"isolatedModules": true,
"resolveJsonModule": true,
/**
* To have warnings / errors of the Svelte compiler at the
* correct position, enable source maps by default.
*/
"sourceMap": true,
"esModuleInterop": true,
"skipLibCheck": true,
/**
* Typecheck JS in `.svelte` and `.js` files by default.
* Disable this if you'd like to use dynamic types.
*/
"checkJs": true
},
/**
* Use global.d.ts instead of compilerOptions.types
* to avoid limiting type declarations.
*/
"include": ["src/**/*.d.ts", "src/**/*.js", "src/**/*.svelte"]
}

1879
front-end/package-lock.json generated Normal file

File diff suppressed because it is too large Load Diff

22
front-end/package.json Normal file
View File

@@ -0,0 +1,22 @@
{
"name": "bot-ui",
"private": true,
"version": "0.0.0",
"type": "module",
"scripts": {
"dev": "vite",
"build": "vite build",
"preview": "vite preview"
},
"devDependencies": {
"@sveltejs/vite-plugin-svelte": "^2.4.2",
"autoprefixer": "^10.4.16",
"postcss": "^8.4.31",
"svelte": "^4.0.5",
"tailwindcss": "^3.3.3",
"vite": "^4.4.5"
},
"dependencies": {
"svelte-markdown": "^0.4.0"
}
}

View File

@@ -0,0 +1,6 @@
export default {
plugins: {
tailwindcss: {},
autoprefixer: {},
},
}

View File

@@ -0,0 +1 @@
<svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" class="iconify iconify--logos" width="31.88" height="32" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 257"><defs><linearGradient id="IconifyId1813088fe1fbc01fb466" x1="-.828%" x2="57.636%" y1="7.652%" y2="78.411%"><stop offset="0%" stop-color="#41D1FF"></stop><stop offset="100%" stop-color="#BD34FE"></stop></linearGradient><linearGradient id="IconifyId1813088fe1fbc01fb467" x1="43.376%" x2="50.316%" y1="2.242%" y2="89.03%"><stop offset="0%" stop-color="#FFEA83"></stop><stop offset="8.333%" stop-color="#FFDD35"></stop><stop offset="100%" stop-color="#FFA800"></stop></linearGradient></defs><path fill="url(#IconifyId1813088fe1fbc01fb466)" d="M255.153 37.938L134.897 252.976c-2.483 4.44-8.862 4.466-11.382.048L.875 37.958c-2.746-4.814 1.371-10.646 6.827-9.67l120.385 21.517a6.537 6.537 0 0 0 2.322-.004l117.867-21.483c5.438-.991 9.574 4.796 6.877 9.62Z"></path><path fill="url(#IconifyId1813088fe1fbc01fb467)" d="M185.432.063L96.44 17.501a3.268 3.268 0 0 0-2.634 3.014l-5.474 92.456a3.268 3.268 0 0 0 3.997 3.378l24.777-5.718c2.318-.535 4.413 1.507 3.936 3.838l-7.361 36.047c-.495 2.426 1.782 4.5 4.151 3.78l15.304-4.649c2.372-.72 4.652 1.36 4.15 3.788l-11.698 56.621c-.732 3.542 3.979 5.473 5.943 2.437l1.313-2.028l72.516-144.72c1.215-2.423-.88-5.186-3.54-4.672l-25.505 4.922c-2.396.462-4.435-1.77-3.759-4.114l16.646-57.705c.677-2.35-1.37-4.583-3.769-4.113Z"></path></svg>

After

Width:  |  Height:  |  Size: 1.5 KiB

155
front-end/src/App.svelte Normal file
View File

@@ -0,0 +1,155 @@
<script>
import { tick } from "svelte";
import SvelteMarkdown from "svelte-markdown";
import botImage from "./assets/images/bot.jpeg";
import meImage from "./assets/images/me.jpeg";
import MdLink from "./lib/MdLink.svelte";
let messages = [];
let ragMode = true;
let question = "How can I create a chatbot on top of my local PDF files using langchain?";
let shouldAutoScroll = true;
let input;
let appState = "idle"; // or receiving
let senderImages = { bot: botImage, me: meImage };
async function send() {
if (!question.trim().length) {
return;
}
appState = "receiving";
addMessage("me", question, ragMode);
const messageId = addMessage("bot", "", ragMode);
try {
const evt = new EventSource(
`http://localhost:8504/query-stream?text=${encodeURI(question)}&rag=${ragMode}`
);
question = "";
evt.onmessage = (e) => {
if (e.data) {
const data = JSON.parse(e.data);
if (data.init) {
updateMessage(messageId, "", data.model);
return;
}
updateMessage(messageId, data.token);
}
};
evt.onerror = (e) => {
// Stream will end with an error
// and we want to close the connection on end (otherwise it will keep reconnecting)
evt.close();
};
} catch (e) {
updateMessage(messageId, "Error: " + e.message);
} finally {
appState = "idle";
}
}
function updateMessage(existingId, text, model = null) {
if (!existingId) {
return;
}
const existingIdIndex = messages.findIndex((m) => m.id === existingId);
if (existingIdIndex === -1) {
return;
}
messages[existingIdIndex].text += text;
if (model) {
messages[existingIdIndex].model = model;
}
messages = messages;
}
function addMessage(from, text, rag) {
const newId = Math.random().toString(36).substring(2, 9);
const message = { id: newId, from, text, rag };
messages = messages.concat([message]);
return newId;
}
function scrollToBottom(node, _) {
const scroll = () => node.scrollTo({ top: node.scrollHeight });
scroll();
return { update: () => shouldAutoScroll && scroll() };
}
function scrolling(e) {
shouldAutoScroll = e.target.scrollTop + e.target.clientHeight > e.target.scrollHeight - 55;
}
$: appState === "idle" && input && focus(input);
async function focus(node) {
await tick();
node.focus();
}
// send();
</script>
<main class="h-full text-sm bg-gradient-to-t from-indigo-100 bg-fixed overflow-hidden">
<div on:scroll={scrolling} class="flex h-full flex-col py-12 overflow-y-auto" use:scrollToBottom={messages}>
<div class="w-4/5 mx-auto flex flex-col mb-32">
{#each messages as message (message.id)}
<div
class="max-w-[80%] min-w-[40%] rounded-lg p-4 mb-4 overflow-x-auto bg-white border border-indigo-200"
class:self-end={message.from === "me"}
class:text-right={message.from === "me"}
>
<div class="flex flex-row items-start gap-2">
<div
class:ml-auto={message.from === "me"}
class="relative w-12 h-12 border border-indigo-200 rounded-lg flex justify-center items-center"
>
<img
src={senderImages[message.from]}
alt=""
class="w-12 h-12 absolute top-0 left-0 rounded-lg"
/>
</div>
{#if message.from === "bot"}
<div class="text-sm">
<div>Model: {message.model ? message.model : ""}</div>
<div>RAG: {message.rag ? "Enabled" : "Disabled"}</div>
</div>
{/if}
</div>
<div class="mt-4"><SvelteMarkdown source={message.text} renderers={{ link: MdLink }} /></div>
</div>
{/each}
</div>
<div class="text-sm w-full fixed bottom-16">
<div class="shadow-lg bg-indigo-50 rounded-lg w-1/2 mx-auto">
<div class="rounded-t-lg px-4 py-2 font-light">
<div class="font-semibold">RAG mode</div>
<div class="">
<label class="mr-2">
<input type="radio" bind:group={ragMode} value={false} /> Disabled
</label>
<label>
<input type="radio" bind:group={ragMode} value={true} /> Enabled
</label>
</div>
</div>
<form class="rounded-md w-full bg-white p-2 m-0" on:submit|preventDefault={send}>
<input
disabled={appState === "receiving"}
class="text-lg w-full bg-white focus:outline-none px-4"
bind:value={question}
bind:this={input}
type="text"
/>
</form>
</div>
</div>
</div>
</main>
<style>
:global(pre) {
@apply bg-gray-100 rounded-lg p-4 border border-indigo-200;
}
:global(code) {
@apply text-indigo-500;
}
</style>

28
front-end/src/app.css Normal file
View File

@@ -0,0 +1,28 @@
@tailwind base;
@tailwind components;
@tailwind utilities;
body {
margin: 0;
min-width: 320px;
height: 100dvh;
}
#app {
height: 100%;
}
pre {
line-height: 1;
background-color: rgb(241, 241, 241);
padding: 4px 8px 8px;
border: 1px solid #ccc;
overflow-x: auto;
margin: 0 4px;
border-radius: 2px;
}
ol {
padding: 1em;
list-style: decimal;
}

Binary file not shown.

After

Width:  |  Height:  |  Size: 107 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 124 KiB

View File

@@ -0,0 +1 @@
<svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" class="iconify iconify--logos" width="26.6" height="32" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 308"><path fill="#FF3E00" d="M239.682 40.707C211.113-.182 154.69-12.301 113.895 13.69L42.247 59.356a82.198 82.198 0 0 0-37.135 55.056a86.566 86.566 0 0 0 8.536 55.576a82.425 82.425 0 0 0-12.296 30.719a87.596 87.596 0 0 0 14.964 66.244c28.574 40.893 84.997 53.007 125.787 27.016l71.648-45.664a82.182 82.182 0 0 0 37.135-55.057a86.601 86.601 0 0 0-8.53-55.577a82.409 82.409 0 0 0 12.29-30.718a87.573 87.573 0 0 0-14.963-66.244"></path><path fill="#FFF" d="M106.889 270.841c-23.102 6.007-47.497-3.036-61.103-22.648a52.685 52.685 0 0 1-9.003-39.85a49.978 49.978 0 0 1 1.713-6.693l1.35-4.115l3.671 2.697a92.447 92.447 0 0 0 28.036 14.007l2.663.808l-.245 2.659a16.067 16.067 0 0 0 2.89 10.656a17.143 17.143 0 0 0 18.397 6.828a15.786 15.786 0 0 0 4.403-1.935l71.67-45.672a14.922 14.922 0 0 0 6.734-9.977a15.923 15.923 0 0 0-2.713-12.011a17.156 17.156 0 0 0-18.404-6.832a15.78 15.78 0 0 0-4.396 1.933l-27.35 17.434a52.298 52.298 0 0 1-14.553 6.391c-23.101 6.007-47.497-3.036-61.101-22.649a52.681 52.681 0 0 1-9.004-39.849a49.428 49.428 0 0 1 22.34-33.114l71.664-45.677a52.218 52.218 0 0 1 14.563-6.398c23.101-6.007 47.497 3.036 61.101 22.648a52.685 52.685 0 0 1 9.004 39.85a50.559 50.559 0 0 1-1.713 6.692l-1.35 4.116l-3.67-2.693a92.373 92.373 0 0 0-28.037-14.013l-2.664-.809l.246-2.658a16.099 16.099 0 0 0-2.89-10.656a17.143 17.143 0 0 0-18.398-6.828a15.786 15.786 0 0 0-4.402 1.935l-71.67 45.674a14.898 14.898 0 0 0-6.73 9.975a15.9 15.9 0 0 0 2.709 12.012a17.156 17.156 0 0 0 18.404 6.832a15.841 15.841 0 0 0 4.402-1.935l27.345-17.427a52.147 52.147 0 0 1 14.552-6.397c23.101-6.006 47.497 3.037 61.102 22.65a52.681 52.681 0 0 1 9.003 39.848a49.453 49.453 0 0 1-22.34 33.12l-71.664 45.673a52.218 52.218 0 0 1-14.563 6.398"></path></svg>

After

Width:  |  Height:  |  Size: 1.9 KiB

View File

@@ -0,0 +1,7 @@
<script>
export let href = "";
export let title = "";
export let text = "";
</script>
[<a {href} {title}>{text}</a>]

8
front-end/src/main.js Normal file
View File

@@ -0,0 +1,8 @@
import './app.css'
import App from './App.svelte'
const app = new App({
target: document.getElementById('app'),
})
export default app

2
front-end/src/vite-env.d.ts vendored Normal file
View File

@@ -0,0 +1,2 @@
/// <reference types="svelte" />
/// <reference types="vite/client" />

View File

@@ -0,0 +1,7 @@
import { vitePreprocess } from '@sveltejs/vite-plugin-svelte'
export default {
// Consult https://svelte.dev/docs#compile-time-svelte-preprocess
// for more information about preprocessors
preprocess: vitePreprocess(),
}

View File

@@ -0,0 +1,12 @@
/** @type {import('tailwindcss').Config} */
export default {
content: [
"./index.html",
"./src/**/*.{svelte,js,ts,jsx,tsx}",
],
theme: {
extend: {},
},
plugins: [],
}

11
front-end/vite.config.js Normal file
View File

@@ -0,0 +1,11 @@
import { defineConfig } from 'vite'
import { svelte } from '@sveltejs/vite-plugin-svelte'
// https://vitejs.dev/config/
export default defineConfig({
server: {
host: '0.0.0.0',
port: 8505,
},
plugins: [svelte()],
})

View File

@@ -6,5 +6,9 @@ neo4j
streamlit
sentence_transformers==2.2.2
Pillow
fastapi
PyPDF2
torch==2.0.1
torch==2.0.1
pydantic
uvicorn
sse-starlette