Merge pull request #62 from mborne/open-webui-proxy

ollama / open-webui - split and add proxy support

ollama/.env

@@ -0,0 +1 @@
+DEVBOX_HOSTNAME=dev.localhost

ollama/README.md

@@ -0,0 +1,55 @@
+# ollama
+
+Containers running [Ollama](https://hub.docker.com/r/ollama/ollama)
+
+## Usage with docker
+
+* Ensure that GPU support is enabled in docker (or adapt [docker-compose.yaml](docker-compose.yaml)) :
+
+```bash
+docker run --gpus all nvcr.io/nvidia/k8s/cuda-sample:nbody nbody -gpu -benchmark
+```
+
+* To use Ollama CLI :
+
+```bash
+# pull models from https://ollama.com/library
+docker compose exec ollama ollama pull llama3
+docker compose exec ollama ollama pull gemma2
+# interactive model
+docker compose exec ollama ollama run llama3
+```
+
+* To use [Ollama API](https://github.com/ollama/ollama/blob/main/docs/api.md#api) :
+
+```bash
+# list models
+curl -sS http://localhost:11434/api/tags | jq -r '.models[].name'
+
+# pull model from https://ollama.com/library
+curl http://localhost:11434/api/pull -d '{
+  "name": "llama3"
+}'
+
+# use model
+curl http://localhost:11434/api/generate -d '{
+  "model": "llama3",
+  "prompt": "Why is the sky blue?"
+}'
+```
+
+* To create **custom model** from [OLLAMA Modelfile](https://github.com/ollama/ollama/tree/main?tab=readme-ov-file#customize-a-prompt), a sample [models/geoassistant](models/geoassistant/README.md) is available :
+
+```bash
+docker compose exec ollama /bin/bash
+ollama create geoassistant -f /models/geoassistant/Modelfile
+ollama run geoassistant
+# Do you know the most visited museums in Paris?
+```
+
+## Ressources
+
+* [ollama](https://github.com/ollama/ollama/tree/main?tab=readme-ov-file#ollama)
+* [hub.docker.com - ollama/ollama](https://hub.docker.com/r/ollama/ollama)
+* [ollama - API](https://github.com/ollama/ollama/blob/main/docs/api.md#api)
+* [mborne/toolbox - cuda-toolkit](https://github.com/mborne/toolbox/tree/master/cuda-toolkit#ressources)

ollama/docker-compose.yaml

@@ -0,0 +1,38 @@
+services:
+  # https://hub.docker.com/r/ollama/ollama
+  ollama:
+    image: ollama/ollama:0.2.6
+    container_name: ollama
+    ports:
+    - 11434:11434
+    volumes:
+    - ollama:/root/.ollama
+    - ./models:/models:ro
+    environment:
+    #- OLLAMA_DEBUG=1
+    - HTTP_PROXY
+    - HTTPS_PROXY
+    - NO_PROXY=0.0.0.0,ollama,${NO_PROXY:-127.0.0.1,localhost}
+    deploy:
+     resources:
+       reservations:
+         devices:
+         - driver: nvidia
+           capabilities: ["gpu"]
+           count: all
+    labels:
+      - "traefik.enable=true"
+      # https://ollama.dev.localhost
+      - "traefik.http.routers.ollama.rule=Host(`ollama.${DEVBOX_HOSTNAME}`)"
+      - "traefik.http.routers.ollama.service=ollama-service@docker"
+      - "traefik.http.services.ollama-service.loadbalancer.server.port=11434"
+    restart: unless-stopped
+
+volumes:
+  ollama:
+    name: ollama
+
+networks:
+  default:
+    name: devbox
+    external: true

open-webui/.env

@@ -2,3 +2,4 @@ DEVBOX_HOSTNAME=dev.localhost
 WEBUI_AUTH=False
 OPENAI_API_KEY=
 PIPELINES_API_KEY=0p3n-w3bu!
+OLLAMA_BASE_URL=http://ollama:11434

open-webui/README.md

@@ -1,6 +1,10 @@
 # Open WebUI
 
-Containers running [Open WebUI](https://github.com/open-webui/open-webui?tab=readme-ov-file#open-webui-formerly-ollama-webui-) and [ollama](https://hub.docker.com/r/ollama/ollama) to get started with LLM locally.
+Container running [Open WebUI](https://github.com/open-webui/open-webui?tab=readme-ov-file#open-webui-formerly-ollama-webui-) for [Ollama](../ollama/README.md).
+
+## Requirements
+
+* [ollama](../ollama/README.md)
 
 ## Usage with docker
 
@@ -12,49 +16,14 @@ docker run --gpus all nvcr.io/nvidia/k8s/cuda-sample:nbody nbody -gpu -benchmark
 
 * Start : `docker compose up -d`
 * Open https://open-webui.dev.localhost
-* To use [ollama API](https://github.com/ollama/ollama/blob/main/docs/api.md#api) :
-
-```bash
-# list models
-curl -sS http://localhost:11434/api/tags | jq -r '.models[].name'
-
-# pull model from https://ollama.com/library
-curl http://localhost:11434/api/pull -d '{
-  "name": "llama3"
-}'
-
-# use model
-curl http://localhost:11434/api/generate -d '{
-  "model": "llama3",
-  "prompt": "Why is the sky blue?"
-}'
-```
-
-## Custom model
-
-To create custom model from [OLLAMA Modelfile](https://github.com/ollama/ollama/tree/main?tab=readme-ov-file#customize-a-prompt), a sample [models/geoassistant](models/geoassistant/README.md) is available :
-
-```bash
-docker compose exec ollama
-ollama create geoassistant -f /models/geoassistant/Modelfile
-ollama run geoassistant
-# Do you know the most visited museums in Paris?
-```
 
 
 ## Resources
 
 * [Open WebUI - Getting Started](https://docs.openwebui.com/getting-started/)
 * [mborne/toolbox - cuda-toolkit](https://github.com/mborne/toolbox/tree/master/cuda-toolkit#ressources)
+* [Pipelines](https://docs.openwebui.com/pipelines) :
+  * https://docs.openwebui.com/pipelines/#-quick-start-with-docker
+  * https://ikasten.io/2024/06/03/getting-started-with-openwebui-pipelines/
+  * https://raw.githubusercontent.com/open-webui/pipelines/main/examples/filters/function_calling_filter_pipeline.py
 
-[OLLAMA](https://github.com/ollama/ollama) :
-
-* [ollama](https://github.com/ollama/ollama/tree/main?tab=readme-ov-file#ollama)
-* [hub.docker.com - ollama/ollama](https://hub.docker.com/r/ollama/ollama)
-* [ollama - API](https://github.com/ollama/ollama/blob/main/docs/api.md#api)
-
-[Pipelines](https://docs.openwebui.com/pipelines) :
-
-* https://docs.openwebui.com/pipelines/#-quick-start-with-docker
-* https://ikasten.io/2024/06/03/getting-started-with-openwebui-pipelines/
-* https://raw.githubusercontent.com/open-webui/pipelines/main/examples/filters/function_calling_filter_pipeline.py

open-webui/docker-compose.yaml

@@ -8,20 +8,23 @@ services:
     - 3000:8080
     environment:
     - WEBUI_AUTH=False
-    - OLLAMA_BASE_URL=http://ollama:11434
+    - OLLAMA_BASE_URL=${OLLAMA_BASE_URL}
     - GLOBAL_LOG_LEVEL="DEBUG"
     # add pipelines
     - OPENAI_API_BASE_URL=https://api.openai.com/v1;http://pipelines:9099
     - OPENAI_API_KEYS=${OPENAI_API_KEY};${PIPELINES_API_KEY}
+    - HTTP_PROXY
+    - HTTPS_PROXY
+    - NO_PROXY=ollama,pipelines,$NO_PROXY
     volumes:
     - open-webui:/app/backend/data
     deploy:
-      resources:
-        reservations:
-          devices:
-          - driver: nvidia
-            capabilities: ["gpu"]
-            count: all
+     resources:
+       reservations:
+         devices:
+         - driver: nvidia
+           capabilities: ["gpu"]
+           count: all
     labels:
       - "traefik.enable=true"
       # https://open-webui.dev.localhost
@@ -30,30 +33,6 @@ services:
       - "traefik.http.services.open-webui-service.loadbalancer.server.port=8080"
     restart: unless-stopped
 
-  # https://hub.docker.com/r/ollama/ollama
-  ollama:
-    image: ollama/ollama:latest
-    container_name: ollama
-    ports:
-    - 11434:11434
-    volumes:
-    - ollama:/root/.ollama
-    - ./models:/models:ro
-    deploy:
-      resources:
-        reservations:
-          devices:
-          - driver: nvidia
-            capabilities: ["gpu"]
-            count: all
-    labels:
-      - "traefik.enable=true"
-      # https://ollama.dev.localhost
-      - "traefik.http.routers.ollama.rule=Host(`ollama.${DEVBOX_HOSTNAME}`)"
-      - "traefik.http.routers.ollama.service=ollama-service@docker"
-      - "traefik.http.services.ollama-service.loadbalancer.server.port=11434"
-    restart: unless-stopped
-
   # https://docs.openwebui.com/pipelines/#-quick-start-with-docker
   # https://ikasten.io/2024/06/03/getting-started-with-openwebui-pipelines/
   # https://raw.githubusercontent.com/open-webui/pipelines/main/examples/filters/function_calling_filter_pipeline.py
@@ -69,8 +48,6 @@ services:
     restart: unless-stopped
 
 volumes:
-  ollama:
-    name: ollama
   open-webui:
     name: open-webui
   pipelines: