Unverified Commit 9763d885 authored by lainedfles's avatar lainedfles Committed by GitHub
Browse files

Merge Updates & Dockerfile improvements

parent fdef2abd
name: Python CI
on:
push:
branches: ['main']
branches:
- main
- dev
pull_request:
branches:
- main
- dev
jobs:
build:
name: 'Format Backend'
env:
PUBLIC_API_BASE_URL: ''
runs-on: ubuntu-latest
strategy:
matrix:
node-version:
- latest
python-version: [3.11]
steps:
- uses: actions/checkout@v4
- name: Use Python
uses: actions/setup-python@v4
- name: Use Bun
uses: oven-sh/setup-bun@v1
- name: Set up Python
uses: actions/setup-python@v2
with:
python-version: ${{ matrix.python-version }}
- name: Install dependencies
run: |
python -m pip install --upgrade pip
pip install yapf
pip install black
- name: Format backend
run: bun run format:backend
run: npm run format:backend
- name: Check for changes after format
run: git diff --exit-code
name: Bun CI
name: Frontend Build
on:
push:
branches: ['main']
branches:
- main
- dev
pull_request:
branches:
- main
- dev
jobs:
build:
name: 'Format & Build Frontend'
env:
PUBLIC_API_BASE_URL: ''
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- name: Use Bun
uses: oven-sh/setup-bun@v1
- run: bun --version
- name: Install frontend dependencies
run: bun install
- name: Format frontend
run: bun run format
- name: Build frontend
run: bun run build
- name: Checkout Repository
uses: actions/checkout@v4
- name: Setup Node.js
uses: actions/setup-node@v3
with:
node-version: '20' # Or specify any other version you want to use
- name: Install Dependencies
run: npm install
- name: Format Frontend
run: npm run format
- name: Check for Changes After Format
run: git diff --exit-code
- name: Build Frontend
run: npm run build
......@@ -166,7 +166,7 @@ cython_debug/
# be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
# and can be added to the global gitignore or merged into this file. For a more nuclear
# option (not recommended) you can uncomment the following to ignore the entire idea folder.
#.idea/
.idea/
# Logs
logs
......
......@@ -5,6 +5,86 @@ All notable changes to this project will be documented in this file.
The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
## [0.1.116] - 2024-03-31
### Added
- **🔄 Enhanced UI**: Model selector now conveniently located in the navbar, enabling seamless switching between multiple models during conversations.
- **🔍 Improved Model Selector**: Directly pull a model from the selector/Models now display detailed information for better understanding.
- **💬 Webhook Support**: Now compatible with Google Chat and Microsoft Teams.
- **🌐 Localization**: Korean translation (I18n) now available.
- **🌑 Dark Theme**: OLED dark theme introduced for reduced strain during prolonged usage.
- **🏷️ Tag Autocomplete**: Dropdown feature added for effortless chat tagging.
### Fixed
- **🔽 Auto-Scrolling**: Addressed OpenAI auto-scrolling issue.
- **🏷️ Tag Validation**: Implemented tag validation to prevent empty string tags.
- **🚫 Model Whitelisting**: Resolved LiteLLM model whitelisting issue.
- **✅ Spelling**: Corrected various spelling issues for improved readability.
## [0.1.115] - 2024-03-24
### Added
- **🔍 Custom Model Selector**: Easily find and select custom models with the new search filter feature.
- **🛑 Cancel Model Download**: Added the ability to cancel model downloads.
- **🎨 Image Generation ComfyUI**: Image generation now supports ComfyUI.
- **🌟 Updated Light Theme**: Updated the light theme for a fresh look.
- **🌍 Additional Language Support**: Now supporting Bulgarian, Italian, Portuguese, Japanese, and Dutch.
### Fixed
- **🔧 Fixed Broken Experimental GGUF Upload**: Resolved issues with experimental GGUF upload functionality.
### Changed
- **🔄 Vector Storage Reset Button**: Moved the reset vector storage button to document settings.
## [0.1.114] - 2024-03-20
### Added
- **🔗 Webhook Integration**: Now you can subscribe to new user sign-up events via webhook. Simply navigate to the admin panel > admin settings > webhook URL.
- **🛡️ Enhanced Model Filtering**: Alongside Ollama, OpenAI proxy model whitelisting, we've added model filtering functionality for LiteLLM proxy.
- **🌍 Expanded Language Support**: Spanish, Catalan, and Vietnamese languages are now available, with improvements made to others.
### Fixed
- **🔧 Input Field Spelling**: Resolved issue with spelling mistakes in input fields.
- **🖊️ Light Mode Styling**: Fixed styling issue with light mode in document adding.
### Changed
- **🔄 Language Sorting**: Languages are now sorted alphabetically by their code for improved organization.
## [0.1.113] - 2024-03-18
### Added
- 🌍 **Localization**: You can now change the UI language in Settings > General. We support Ukrainian, German, Farsi (Persian), Traditional and Simplified Chinese and French translations. You can help us to translate the UI into your language! More info in our [CONTRIBUTION.md](https://github.com/open-webui/open-webui/blob/main/docs/CONTRIBUTING.md#-translations-and-internationalization).
- 🎨 **System-wide Theme**: Introducing a new system-wide theme for enhanced visual experience.
### Fixed
- 🌑 **Dark Background on Select Fields**: Improved readability by adding a dark background to select fields, addressing issues on certain browsers/devices.
- **Multiple OPENAI_API_BASE_URLS Issue**: Resolved issue where multiple base URLs caused conflicts when one wasn't functioning.
- **RAG Encoding Issue**: Fixed encoding problem in RAG.
- **npm Audit Fix**: Addressed npm audit findings.
- **Reduced Scroll Threshold**: Improved auto-scroll experience by reducing the scroll threshold from 50px to 5px.
### Changed
- 🔄 **Sidebar UI Update**: Updated sidebar UI to feature a chat menu dropdown, replacing two icons for improved navigation.
## [0.1.112] - 2024-03-15
### Fixed
- 🗨️ Resolved chat malfunction after image generation.
- 🎨 Fixed various RAG issues.
- 🧪 Rectified experimental broken GGUF upload logic.
## [0.1.111] - 2024-03-10
### Added
......
......@@ -2,6 +2,8 @@
# Initialize device type args
# use build args in the docker build commmand with --build-arg="BUILDARG=true"
ARG USE_CUDA=false
ARG USE_CUDA_VER=cu121
ARG USE_EMBEDDING_MODEL=all-MiniLM-L6-v2
ARG USE_MPS=false
ARG INCLUDE_OLLAMA=false
......@@ -28,8 +30,9 @@ RUN npm run build
######## WebUI backend ########
FROM python:3.11-slim-bookworm as base
# Use args
ARG USE_CUDA
ARG USE_CUDA_VER
ARG USE_EMBEDDING_MODEL
ARG USE_MPS
ARG INCLUDE_OLLAMA
......@@ -39,7 +42,9 @@ ENV ENV=prod \
# pass build args to the build
INCLUDE_OLLAMA_DOCKER=${INCLUDE_OLLAMA} \
USE_MPS_DOCKER=${USE_MPS} \
USE_CUDA_DOCKER=${USE_CUDA}
USE_CUDA_DOCKER=${USE_CUDA} \
USE_CUDA_DOCKER_VER=${USE_CUDA_VER} \
USE_EMBEDDING_MODEL_DOCKER=${USE_EMBEDDING_MODEL}
## Basis URL Config ##
ENV OLLAMA_BASE_URL="/ollama" \
......@@ -61,7 +66,7 @@ ENV WHISPER_MODEL="base" \
# Leaderboard: https://huggingface.co/spaces/mteb/leaderboard
# for better performance and multilangauge support use "intfloat/multilingual-e5-large" (~2.5GB) or "intfloat/multilingual-e5-base" (~1.5GB)
# IMPORTANT: If you change the default model (all-MiniLM-L6-v2) and vice versa, you aren't able to use RAG Chat with your previous documents loaded in the WebUI! You need to re-embed them.
ENV RAG_EMBEDDING_MODEL="all-MiniLM-L6-v2" \
ENV RAG_EMBEDDING_MODEL="$USE_EMBEDDING_MODEL_DOCKER" \
RAG_EMBEDDING_MODEL_DIR="/app/backend/data/cache/embedding/models" \
SENTENCE_TRANSFORMERS_HOME="/app/backend/data/cache/embedding/models" \
# device type for whisper tts and embbeding models - "cpu" (default) or "mps" (apple silicon) - choosing this right can lead to better performance
......@@ -78,8 +83,10 @@ WORKDIR /app/backend
COPY ./backend/requirements.txt ./requirements.txt
RUN if [ "$USE_CUDA" = "true" ]; then \
pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu117 --no-cache-dir && \
pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/$USE_CUDA_DOCKER_VER --no-cache-dir && \
pip3 install -r requirements.txt --no-cache-dir; \
python -c "import os; from faster_whisper import WhisperModel; WhisperModel(os.environ['WHISPER_MODEL'], device='cpu', compute_type='int8', download_root=os.environ['WHISPER_MODEL_DIR'])" && \
python -c "import os; from chromadb.utils import embedding_functions; sentence_transformer_ef = embedding_functions.SentenceTransformerEmbeddingFunction(model_name=os.environ['RAG_EMBEDDING_MODEL'], device='cpu')"; \
elif [ "$USE_MPS" = "true" ]; then \
pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cpu --no-cache-dir && \
pip3 install -r requirements.txt --no-cache-dir && \
......@@ -131,4 +138,4 @@ COPY ./backend .
EXPOSE 8080
CMD [ "bash", "start.sh"]
\ No newline at end of file
CMD [ "bash", "start.sh"]
......@@ -8,6 +8,8 @@ remove:
start:
@docker-compose start
startAndBuild:
docker-compose up -d --build
stop:
@docker-compose stop
......
......@@ -11,7 +11,7 @@
[![Discord](https://img.shields.io/badge/Discord-Open_WebUI-blue?logo=discord&logoColor=white)](https://discord.gg/5rJgQTnV4s)
[![](https://img.shields.io/static/v1?label=Sponsor&message=%E2%9D%A4&logo=GitHub&color=%23fe8e86)](https://github.com/sponsors/tjbck)
User-friendly WebUI for LLMs, supported LLM runners include Ollama and OpenAI-compatible APIs. For more information, be sure to check out our [Open WebUI Documentation](https://docs.openwebui.com/).
Open WebUI is an extensible, feature-rich, and user-friendly self-hosted WebUI designed to operate entirely offline. It supports various LLM runners, including Ollama and OpenAI-compatible APIs. For more information, be sure to check out our [Open WebUI Documentation](https://docs.openwebui.com/).
![Open WebUI Demo](./demo.gif)
......@@ -79,6 +79,8 @@ User-friendly WebUI for LLMs, supported LLM runners include Ollama and OpenAI-co
- 🔒 **Backend Reverse Proxy Support**: Bolster security through direct communication between Open WebUI backend and Ollama. This key feature eliminates the need to expose Ollama over LAN. Requests made to the '/ollama/api' route from the web UI are seamlessly redirected to Ollama from the backend, enhancing overall system security.
- 🌐🌍 **Multilingual Support**: Experience Open WebUI in your preferred language with our internationalization (i18n) support. Join us in expanding our supported languages! We're actively seeking contributors!
- 🌟 **Continuous Updates**: We are committed to improving Open WebUI with regular updates and new features.
## 🔗 Also Check Out Open WebUI Community!
......
import os
import logging
from fastapi import (
FastAPI,
Request,
......@@ -21,11 +22,24 @@ from utils.utils import (
)
from utils.misc import calculate_sha256
from config import CACHE_DIR, UPLOAD_DIR, WHISPER_MODEL, WHISPER_MODEL_DIR, DEVICE_TYPE
from config import (
SRC_LOG_LEVELS,
CACHE_DIR,
UPLOAD_DIR,
WHISPER_MODEL,
WHISPER_MODEL_DIR,
DEVICE_TYPE,
)
log = logging.getLogger(__name__)
log.setLevel(SRC_LOG_LEVELS["AUDIO"])
whisper_device_type = DEVICE_TYPE
if DEVICE_TYPE != "cuda":
if whisper_device_type != "cuda":
whisper_device_type = "cpu"
log.info(f"whisper_device_type: {whisper_device_type}")
app = FastAPI()
app.add_middleware(
......@@ -42,7 +56,7 @@ def transcribe(
file: UploadFile = File(...),
user=Depends(get_current_user),
):
print(file.content_type)
log.info(f"file.content_type: {file.content_type}")
if file.content_type not in ["audio/mpeg", "audio/wav"]:
raise HTTPException(
......@@ -66,7 +80,7 @@ def transcribe(
)
segments, info = model.transcribe(file_path, beam_size=5)
print(
log.info(
"Detected language '%s' with probability %f"
% (info.language, info.language_probability)
)
......@@ -76,7 +90,7 @@ def transcribe(
return {"text": transcript.strip()}
except Exception as e:
print(e)
log.exception(e)
raise HTTPException(
status_code=status.HTTP_400_BAD_REQUEST,
......
......@@ -18,6 +18,8 @@ from utils.utils import (
get_current_user,
get_admin_user,
)
from apps.images.utils.comfyui import ImageGenerationPayload, comfyui_generate_image
from utils.misc import calculate_sha256
from typing import Optional
from pydantic import BaseModel
......@@ -25,9 +27,13 @@ from pathlib import Path
import uuid
import base64
import json
import logging
from config import SRC_LOG_LEVELS, CACHE_DIR, AUTOMATIC1111_BASE_URL, COMFYUI_BASE_URL
from config import CACHE_DIR, AUTOMATIC1111_BASE_URL
log = logging.getLogger(__name__)
log.setLevel(SRC_LOG_LEVELS["IMAGES"])
IMAGE_CACHE_DIR = Path(CACHE_DIR).joinpath("./image/generations/")
IMAGE_CACHE_DIR.mkdir(parents=True, exist_ok=True)
......@@ -49,6 +55,8 @@ app.state.MODEL = ""
app.state.AUTOMATIC1111_BASE_URL = AUTOMATIC1111_BASE_URL
app.state.COMFYUI_BASE_URL = COMFYUI_BASE_URL
app.state.IMAGE_SIZE = "512x512"
app.state.IMAGE_STEPS = 50
......@@ -71,32 +79,48 @@ async def update_config(form_data: ConfigUpdateForm, user=Depends(get_admin_user
return {"engine": app.state.ENGINE, "enabled": app.state.ENABLED}
class UrlUpdateForm(BaseModel):
url: str
class EngineUrlUpdateForm(BaseModel):
AUTOMATIC1111_BASE_URL: Optional[str] = None
COMFYUI_BASE_URL: Optional[str] = None
@app.get("/url")
async def get_automatic1111_url(user=Depends(get_admin_user)):
return {"AUTOMATIC1111_BASE_URL": app.state.AUTOMATIC1111_BASE_URL}
async def get_engine_url(user=Depends(get_admin_user)):
return {
"AUTOMATIC1111_BASE_URL": app.state.AUTOMATIC1111_BASE_URL,
"COMFYUI_BASE_URL": app.state.COMFYUI_BASE_URL,
}
@app.post("/url/update")
async def update_automatic1111_url(
form_data: UrlUpdateForm, user=Depends(get_admin_user)
async def update_engine_url(
form_data: EngineUrlUpdateForm, user=Depends(get_admin_user)
):
if form_data.url == "":
if form_data.AUTOMATIC1111_BASE_URL == None:
app.state.AUTOMATIC1111_BASE_URL = AUTOMATIC1111_BASE_URL
else:
url = form_data.url.strip("/")
url = form_data.AUTOMATIC1111_BASE_URL.strip("/")
try:
r = requests.head(url)
app.state.AUTOMATIC1111_BASE_URL = url
except Exception as e:
raise HTTPException(status_code=400, detail=ERROR_MESSAGES.DEFAULT(e))
if form_data.COMFYUI_BASE_URL == None:
app.state.COMFYUI_BASE_URL = COMFYUI_BASE_URL
else:
url = form_data.COMFYUI_BASE_URL.strip("/")
try:
r = requests.head(url)
app.state.COMFYUI_BASE_URL = url
except Exception as e:
raise HTTPException(status_code=400, detail=ERROR_MESSAGES.DEFAULT(e))
return {
"AUTOMATIC1111_BASE_URL": app.state.AUTOMATIC1111_BASE_URL,
"COMFYUI_BASE_URL": app.state.COMFYUI_BASE_URL,
"status": True,
}
......@@ -186,6 +210,18 @@ def get_models(user=Depends(get_current_user)):
{"id": "dall-e-2", "name": "DALL·E 2"},
{"id": "dall-e-3", "name": "DALL·E 3"},
]
elif app.state.ENGINE == "comfyui":
r = requests.get(url=f"{app.state.COMFYUI_BASE_URL}/object_info")
info = r.json()
return list(
map(
lambda model: {"id": model, "name": model},
info["CheckpointLoaderSimple"]["input"]["required"]["ckpt_name"][0],
)
)
else:
r = requests.get(
url=f"{app.state.AUTOMATIC1111_BASE_URL}/sdapi/v1/sd-models"
......@@ -207,6 +243,8 @@ async def get_default_model(user=Depends(get_admin_user)):
try:
if app.state.ENGINE == "openai":
return {"model": app.state.MODEL if app.state.MODEL else "dall-e-2"}
elif app.state.ENGINE == "comfyui":
return {"model": app.state.MODEL if app.state.MODEL else ""}
else:
r = requests.get(url=f"{app.state.AUTOMATIC1111_BASE_URL}/sdapi/v1/options")
options = r.json()
......@@ -221,10 +259,12 @@ class UpdateModelForm(BaseModel):
def set_model_handler(model: str):
if app.state.ENGINE == "openai":
app.state.MODEL = model
return app.state.MODEL
if app.state.ENGINE == "comfyui":
app.state.MODEL = model
return app.state.MODEL
else:
r = requests.get(url=f"{app.state.AUTOMATIC1111_BASE_URL}/sdapi/v1/options")
options = r.json()
......@@ -268,7 +308,24 @@ def save_b64_image(b64_str):
return image_id
except Exception as e:
print(f"Error saving image: {e}")
log.error(f"Error saving image: {e}")
return None
def save_url_image(url):
image_id = str(uuid.uuid4())
file_path = IMAGE_CACHE_DIR.joinpath(f"{image_id}.png")
try:
r = requests.get(url)
r.raise_for_status()
with open(file_path, "wb") as image_file:
image_file.write(r.content)
return image_id
except Exception as e:
log.exception(f"Error saving image: {e}")
return None
......@@ -278,6 +335,8 @@ def generate_image(
user=Depends(get_current_user),
):
width, height = tuple(map(int, app.state.IMAGE_SIZE.split("x")))
r = None
try:
if app.state.ENGINE == "openai":
......@@ -293,6 +352,7 @@ def generate_image(
"size": form_data.size if form_data.size else app.state.IMAGE_SIZE,
"response_format": "b64_json",
}
r = requests.post(
url=f"https://api.openai.com/v1/images/generations",
json=data,
......@@ -300,7 +360,6 @@ def generate_image(
)
r.raise_for_status()
res = r.json()
images = []
......@@ -315,12 +374,47 @@ def generate_image(
return images
elif app.state.ENGINE == "comfyui":
data = {
"prompt": form_data.prompt,
"width": width,
"height": height,
"n": form_data.n,
}
if app.state.IMAGE_STEPS != None:
data["steps"] = app.state.IMAGE_STEPS
if form_data.negative_prompt != None:
data["negative_prompt"] = form_data.negative_prompt
data = ImageGenerationPayload(**data)
res = comfyui_generate_image(
app.state.MODEL,
data,
user.id,
app.state.COMFYUI_BASE_URL,
)
log.debug(f"res: {res}")
images = []
for image in res["data"]:
image_id = save_url_image(image["url"])
images.append({"url": f"/cache/image/generations/{image_id}.png"})
file_body_path = IMAGE_CACHE_DIR.joinpath(f"{image_id}.json")
with open(file_body_path, "w") as f:
json.dump(data.model_dump(exclude_none=True), f)
log.debug(f"images: {images}")
return images
else:
if form_data.model:
set_model_handler(form_data.model)
width, height = tuple(map(int, app.state.IMAGE_SIZE.split("x")))
data = {
"prompt": form_data.prompt,
"batch_size": form_data.n,
......@@ -341,7 +435,7 @@ def generate_image(
res = r.json()
print(res)
log.debug(f"res: {res}")
images = []
......@@ -356,7 +450,10 @@ def generate_image(
return images
except Exception as e:
print(e)
if r:
print(r.json())
raise HTTPException(status_code=400, detail=ERROR_MESSAGES.DEFAULT(e))
error = e
if r != None:
data = r.json()
if "error" in data:
error = data["error"]["message"]
raise HTTPException(status_code=400, detail=ERROR_MESSAGES.DEFAULT(error))
import websocket # NOTE: websocket-client (https://github.com/websocket-client/websocket-client)
import uuid
import json
import urllib.request
import urllib.parse
import random
import logging
from config import SRC_LOG_LEVELS
log = logging.getLogger(__name__)
log.setLevel(SRC_LOG_LEVELS["COMFYUI"])
from pydantic import BaseModel
from typing import Optional
COMFYUI_DEFAULT_PROMPT = """
{
"3": {
"inputs": {
"seed": 0,
"steps": 20,
"cfg": 8,
"sampler_name": "euler",
"scheduler": "normal",
"denoise": 1,
"model": [
"4",
0
],
"positive": [
"6",
0
],
"negative": [
"7",
0
],
"latent_image": [
"5",
0
]
},
"class_type": "KSampler",
"_meta": {
"title": "KSampler"
}
},
"4": {
"inputs": {
"ckpt_name": "model.safetensors"
},
"class_type": "CheckpointLoaderSimple",
"_meta": {
"title": "Load Checkpoint"
}
},
"5": {
"inputs": {
"width": 512,
"height": 512,
"batch_size": 1
},
"class_type": "EmptyLatentImage",
"_meta": {
"title": "Empty Latent Image"
}
},
"6": {
"inputs": {
"text": "Prompt",
"clip": [
"4",
1
]
},
"class_type": "CLIPTextEncode",
"_meta": {
"title": "CLIP Text Encode (Prompt)"
}
},
"7": {
"inputs": {
"text": "Negative Prompt",
"clip": [
"4",
1
]
},
"class_type": "CLIPTextEncode",
"_meta": {
"title": "CLIP Text Encode (Prompt)"
}
},
"8": {
"inputs": {
"samples": [
"3",
0
],
"vae": [
"4",
2
]
},
"class_type": "VAEDecode",
"_meta": {
"title": "VAE Decode"
}
},
"9": {
"inputs": {
"filename_prefix": "ComfyUI",
"images": [
"8",
0
]
},
"class_type": "SaveImage",
"_meta": {
"title": "Save Image"
}
}
}
"""
def queue_prompt(prompt, client_id, base_url):
log.info("queue_prompt")
p = {"prompt": prompt, "client_id": client_id}
data = json.dumps(p).encode("utf-8")
req = urllib.request.Request(f"{base_url}/prompt", data=data)
return json.loads(urllib.request.urlopen(req).read())
def get_image(filename, subfolder, folder_type, base_url):
log.info("get_image")
data = {"filename": filename, "subfolder": subfolder, "type": folder_type}
url_values = urllib.parse.urlencode(data)
with urllib.request.urlopen(f"{base_url}/view?{url_values}") as response:
return response.read()
def get_image_url(filename, subfolder, folder_type, base_url):
log.info("get_image")
data = {"filename": filename, "subfolder": subfolder, "type": folder_type}
url_values = urllib.parse.urlencode(data)
return f"{base_url}/view?{url_values}"
def get_history(prompt_id, base_url):
log.info("get_history")
with urllib.request.urlopen(f"{base_url}/history/{prompt_id}") as response:
return json.loads(response.read())
def get_images(ws, prompt, client_id, base_url):
prompt_id = queue_prompt(prompt, client_id, base_url)["prompt_id"]
output_images = []
while True:
out = ws.recv()
if isinstance(out, str):
message = json.loads(out)
if message["type"] == "executing":
data = message["data"]
if data["node"] is None and data["prompt_id"] == prompt_id:
break # Execution is done
else:
continue # previews are binary data
history = get_history(prompt_id, base_url)[prompt_id]
for o in history["outputs"]:
for node_id in history["outputs"]:
node_output = history["outputs"][node_id]
if "images" in node_output:
for image in node_output["images"]:
url = get_image_url(
image["filename"], image["subfolder"], image["type"], base_url
)
output_images.append({"url": url})
return {"data": output_images}
class ImageGenerationPayload(BaseModel):
prompt: str
negative_prompt: Optional[str] = ""
steps: Optional[int] = None
seed: Optional[int] = None
width: int
height: int
n: int = 1
def comfyui_generate_image(
model: str, payload: ImageGenerationPayload, client_id, base_url
):
host = base_url.replace("http://", "").replace("https://", "")
comfyui_prompt = json.loads(COMFYUI_DEFAULT_PROMPT)
comfyui_prompt["4"]["inputs"]["ckpt_name"] = model
comfyui_prompt["5"]["inputs"]["batch_size"] = payload.n
comfyui_prompt["5"]["inputs"]["width"] = payload.width
comfyui_prompt["5"]["inputs"]["height"] = payload.height
# set the text prompt for our positive CLIPTextEncode
comfyui_prompt["6"]["inputs"]["text"] = payload.prompt
comfyui_prompt["7"]["inputs"]["text"] = payload.negative_prompt
if payload.steps:
comfyui_prompt["3"]["inputs"]["steps"] = payload.steps
comfyui_prompt["3"]["inputs"]["seed"] = (
payload.seed if payload.seed else random.randint(0, 18446744073709551614)
)
try:
ws = websocket.WebSocket()
ws.connect(f"ws://{host}/ws?clientId={client_id}")
log.info("WebSocket connection established.")
except Exception as e:
log.exception(f"Failed to connect to WebSocket server: {e}")
return None
try:
images = get_images(ws, comfyui_prompt, client_id, base_url)
except Exception as e:
log.exception(f"Error while receiving images: {e}")
images = None
ws.close()
return images
import logging
from litellm.proxy.proxy_server import ProxyConfig, initialize
from litellm.proxy.proxy_server import app
from fastapi import FastAPI, Request, Depends, status
from fastapi import FastAPI, Request, Depends, status, Response
from fastapi.responses import JSONResponse
from starlette.middleware.base import BaseHTTPMiddleware, RequestResponseEndpoint
from starlette.responses import StreamingResponse
import json
from utils.utils import get_http_authorization_cred, get_current_user
from config import ENV
from config import SRC_LOG_LEVELS, ENV
log = logging.getLogger(__name__)
log.setLevel(SRC_LOG_LEVELS["LITELLM"])
from config import (
MODEL_FILTER_ENABLED,
MODEL_FILTER_LIST,
)
proxy_config = ProxyConfig()
......@@ -26,16 +43,58 @@ async def on_startup():
await startup()
app.state.MODEL_FILTER_ENABLED = MODEL_FILTER_ENABLED
app.state.MODEL_FILTER_LIST = MODEL_FILTER_LIST
@app.middleware("http")
async def auth_middleware(request: Request, call_next):
auth_header = request.headers.get("Authorization", "")
request.state.user = None
if ENV != "dev":
try:
user = get_current_user(get_http_authorization_cred(auth_header))
print(user)
except Exception as e:
return JSONResponse(status_code=400, content={"detail": str(e)})
try:
user = get_current_user(get_http_authorization_cred(auth_header))
log.debug(f"user: {user}")
request.state.user = user
except Exception as e:
return JSONResponse(status_code=400, content={"detail": str(e)})
response = await call_next(request)
return response
class ModifyModelsResponseMiddleware(BaseHTTPMiddleware):
async def dispatch(
self, request: Request, call_next: RequestResponseEndpoint
) -> Response:
response = await call_next(request)
user = request.state.user
if "/models" in request.url.path:
if isinstance(response, StreamingResponse):
# Read the content of the streaming response
body = b""
async for chunk in response.body_iterator:
body += chunk
data = json.loads(body.decode("utf-8"))
if app.state.MODEL_FILTER_ENABLED:
if user and user.role == "user":
data["data"] = list(
filter(
lambda model: model["id"]
in app.state.MODEL_FILTER_LIST,
data["data"],
)
)
# Modified Flag
data["modified"] = True
return JSONResponse(content=data)
return response
app.add_middleware(ModifyModelsResponseMiddleware)
This diff is collapsed.
......@@ -6,6 +6,7 @@ import requests
import aiohttp
import asyncio
import json
import logging
from pydantic import BaseModel
......@@ -19,6 +20,7 @@ from utils.utils import (
get_admin_user,
)
from config import (
SRC_LOG_LEVELS,
OPENAI_API_BASE_URLS,
OPENAI_API_KEYS,
CACHE_DIR,
......@@ -31,6 +33,9 @@ from typing import List, Optional
import hashlib
from pathlib import Path
log = logging.getLogger(__name__)
log.setLevel(SRC_LOG_LEVELS["OPENAI"])
app = FastAPI()
app.add_middleware(
CORSMiddleware,
......@@ -111,6 +116,7 @@ async def speech(request: Request, user=Depends(get_verified_user)):
headers["Authorization"] = f"Bearer {app.state.OPENAI_API_KEYS[idx]}"
headers["Content-Type"] = "application/json"
r = None
try:
r = requests.post(
url=f"{app.state.OPENAI_API_BASE_URLS[idx]}/audio/speech",
......@@ -133,7 +139,7 @@ async def speech(request: Request, user=Depends(get_verified_user)):
return FileResponse(file_path)
except Exception as e:
print(e)
log.exception(e)
error_detail = "Open WebUI: Server Connection Error"
if r is not None:
try:
......@@ -143,7 +149,9 @@ async def speech(request: Request, user=Depends(get_verified_user)):
except:
error_detail = f"External: {e}"
raise HTTPException(status_code=r.status_code, detail=error_detail)
raise HTTPException(
status_code=r.status_code if r else 500, detail=error_detail
)
except ValueError:
raise HTTPException(status_code=401, detail=ERROR_MESSAGES.OPENAI_NOT_FOUND)
......@@ -157,7 +165,7 @@ async def fetch_url(url, key):
return await response.json()
except Exception as e:
# Handle connection error here
print(f"Connection error: {e}")
log.error(f"Connection error: {e}")
return None
......@@ -165,20 +173,21 @@ def merge_models_lists(model_lists):
merged_list = []
for idx, models in enumerate(model_lists):
merged_list.extend(
[
{**model, "urlIdx": idx}
for model in models
if "api.openai.com" not in app.state.OPENAI_API_BASE_URLS[idx]
or "gpt" in model["id"]
]
)
if models is not None and "error" not in models:
merged_list.extend(
[
{**model, "urlIdx": idx}
for model in models
if "api.openai.com" not in app.state.OPENAI_API_BASE_URLS[idx]
or "gpt" in model["id"]
]
)
return merged_list
async def get_all_models():
print("get_all_models")
log.info("get_all_models()")
if len(app.state.OPENAI_API_KEYS) == 1 and app.state.OPENAI_API_KEYS[0] == "":
models = {"data": []}
......@@ -187,15 +196,24 @@ async def get_all_models():
fetch_url(f"{url}/models", app.state.OPENAI_API_KEYS[idx])
for idx, url in enumerate(app.state.OPENAI_API_BASE_URLS)
]
responses = await asyncio.gather(*tasks)
responses = list(
filter(lambda x: x is not None and "error" not in x, responses)
)
models = {
"data": merge_models_lists(
list(map(lambda response: response["data"], responses))
list(
map(
lambda response: (
response["data"]
if response and "data" in response
else None
),
responses,
)
)
)
}
log.info(f"models: {models}")
app.state.MODELS = {model["id"]: model for model in models["data"]}
return models
......@@ -218,6 +236,9 @@ async def get_models(url_idx: Optional[int] = None, user=Depends(get_current_use
return models
else:
url = app.state.OPENAI_API_BASE_URLS[url_idx]
r = None
try:
r = requests.request(method="GET", url=f"{url}/models")
r.raise_for_status()
......@@ -230,7 +251,7 @@ async def get_models(url_idx: Optional[int] = None, user=Depends(get_current_use
return response_data
except Exception as e:
print(e)
log.exception(e)
error_detail = "Open WebUI: Server Connection Error"
if r is not None:
try:
......@@ -264,7 +285,7 @@ async def proxy(path: str, request: Request, user=Depends(get_verified_user)):
if body.get("model") == "gpt-4-vision-preview":
if "max_tokens" not in body:
body["max_tokens"] = 4000
print("Modified body_dict:", body)
log.debug("Modified body_dict:", body)
# Fix for ChatGPT calls failing because the num_ctx key is in body
if "num_ctx" in body:
......@@ -276,7 +297,7 @@ async def proxy(path: str, request: Request, user=Depends(get_verified_user)):
# Convert the modified body back to JSON
body = json.dumps(body)
except json.JSONDecodeError as e:
print("Error loading request body into a dictionary:", e)
log.error("Error loading request body into a dictionary:", e)
url = app.state.OPENAI_API_BASE_URLS[idx]
key = app.state.OPENAI_API_KEYS[idx]
......@@ -290,6 +311,8 @@ async def proxy(path: str, request: Request, user=Depends(get_verified_user)):
headers["Authorization"] = f"Bearer {key}"
headers["Content-Type"] = "application/json"
r = None
try:
r = requests.request(
method=request.method,
......@@ -312,7 +335,7 @@ async def proxy(path: str, request: Request, user=Depends(get_verified_user)):
response_data = r.json()
return response_data
except Exception as e:
print(e)
log.exception(e)
error_detail = "Open WebUI: Server Connection Error"
if r is not None:
try:
......@@ -322,4 +345,6 @@ async def proxy(path: str, request: Request, user=Depends(get_verified_user)):
except:
error_detail = f"External: {e}"
raise HTTPException(status_code=r.status_code, detail=error_detail)
raise HTTPException(
status_code=r.status_code if r else 500, detail=error_detail
)
......@@ -8,7 +8,7 @@ from fastapi import (
Form,
)
from fastapi.middleware.cors import CORSMiddleware
import os, shutil
import os, shutil, logging
from pathlib import Path
from typing import List
......@@ -21,6 +21,7 @@ from langchain_community.document_loaders import (
TextLoader,
PyPDFLoader,
CSVLoader,
BSHTMLLoader,
Docx2txtLoader,
UnstructuredEPubLoader,
UnstructuredWordDocumentLoader,
......@@ -54,6 +55,7 @@ from utils.misc import (
)
from utils.utils import get_current_user, get_admin_user
from config import (
SRC_LOG_LEVELS,
UPLOAD_DIR,
DOCS_DIR,
RAG_EMBEDDING_MODEL,
......@@ -66,6 +68,9 @@ from config import (
from constants import ERROR_MESSAGES
log = logging.getLogger(__name__)
log.setLevel(SRC_LOG_LEVELS["RAG"])
#
# if RAG_EMBEDDING_MODEL:
# sentence_transformer_ef = SentenceTransformer(
......@@ -111,39 +116,6 @@ class StoreWebForm(CollectionNameForm):
url: str
def store_data_in_vector_db(data, collection_name, overwrite: bool = False) -> bool:
text_splitter = RecursiveCharacterTextSplitter(
chunk_size=app.state.CHUNK_SIZE, chunk_overlap=app.state.CHUNK_OVERLAP
)
docs = text_splitter.split_documents(data)
texts = [doc.page_content for doc in docs]
metadatas = [doc.metadata for doc in docs]
try:
if overwrite:
for collection in CHROMA_CLIENT.list_collections():
if collection_name == collection.name:
print(f"deleting existing collection {collection_name}")
CHROMA_CLIENT.delete_collection(name=collection_name)
collection = CHROMA_CLIENT.create_collection(
name=collection_name,
embedding_function=app.state.sentence_transformer_ef,
)
collection.add(
documents=texts, metadatas=metadatas, ids=[str(uuid.uuid1()) for _ in texts]
)
return True
except Exception as e:
print(e)
if e.__class__.__name__ == "UniqueConstraintError":
return True
return False
@app.get("/")
async def get_status():
return {
......@@ -273,7 +245,7 @@ def query_doc_handler(
embedding_function=app.state.sentence_transformer_ef,
)
except Exception as e:
print(e)
log.exception(e)
raise HTTPException(
status_code=status.HTTP_400_BAD_REQUEST,
detail=ERROR_MESSAGES.DEFAULT(e),
......@@ -317,13 +289,69 @@ def store_web(form_data: StoreWebForm, user=Depends(get_current_user)):
"filename": form_data.url,
}
except Exception as e:
print(e)
log.exception(e)
raise HTTPException(
status_code=status.HTTP_400_BAD_REQUEST,
detail=ERROR_MESSAGES.DEFAULT(e),
)
def store_data_in_vector_db(data, collection_name, overwrite: bool = False) -> bool:
text_splitter = RecursiveCharacterTextSplitter(
chunk_size=app.state.CHUNK_SIZE,
chunk_overlap=app.state.CHUNK_OVERLAP,
add_start_index=True,
)
docs = text_splitter.split_documents(data)
if len(docs) > 0:
return store_docs_in_vector_db(docs, collection_name, overwrite), None
else:
raise ValueError(ERROR_MESSAGES.EMPTY_CONTENT)
def store_text_in_vector_db(
text, metadata, collection_name, overwrite: bool = False
) -> bool:
text_splitter = RecursiveCharacterTextSplitter(
chunk_size=app.state.CHUNK_SIZE,
chunk_overlap=app.state.CHUNK_OVERLAP,
add_start_index=True,
)
docs = text_splitter.create_documents([text], metadatas=[metadata])
return store_docs_in_vector_db(docs, collection_name, overwrite)
def store_docs_in_vector_db(docs, collection_name, overwrite: bool = False) -> bool:
texts = [doc.page_content for doc in docs]
metadatas = [doc.metadata for doc in docs]
try:
if overwrite:
for collection in CHROMA_CLIENT.list_collections():
if collection_name == collection.name:
log.info(f"deleting existing collection {collection_name}")
CHROMA_CLIENT.delete_collection(name=collection_name)
collection = CHROMA_CLIENT.create_collection(
name=collection_name,
embedding_function=app.state.sentence_transformer_ef,
)
collection.add(
documents=texts, metadatas=metadatas, ids=[str(uuid.uuid1()) for _ in texts]
)
return True
except Exception as e:
log.exception(e)
if e.__class__.__name__ == "UniqueConstraintError":
return True
return False
def get_loader(filename: str, file_content_type: str, file_path: str):
file_ext = filename.split(".")[-1].lower()
known_type = True
......@@ -381,6 +409,8 @@ def get_loader(filename: str, file_content_type: str, file_path: str):
loader = UnstructuredRSTLoader(file_path, mode="elements")
elif file_ext == "xml":
loader = UnstructuredXMLLoader(file_path)
elif file_ext in ["htm", "html"]:
loader = BSHTMLLoader(file_path, open_encoding="unicode_escape")
elif file_ext == "md":
loader = UnstructuredMarkdownLoader(file_path)
elif file_content_type == "application/epub+zip":
......@@ -399,9 +429,9 @@ def get_loader(filename: str, file_content_type: str, file_path: str):
elif file_ext in known_source_ext or (
file_content_type and file_content_type.find("text/") >= 0
):
loader = TextLoader(file_path)
loader = TextLoader(file_path, autodetect_encoding=True)
else:
loader = TextLoader(file_path)
loader = TextLoader(file_path, autodetect_encoding=True)
known_type = False
return loader, known_type
......@@ -415,7 +445,7 @@ def store_doc(
):
# "https://www.gutenberg.org/files/1727/1727-h/1727-h.htm"
print(file.content_type)
log.info(f"file.content_type: {file.content_type}")
try:
filename = file.filename
file_path = f"{UPLOAD_DIR}/{filename}"
......@@ -431,22 +461,24 @@ def store_doc(
loader, known_type = get_loader(file.filename, file.content_type, file_path)
data = loader.load()
result = store_data_in_vector_db(data, collection_name)
if result:
return {
"status": True,
"collection_name": collection_name,
"filename": filename,
"known_type": known_type,
}
else:
try:
result = store_data_in_vector_db(data, collection_name)
if result:
return {
"status": True,
"collection_name": collection_name,
"filename": filename,
"known_type": known_type,
}
except Exception as e:
raise HTTPException(
status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
detail=ERROR_MESSAGES.DEFAULT(),
detail=e,
)
except Exception as e:
print(e)
log.exception(e)
if "No pandoc was found" in str(e):
raise HTTPException(
status_code=status.HTTP_400_BAD_REQUEST,
......@@ -459,6 +491,37 @@ def store_doc(
)
class TextRAGForm(BaseModel):
name: str
content: str
collection_name: Optional[str] = None
@app.post("/text")
def store_text(
form_data: TextRAGForm,
user=Depends(get_current_user),
):
collection_name = form_data.collection_name
if collection_name == None:
collection_name = calculate_sha256_string(form_data.content)
result = store_text_in_vector_db(
form_data.content,
metadata={"name": form_data.name, "created_by": user.id},
collection_name=collection_name,
)
if result:
return {"status": True, "collection_name": collection_name}
else:
raise HTTPException(
status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
detail=ERROR_MESSAGES.DEFAULT(),
)
@app.get("/scan")
def scan_docs_dir(user=Depends(get_admin_user)):
for path in Path(DOCS_DIR).rglob("./**/*"):
......@@ -477,41 +540,45 @@ def scan_docs_dir(user=Depends(get_admin_user)):
)
data = loader.load()
result = store_data_in_vector_db(data, collection_name)
if result:
sanitized_filename = sanitize_filename(filename)
doc = Documents.get_doc_by_name(sanitized_filename)
if doc == None:
doc = Documents.insert_new_doc(
user.id,
DocumentForm(
**{
"name": sanitized_filename,
"title": filename,
"collection_name": collection_name,
"filename": filename,
"content": (
json.dumps(
{
"tags": list(
map(
lambda name: {"name": name},
tags,
try:
result = store_data_in_vector_db(data, collection_name)
if result:
sanitized_filename = sanitize_filename(filename)
doc = Documents.get_doc_by_name(sanitized_filename)
if doc == None:
doc = Documents.insert_new_doc(
user.id,
DocumentForm(
**{
"name": sanitized_filename,
"title": filename,
"collection_name": collection_name,
"filename": filename,
"content": (
json.dumps(
{
"tags": list(
map(
lambda name: {"name": name},
tags,
)
)
)
}
)
if len(tags)
else "{}"
),
}
),
)
}
)
if len(tags)
else "{}"
),
}
),
)
except Exception as e:
log.exception(e)
pass
except Exception as e:
print(e)
log.exception(e)
return True
......@@ -532,11 +599,11 @@ def reset(user=Depends(get_admin_user)) -> bool:
elif os.path.isdir(file_path):
shutil.rmtree(file_path)
except Exception as e:
print("Failed to delete %s. Reason: %s" % (file_path, e))
log.error("Failed to delete %s. Reason: %s" % (file_path, e))
try:
CHROMA_CLIENT.reset()
except Exception as e:
print(e)
log.exception(e)
return True
import re
import logging
from typing import List
from config import CHROMA_CLIENT
from config import SRC_LOG_LEVELS, CHROMA_CLIENT
log = logging.getLogger(__name__)
log.setLevel(SRC_LOG_LEVELS["RAG"])
def query_doc(collection_name: str, query: str, k: int, embedding_function):
......@@ -91,14 +95,13 @@ def query_collection(
def rag_template(template: str, context: str, query: str):
template = re.sub(r"\[context\]", context, template)
template = re.sub(r"\[query\]", query, template)
template = template.replace("[context]", context)
template = template.replace("[query]", query)
return template
def rag_messages(docs, messages, template, k, embedding_function):
print(docs)
log.debug(f"docs: {docs}")
last_user_message_idx = None
for i in range(len(messages) - 1, -1, -1):
......@@ -138,6 +141,8 @@ def rag_messages(docs, messages, template, k, embedding_function):
k=k,
embedding_function=embedding_function,
)
elif doc["type"] == "text":
context = doc["content"]
else:
context = query_doc(
collection_name=doc["collection_name"],
......@@ -146,11 +151,13 @@ def rag_messages(docs, messages, template, k, embedding_function):
embedding_function=embedding_function,
)
except Exception as e:
print(e)
log.exception(e)
context = None
relevant_contexts.append(context)
log.debug(f"relevant_contexts: {relevant_contexts}")
context_string = ""
for context in relevant_contexts:
if context:
......
from peewee import *
from config import DATA_DIR
from config import SRC_LOG_LEVELS, DATA_DIR
import os
import logging
log = logging.getLogger(__name__)
log.setLevel(SRC_LOG_LEVELS["DB"])
# Check if the file exists
if os.path.exists(f"{DATA_DIR}/ollama.db"):
# Rename the file
os.rename(f"{DATA_DIR}/ollama.db", f"{DATA_DIR}/webui.db")
print("File renamed successfully.")
log.info("File renamed successfully.")
else:
pass
......
......@@ -19,6 +19,7 @@ from config import (
DEFAULT_USER_ROLE,
ENABLE_SIGNUP,
USER_PERMISSIONS,
WEBHOOK_URL,
)
app = FastAPI()
......@@ -32,6 +33,7 @@ app.state.DEFAULT_MODELS = DEFAULT_MODELS
app.state.DEFAULT_PROMPT_SUGGESTIONS = DEFAULT_PROMPT_SUGGESTIONS
app.state.DEFAULT_USER_ROLE = DEFAULT_USER_ROLE
app.state.USER_PERMISSIONS = USER_PERMISSIONS
app.state.WEBHOOK_URL = WEBHOOK_URL
app.add_middleware(
......
......@@ -2,6 +2,7 @@ from pydantic import BaseModel
from typing import List, Union, Optional
import time
import uuid
import logging
from peewee import *
from apps.web.models.users import UserModel, Users
......@@ -9,6 +10,11 @@ from utils.utils import verify_password
from apps.web.internal.db import DB
from config import SRC_LOG_LEVELS
log = logging.getLogger(__name__)
log.setLevel(SRC_LOG_LEVELS["MODELS"])
####################
# DB MODEL
####################
......@@ -86,7 +92,7 @@ class AuthsTable:
def insert_new_auth(
self, email: str, password: str, name: str, role: str = "pending"
) -> Optional[UserModel]:
print("insert_new_auth")
log.info("insert_new_auth")
id = str(uuid.uuid4())
......@@ -103,7 +109,7 @@ class AuthsTable:
return None
def authenticate_user(self, email: str, password: str) -> Optional[UserModel]:
print("authenticate_user", email)
log.info(f"authenticate_user: {email}")
try:
auth = Auth.get(Auth.email == email, Auth.active == True)
if auth:
......
......@@ -95,20 +95,6 @@ class ChatTable:
except:
return None
def update_chat_by_id(self, id: str, chat: dict) -> Optional[ChatModel]:
try:
query = Chat.update(
chat=json.dumps(chat),
title=chat["title"] if "title" in chat else "New Chat",
timestamp=int(time.time()),
).where(Chat.id == id)
query.execute()
chat = Chat.get(Chat.id == id)
return ChatModel(**model_to_dict(chat))
except:
return None
def get_chat_lists_by_user_id(
self, user_id: str, skip: int = 0, limit: int = 50
) -> List[ChatModel]:
......
......@@ -3,6 +3,7 @@ from peewee import *
from playhouse.shortcuts import model_to_dict
from typing import List, Union, Optional
import time
import logging
from utils.utils import decode_token
from utils.misc import get_gravatar_url
......@@ -11,6 +12,11 @@ from apps.web.internal.db import DB
import json
from config import SRC_LOG_LEVELS
log = logging.getLogger(__name__)
log.setLevel(SRC_LOG_LEVELS["MODELS"])
####################
# Documents DB Schema
####################
......@@ -118,7 +124,7 @@ class DocumentsTable:
doc = Document.get(Document.name == form_data.name)
return DocumentModel(**model_to_dict(doc))
except Exception as e:
print(e)
log.exception(e)
return None
def update_doc_content_by_name(
......@@ -138,7 +144,7 @@ class DocumentsTable:
doc = Document.get(Document.name == name)
return DocumentModel(**model_to_dict(doc))
except Exception as e:
print(e)
log.exception(e)
return None
def delete_doc_by_name(self, name: str) -> bool:
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment