Merge branch 'main' into dev

README.md

@@ -13,7 +13,7 @@
 
 ChatGPT-Style Web Interface for Ollama 🦙
 
-**Disclaimer:** *ollama-webui is a community-driven project and is not affiliated with the Ollama team in any way. This initiative is independent, and any inquiries or feedback should be directed to [our community on Discord](https://discord.gg/5rJgQTnV4s). We kindly request users to refrain from contacting or harassing the Ollama team regarding this project.*
+**Disclaimer:** _ollama-webui is a community-driven project and is not affiliated with the Ollama team in any way. This initiative is independent, and any inquiries or feedback should be directed to [our community on Discord](https://discord.gg/5rJgQTnV4s). We kindly request users to refrain from contacting or harassing the Ollama team regarding this project._
 
 ![Ollama Web UI Demo](./demo.gif)
 
@@ -35,6 +35,8 @@ Also check our sibling project, [OllamaHub](https://ollamahub.com/), where you c
 
 - 📥🗑️ **Download/Delete Models**: Easily download or remove models directly from the web UI.
 
+- ⬆️ **GGUF File Model Creation**: Effortlessly create Ollama models by uploading GGUF files directly from the web UI. Streamlined process with options to upload from your machine or download GGUF files from Hugging Face.
+
 - 🤖 **Multiple Model Support**: Seamlessly switch between different chat models for diverse interactions.
 
 - 🔄 **Multi-Modal Support**: Seamlessly engage with models that support multimodal interactions, including images (e.g., LLava).

backend/.gitignore

 __pycache__
 .env
+_old
+uploads
\ No newline at end of file

backend/apps/web/main.py

 from fastapi import FastAPI, Request, Depends, HTTPException
 from fastapi.middleware.cors import CORSMiddleware
 
-from apps.web.routers import auths, users
+from apps.web.routers import auths, users, utils
 from config import WEBUI_VERSION, WEBUI_AUTH
 
 app = FastAPI()
@@ -19,6 +19,7 @@ app.add_middleware(
 
 app.include_router(auths.router, prefix="/auths", tags=["auths"])
 app.include_router(users.router, prefix="/users", tags=["users"])
+app.include_router(utils.router, prefix="/utils", tags=["utils"])
 
 
 @app.get("/")

backend/apps/web/routers/utils.py

+from fastapi import APIRouter, UploadFile, File, BackgroundTasks
+from fastapi import Depends, HTTPException, status
+from starlette.responses import StreamingResponse
+
+from pydantic import BaseModel
+
+import requests
+import os
+import aiohttp
+import json
+
+
+from utils.misc import calculate_sha256
+
+from config import OLLAMA_API_BASE_URL
+
+
+router = APIRouter()
+
+
+class UploadBlobForm(BaseModel):
+    filename: str
+
+
+from urllib.parse import urlparse
+
+
+def parse_huggingface_url(hf_url):
+    try:
+        # Parse the URL
+        parsed_url = urlparse(hf_url)
+
+        # Get the path and split it into components
+        path_components = parsed_url.path.split("/")
+
+        # Extract the desired output
+        user_repo = "/".join(path_components[1:3])
+        model_file = path_components[-1]
+
+        return model_file
+    except ValueError:
+        return None
+
+
+async def download_file_stream(url, file_path, file_name, chunk_size=1024 * 1024):
+    done = False
+
+    if os.path.exists(file_path):
+        current_size = os.path.getsize(file_path)
+    else:
+        current_size = 0
+
+    headers = {"Range": f"bytes={current_size}-"} if current_size > 0 else {}
+
+    timeout = aiohttp.ClientTimeout(total=600)  # Set the timeout
+
+    async with aiohttp.ClientSession(timeout=timeout) as session:
+        async with session.get(url, headers=headers) as response:
+            total_size = int(response.headers.get("content-length", 0)) + current_size
+
+            with open(file_path, "ab+") as file:
+                async for data in response.content.iter_chunked(chunk_size):
+                    current_size += len(data)
+                    file.write(data)
+
+                    done = current_size == total_size
+                    progress = round((current_size / total_size) * 100, 2)
+                    yield f'data: {{"progress": {progress}, "completed": {current_size}, "total": {total_size}}}\n\n'
+
+                if done:
+                    file.seek(0)
+                    hashed = calculate_sha256(file)
+                    file.seek(0)
+
+                    url = f"{OLLAMA_API_BASE_URL}/blobs/sha256:{hashed}"
+                    response = requests.post(url, data=file)
+
+                    if response.ok:
+                        res = {
+                            "done": done,
+                            "blob": f"sha256:{hashed}",
+                            "name": file_name,
+                        }
+                        os.remove(file_path)
+
+                        yield f"data: {json.dumps(res)}\n\n"
+                    else:
+                        raise "Ollama: Could not create blob, Please try again."
+
+
+@router.get("/download")
+async def download(
+    url: str,
+):
+    # url = "https://huggingface.co/TheBloke/stablelm-zephyr-3b-GGUF/resolve/main/stablelm-zephyr-3b.Q2_K.gguf"
+    file_name = parse_huggingface_url(url)
+
+    if file_name:
+        os.makedirs("./uploads", exist_ok=True)
+        file_path = os.path.join("./uploads", f"{file_name}")
+
+        return StreamingResponse(
+            download_file_stream(url, file_path, file_name),
+            media_type="text/event-stream",
+        )
+    else:
+        return None
+
+
+@router.post("/upload")
+async def upload(file: UploadFile = File(...)):
+    os.makedirs("./uploads", exist_ok=True)
+    file_path = os.path.join("./uploads", file.filename)
+
+    async def file_write_stream():
+        total = 0
+        total_size = file.size
+        chunk_size = 1024 * 1024
+
+        done = False
+        try:
+            with open(file_path, "wb+") as f:
+                while True:
+                    chunk = file.file.read(chunk_size)
+                    if not chunk:
+                        break
+                    f.write(chunk)
+                    total += len(chunk)
+                    done = total_size == total
+                    progress = round((total / total_size) * 100, 2)
+
+                    res = {
+                        "progress": progress,
+                        "total": total_size,
+                        "completed": total,
+                    }
+
+                    yield f"data: {json.dumps(res)}\n\n"
+
+                if done:
+                    f.seek(0)
+                    hashed = calculate_sha256(f)
+                    f.seek(0)
+
+                    url = f"{OLLAMA_API_BASE_URL}/blobs/sha256:{hashed}"
+                    response = requests.post(url, data=f)
+
+                    if response.ok:
+                        res = {
+                            "done": done,
+                            "blob": f"sha256:{hashed}",
+                            "name": file.filename,
+                        }
+                        os.remove(file_path)
+
+                        yield f"data: {json.dumps(res)}\n\n"
+                    else:
+                        raise "Ollama: Could not create blob, Please try again."
+
+        except Exception as e:
+            res = {"error": str(e)}
+            yield f"data: {json.dumps(res)}\n\n"
+
+    return StreamingResponse(file_write_stream(), media_type="text/event-stream")

backend/config.py

@@ -30,7 +30,7 @@ if ENV == "prod":
 # WEBUI_VERSION
 ####################################
 
-WEBUI_VERSION = os.environ.get("WEBUI_VERSION", "v1.0.0-alpha.34")
+WEBUI_VERSION = os.environ.get("WEBUI_VERSION", "v1.0.0-alpha.40")
 
 ####################################
 # WEBUI_AUTH

backend/requirements.txt

@@ -12,6 +12,7 @@ passlib[bcrypt]
 uuid
 
 requests
+aiohttp
 pymongo
 bcrypt
 

backend/utils/misc.py

@@ -13,3 +13,11 @@ def get_gravatar_url(email):
 
     # Grab the actual image URL
     return f"https://www.gravatar.com/avatar/{hash_hex}?d=mp"
+
+
+def calculate_sha256(file):
+    sha256 = hashlib.sha256()
+    # Read the file in chunks to efficiently handle large files
+    for chunk in iter(lambda: file.read(8192), b""):
+        sha256.update(chunk)
+    return sha256.hexdigest()

src/lib/components/chat/MessageInput.svelte

@@ -155,7 +155,7 @@
 <div class="fixed bottom-0 w-full">
 	<div class="px-2.5 pt-2.5 -mb-0.5 mx-auto inset-x-0 bg-transparent flex justify-center">
 		{#if messages.length == 0 && suggestionPrompts.length !== 0}
-			<div class="max-w-3xl">
+			<div class="max-w-3xl w-full">
 				<Suggestions {suggestionPrompts} {submitPrompt} />
 			</div>
 		{/if}

src/lib/components/chat/MessageInput/Suggestions.svelte

@@ -3,7 +3,7 @@
 	export let suggestionPrompts = [];
 </script>
 
-<div class=" flex flex-wrap-reverse mb-3 md:p-1 text-left">
+<div class=" flex flex-wrap-reverse mb-3 md:p-1 text-left w-full">
 	{#each suggestionPrompts as prompt, promptIdx}
 		<div class="{promptIdx > 1 ? 'hidden sm:inline-flex' : ''} basis-full sm:basis-1/2 p-[5px]">
 			<button

src/lib/constants.ts

@@ -8,7 +8,8 @@ export const OLLAMA_API_BASE_URL =
 			: `http://localhost:11434/api`
 		: PUBLIC_API_BASE_URL;
 
-export const WEBUI_API_BASE_URL = dev ? `http://${location.hostname}:8080/api/v1` : `/api/v1`;
+export const WEBUI_BASE_URL = dev ? `http://${location.hostname}:8080` : ``;
+export const WEBUI_API_BASE_URL = `${WEBUI_BASE_URL}/api/v1`;
 
 export const WEB_UI_VERSION = 'v1.0.0-alpha-static';
 

src/routes/(app)/+layout.svelte

@@ -55,7 +55,9 @@
 		// If OpenAI API Key exists
 		if ($settings.OPENAI_API_KEY) {
 			// Validate OPENAI_API_KEY
-			const openaiModelRes = await fetch(`https://api.openai.com/v1/models`, {
+
+			const API_BASE_URL = $settings.OPENAI_API_BASE_URL ?? 'https://api.openai.com/v1';
+			const openaiModelRes = await fetch(`${API_BASE_URL}/models`, {
 				method: 'GET',
 				headers: {
 					'Content-Type': 'application/json',
@@ -72,15 +74,19 @@
 					return null;
 				});
 
-			const openAIModels = openaiModelRes?.data ?? null;
+			const openAIModels = Array.isArray(openaiModelRes)
+				? openaiModelRes
+				: openaiModelRes?.data ?? null;
 
 			models.push(
 				...(openAIModels
 					? [
 							{ name: 'hr' },
 							...openAIModels
-								.map((model) => ({ name: model.id, label: 'OpenAI' }))
-								.filter((model) => model.name.includes('gpt'))
+								.map((model) => ({ name: model.id, external: true }))
+								.filter((model) =>
+									API_BASE_URL.includes('openai') ? model.name.includes('gpt') : true
+								)
 					  ]
 					: [])
 			);

src/routes/(app)/+page.svelte

@@ -7,7 +7,7 @@
 	import { splitStream } from '$lib/utils';
 	import { goto } from '$app/navigation';
 
-	import { config, modelfiles, user, settings, db, chats, chatId } from '$lib/stores';
+	import { config, models, modelfiles, user, settings, db, chats, chatId } from '$lib/stores';
 
 	import MessageInput from '$lib/components/chat/MessageInput.svelte';
 	import Messages from '$lib/components/chat/Messages.svelte';
@@ -130,7 +130,8 @@
 	const sendPrompt = async (userPrompt, parentId, _chatId) => {
 		await Promise.all(
 			selectedModels.map(async (model) => {
-				if (model.includes('gpt-')) {
+				console.log(model);
+				if ($models.filter((m) => m.name === model)[0].external) {
 					await sendPromptOpenAI(model, userPrompt, parentId, _chatId);
 				} else {
 					await sendPromptOllama(model, userPrompt, parentId, _chatId);
@@ -244,6 +245,13 @@
 								}
 							} else {
 								responseMessage.done = true;
+
+								if (responseMessage.content == '') {
+									responseMessage.error = true;
+									responseMessage.content =
+										'Oops! No text generated from Ollama, Please try again.';
+								}
+
 								responseMessage.context = data.context ?? null;
 								responseMessage.info = {
 									total_duration: data.total_duration,
@@ -364,15 +372,15 @@
 					];
 				}
 
-				await tick();
-
 				window.scrollTo({ top: document.body.scrollHeight });
 
-				const res = await fetch(`https://api.openai.com/v1/chat/completions`, {
+				const res = await fetch(
+					`${$settings.OPENAI_API_BASE_URL ?? 'https://api.openai.com/v1'}/chat/completions`,
+					{
 						method: 'POST',
 						headers: {
-						'Content-Type': 'application/json',
-						Authorization: `Bearer ${$settings.OPENAI_API_KEY}`
+							Authorization: `Bearer ${$settings.OPENAI_API_KEY}`,
+							'Content-Type': 'application/json'
 						},
 						body: JSON.stringify({
 							model: model,
@@ -413,8 +421,13 @@
 							num_ctx: $settings.num_ctx ?? undefined,
 							frequency_penalty: $settings.repeat_penalty ?? undefined
 						})
+					}
+				).catch((err) => {
+					console.log(err);
+					return null;
 				});
 
+				if (res && res.ok) {
 					const reader = res.body
 						.pipeThrough(new TextDecoderStream())
 						.pipeThrough(splitStream('\n'))
@@ -454,6 +467,17 @@
 							console.log(error);
 						}
 
+						if ($settings.notificationEnabled && !document.hasFocus()) {
+							const notification = new Notification(`OpenAI ${model}`, {
+								body: responseMessage.content,
+								icon: '/favicon.png'
+							});
+						}
+
+						if ($settings.responseAutoCopy) {
+							copyToClipboard(responseMessage.content);
+						}
+
 						if (autoScroll) {
 							window.scrollTo({ top: document.body.scrollHeight });
 						}
@@ -475,22 +499,36 @@
 							history: history
 						});
 					}
-
-				stopResponseFlag = false;
-
-				await tick();
-
-				if ($settings.notificationEnabled && !document.hasFocus()) {
-					const notification = new Notification(`OpenAI ${model}`, {
-						body: responseMessage.content,
-						icon: '/favicon.png'
-					});
+				} else {
+					if (res !== null) {
+						const error = await res.json();
+						console.log(error);
+						if ('detail' in error) {
+							toast.error(error.detail);
+							responseMessage.content = error.detail;
+						} else {
+							if ('message' in error.error) {
+								toast.error(error.error.message);
+								responseMessage.content = error.error.message;
+							} else {
+								toast.error(error.error);
+								responseMessage.content = error.error;
+							}
+						}
+					} else {
+						toast.error(`Uh-oh! There was an issue connecting to ${model}.`);
+						responseMessage.content = `Uh-oh! There was an issue connecting to ${model}.`;
 					}
 
-				if ($settings.responseAutoCopy) {
-					copyToClipboard(responseMessage.content);
+					responseMessage.error = true;
+					responseMessage.content = `Uh-oh! There was an issue connecting to ${model}.`;
+					responseMessage.done = true;
+					messages = messages;
 				}
 
+				stopResponseFlag = false;
+				await tick();
+
 				if (autoScroll) {
 					window.scrollTo({ top: document.body.scrollHeight });
 				}

src/routes/(app)/c/[id]/+page.svelte

@@ -6,7 +6,7 @@
 	import { onMount, tick } from 'svelte';
 	import { convertMessagesToHistory, splitStream } from '$lib/utils';
 	import { goto } from '$app/navigation';
-	import { config, modelfiles, user, settings, db, chats, chatId } from '$lib/stores';
+	import { config, models, modelfiles, user, settings, db, chats, chatId } from '$lib/stores';
 
 	import MessageInput from '$lib/components/chat/MessageInput.svelte';
 	import Messages from '$lib/components/chat/Messages.svelte';
@@ -144,7 +144,8 @@
 	const sendPrompt = async (userPrompt, parentId, _chatId) => {
 		await Promise.all(
 			selectedModels.map(async (model) => {
-				if (model.includes('gpt-')) {
+				console.log(model);
+				if ($models.filter((m) => m.name === model)[0].external) {
 					await sendPromptOpenAI(model, userPrompt, parentId, _chatId);
 				} else {
 					await sendPromptOllama(model, userPrompt, parentId, _chatId);
@@ -258,6 +259,13 @@
 								}
 							} else {
 								responseMessage.done = true;
+
+								if (responseMessage.content == '') {
+									responseMessage.error = true;
+									responseMessage.content =
+										'Oops! No text generated from Ollama, Please try again.';
+								}
+
 								responseMessage.context = data.context ?? null;
 								responseMessage.info = {
 									total_duration: data.total_duration,
@@ -378,15 +386,15 @@
 					];
 				}
 
-				await tick();
-
 				window.scrollTo({ top: document.body.scrollHeight });
 
-				const res = await fetch(`https://api.openai.com/v1/chat/completions`, {
+				const res = await fetch(
+					`${$settings.OPENAI_API_BASE_URL ?? 'https://api.openai.com/v1'}/chat/completions`,
+					{
 						method: 'POST',
 						headers: {
-						'Content-Type': 'application/json',
-						Authorization: `Bearer ${$settings.OPENAI_API_KEY}`
+							Authorization: `Bearer ${$settings.OPENAI_API_KEY}`,
+							'Content-Type': 'application/json'
 						},
 						body: JSON.stringify({
 							model: model,
@@ -427,8 +435,13 @@
 							num_ctx: $settings.num_ctx ?? undefined,
 							frequency_penalty: $settings.repeat_penalty ?? undefined
 						})
+					}
+				).catch((err) => {
+					console.log(err);
+					return null;
 				});
 
+				if (res && res.ok) {
 					const reader = res.body
 						.pipeThrough(new TextDecoderStream())
 						.pipeThrough(splitStream('\n'))
@@ -468,6 +481,17 @@
 							console.log(error);
 						}
 
+						if ($settings.notificationEnabled && !document.hasFocus()) {
+							const notification = new Notification(`OpenAI ${model}`, {
+								body: responseMessage.content,
+								icon: '/favicon.png'
+							});
+						}
+
+						if ($settings.responseAutoCopy) {
+							copyToClipboard(responseMessage.content);
+						}
+
 						if (autoScroll) {
 							window.scrollTo({ top: document.body.scrollHeight });
 						}
@@ -489,22 +513,36 @@
 							history: history
 						});
 					}
-
-				stopResponseFlag = false;
-
-				await tick();
-
-				if ($settings.notificationEnabled && !document.hasFocus()) {
-					const notification = new Notification(`OpenAI ${model}`, {
-						body: responseMessage.content,
-						icon: '/favicon.png'
-					});
+				} else {
+					if (res !== null) {
+						const error = await res.json();
+						console.log(error);
+						if ('detail' in error) {
+							toast.error(error.detail);
+							responseMessage.content = error.detail;
+						} else {
+							if ('message' in error.error) {
+								toast.error(error.error.message);
+								responseMessage.content = error.error.message;
+							} else {
+								toast.error(error.error);
+								responseMessage.content = error.error;
+							}
+						}
+					} else {
+						toast.error(`Uh-oh! There was an issue connecting to ${model}.`);
+						responseMessage.content = `Uh-oh! There was an issue connecting to ${model}.`;
 					}
 
-				if ($settings.responseAutoCopy) {
-					copyToClipboard(responseMessage.content);
+					responseMessage.error = true;
+					responseMessage.content = `Uh-oh! There was an issue connecting to ${model}.`;
+					responseMessage.done = true;
+					messages = messages;
 				}
 
+				stopResponseFlag = false;
+				await tick();
+
 				if (autoScroll) {
 					window.scrollTo({ top: document.body.scrollHeight });
 				}