Merge pull request #1555 from open-webui/dev

0.1.119

CHANGELOG.md

@@ -5,6 +5,26 @@ 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).
 
+## [0.1.119] - 2024-04-16
+
+### Added
+
+- **🌟 Enhanced RAG Embedding Support**: Ollama, and OpenAI models can now be used for RAG embedding model.
+- **🔄 Seamless Integration**: Copy 'ollama run <model name>' directly from Ollama page to easily select and pull models.
+- **🏷️ Tagging Feature**: Add tags to chats directly via the sidebar chat menu.
+- **📱 Mobile Accessibility**: Swipe left and right on mobile to effortlessly open and close the sidebar.
+- **🔍 Improved Navigation**: Admin panel now supports pagination for user list.
+- **🌍 Additional Language Support**: Added Polish language support.
+
+### Fixed
+
+- **🌍 Language Enhancements**: Vietnamese and Spanish translations have been improved.
+- **🔧 Helm Fixes**: Resolved issues with Helm trailing slash and manifest.json.
+
+### Changed
+
+- **🐳 Docker Optimization**: Updated docker image build process to utilize 'uv' for significantly faster builds compared to 'pip3'.
+
 ## [0.1.118] - 2024-04-10
 
 ### Added

Dockerfile

@@ -93,15 +93,16 @@ RUN if [ "$USE_OLLAMA" = "true" ]; then \
 # install python dependencies
 COPY ./backend/requirements.txt ./requirements.txt
 
-RUN if [ "$USE_CUDA" = "true" ]; then \
+RUN pip3 install uv && \
+    if [ "$USE_CUDA" = "true" ]; then \
         # If you use CUDA the whisper and embedding model will be downloaded on first use
         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 && \
+        uv pip install --system -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')"; \
     else \
         pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cpu --no-cache-dir && \
-        pip3 install -r requirements.txt --no-cache-dir && \
+        uv pip install --system -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')"; \
     fi

backend/apps/audio/main.py

@@ -28,6 +28,7 @@
     UPLOAD_DIR,
     WHISPER_MODEL,
     WHISPER_MODEL_DIR,
+    WHISPER_MODEL_AUTO_UPDATE,
     DEVICE_TYPE,
 )
 
@@ -69,12 +70,24 @@ def transcribe(
             f.write(contents)
             f.close()
 
-        model = WhisperModel(
-            WHISPER_MODEL,
-            device=whisper_device_type,
-            compute_type="int8",
-            download_root=WHISPER_MODEL_DIR,
-        )
+        whisper_kwargs = {
+            "model_size_or_path": WHISPER_MODEL,
+            "device": whisper_device_type,
+            "compute_type": "int8",
+            "download_root": WHISPER_MODEL_DIR,
+            "local_files_only": not WHISPER_MODEL_AUTO_UPDATE,
+        }
+
+        log.debug(f"whisper_kwargs: {whisper_kwargs}")
+
+        try:
+            model = WhisperModel(**whisper_kwargs)
+        except:
+            log.warning(
+                "WhisperModel initialization failed, attempting download with local_files_only=False"
+            )
+            whisper_kwargs["local_files_only"] = False
+            model = WhisperModel(**whisper_kwargs)
 
         segments, info = model.transcribe(file_path, beam_size=5)
         log.info(

backend/apps/images/main.py

@@ -29,7 +29,13 @@
 import json
 import logging
 
-from config import SRC_LOG_LEVELS, CACHE_DIR, AUTOMATIC1111_BASE_URL, COMFYUI_BASE_URL
+from config import (
+    SRC_LOG_LEVELS,
+    CACHE_DIR,
+    ENABLE_IMAGE_GENERATION,
+    AUTOMATIC1111_BASE_URL,
+    COMFYUI_BASE_URL,
+)
 
 
 log = logging.getLogger(__name__)
@@ -48,7 +54,7 @@
 )
 
 app.state.ENGINE = ""
-app.state.ENABLED = False
+app.state.ENABLED = ENABLE_IMAGE_GENERATION
 
 app.state.OPENAI_API_KEY = ""
 app.state.MODEL = ""

backend/apps/ollama/main.py

@@ -612,8 +612,13 @@ async def generate_embeddings(
     user=Depends(get_current_user),
 ):
     if url_idx == None:
-        if form_data.model in app.state.MODELS:
-            url_idx = random.choice(app.state.MODELS[form_data.model]["urls"])
+        model = form_data.model
+
+        if ":" not in model:
+            model = f"{model}:latest"
+
+        if model in app.state.MODELS:
+            url_idx = random.choice(app.state.MODELS[model]["urls"])
         else:
             raise HTTPException(
                 status_code=400,
@@ -649,6 +654,60 @@ async def generate_embeddings(
         )
 
 
+def generate_ollama_embeddings(
+    form_data: GenerateEmbeddingsForm,
+    url_idx: Optional[int] = None,
+):
+
+    log.info(f"generate_ollama_embeddings {form_data}")
+
+    if url_idx == None:
+        model = form_data.model
+
+        if ":" not in model:
+            model = f"{model}:latest"
+
+        if model in app.state.MODELS:
+            url_idx = random.choice(app.state.MODELS[model]["urls"])
+        else:
+            raise HTTPException(
+                status_code=400,
+                detail=ERROR_MESSAGES.MODEL_NOT_FOUND(form_data.model),
+            )
+
+    url = app.state.OLLAMA_BASE_URLS[url_idx]
+    log.info(f"url: {url}")
+
+    try:
+        r = requests.request(
+            method="POST",
+            url=f"{url}/api/embeddings",
+            data=form_data.model_dump_json(exclude_none=True).encode(),
+        )
+        r.raise_for_status()
+
+        data = r.json()
+
+        log.info(f"generate_ollama_embeddings {data}")
+
+        if "embedding" in data:
+            return data["embedding"]
+        else:
+            raise "Something went wrong :/"
+    except Exception as e:
+        log.exception(e)
+        error_detail = "Open WebUI: Server Connection Error"
+        if r is not None:
+            try:
+                res = r.json()
+                if "error" in res:
+                    error_detail = f"Ollama: {res['error']}"
+            except:
+                error_detail = f"Ollama: {e}"
+
+        raise error_detail
+
+
 class GenerateCompletionForm(BaseModel):
     model: str
     prompt: str
@@ -672,8 +731,13 @@ async def generate_completion(
 ):
 
     if url_idx == None:
-        if form_data.model in app.state.MODELS:
-            url_idx = random.choice(app.state.MODELS[form_data.model]["urls"])
+        model = form_data.model
+
+        if ":" not in model:
+            model = f"{model}:latest"
+
+        if model in app.state.MODELS:
+            url_idx = random.choice(app.state.MODELS[model]["urls"])
         else:
             raise HTTPException(
                 status_code=400,
@@ -770,8 +834,13 @@ async def generate_chat_completion(
 ):
 
     if url_idx == None:
-        if form_data.model in app.state.MODELS:
-            url_idx = random.choice(app.state.MODELS[form_data.model]["urls"])
+        model = form_data.model
+
+        if ":" not in model:
+            model = f"{model}:latest"
+
+        if model in app.state.MODELS:
+            url_idx = random.choice(app.state.MODELS[model]["urls"])
         else:
             raise HTTPException(
                 status_code=400,
@@ -874,8 +943,13 @@ async def generate_openai_chat_completion(
 ):
 
     if url_idx == None:
-        if form_data.model in app.state.MODELS:
-            url_idx = random.choice(app.state.MODELS[form_data.model]["urls"])
+        model = form_data.model
+
+        if ":" not in model:
+            model = f"{model}:latest"
+
+        if model in app.state.MODELS:
+            url_idx = random.choice(app.state.MODELS[model]["urls"])
         else:
             raise HTTPException(
                 status_code=400,