From afa591afb2f9715b1dcbd4df798c5feb66eab2a9 Mon Sep 17 00:00:00 2001
From: "Timothy J. Baek" <timothyjrbeck@gmail.com>
Date: Wed, 20 Mar 2024 20:58:23 -0700
Subject: [PATCH] Create Dockerfile-ollama

---
 Dockerfile-ollama | 87 +++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 87 insertions(+)
 create mode 100644 Dockerfile-ollama

diff --git a/Dockerfile-ollama b/Dockerfile-ollama
new file mode 100644
index 000000000..35066e748
--- /dev/null
+++ b/Dockerfile-ollama
@@ -0,0 +1,87 @@
+# syntax=docker/dockerfile:1
+
+FROM node:alpine as build
+
+WORKDIR /app
+
+# wget embedding model weight from alpine (does not exist from slim-buster)
+RUN wget "https://chroma-onnx-models.s3.amazonaws.com/all-MiniLM-L6-v2/onnx.tar.gz" -O - | \
+    tar -xzf - -C /app
+
+COPY package.json package-lock.json ./
+RUN npm ci
+
+COPY . .
+RUN npm run build
+
+
+FROM python:3.11-slim-bookworm as base
+
+ENV ENV=prod
+ENV PORT ""
+
+ENV OLLAMA_BASE_URL "http://localhost:11434"
+
+ENV OPENAI_API_BASE_URL ""
+ENV OPENAI_API_KEY ""
+
+ENV WEBUI_SECRET_KEY ""
+
+ENV SCARF_NO_ANALYTICS true
+ENV DO_NOT_TRACK true
+
+######## Preloaded models ########
+# whisper TTS Settings
+ENV WHISPER_MODEL="base"
+ENV WHISPER_MODEL_DIR="/app/backend/data/cache/whisper/models"
+
+# RAG Embedding Model Settings
+# any sentence transformer model; models to use can be found at https://huggingface.co/models?library=sentence-transformers
+# Leaderboard: https://huggingface.co/spaces/mteb/leaderboard 
+# for better persormance 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"
+# device type for whisper tts and embbeding models - "cpu" (default), "cuda" (nvidia gpu and CUDA required) or "mps" (apple silicon) - choosing this right can lead to better performance
+ENV RAG_EMBEDDING_MODEL_DEVICE_TYPE="cpu"
+ENV RAG_EMBEDDING_MODEL_DIR="/app/backend/data/cache/embedding/models"
+ENV SENTENCE_TRANSFORMERS_HOME $RAG_EMBEDDING_MODEL_DIR
+
+######## Preloaded models ########
+
+WORKDIR /app/backend
+
+# install python dependencies
+COPY ./backend/requirements.txt ./requirements.txt
+
+RUN apt-get update && apt-get install ffmpeg libsm6 libxext6 curl -y && \
+    curl -fsSL https://ollama.com/install.sh | sh && \
+    apt-get purge curl -y && \
+    apt-get autoremove -y
+
+RUN pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cpu --no-cache-dir
+RUN pip3 install -r requirements.txt --no-cache-dir
+
+# Install pandoc and netcat
+# RUN python -c "import pypandoc; pypandoc.download_pandoc()"
+RUN apt-get update \
+    && apt-get install -y pandoc netcat-openbsd \
+    && rm -rf /var/lib/apt/lists/*
+
+# preload embedding model
+RUN python -c "import os; from chromadb.utils import embedding_functions; sentence_transformer_ef = embedding_functions.SentenceTransformerEmbeddingFunction(model_name=os.environ['RAG_EMBEDDING_MODEL'], device=os.environ['RAG_EMBEDDING_MODEL_DEVICE_TYPE'])"
+# preload tts model
+RUN python -c "import os; from faster_whisper import WhisperModel; WhisperModel(os.environ['WHISPER_MODEL'], device='auto', compute_type='int8', download_root=os.environ['WHISPER_MODEL_DIR'])"
+
+# copy embedding weight from build
+RUN mkdir -p /root/.cache/chroma/onnx_models/all-MiniLM-L6-v2
+COPY --from=build /app/onnx /root/.cache/chroma/onnx_models/all-MiniLM-L6-v2/onnx
+
+# copy built frontend files
+COPY --from=build /app/build /app/build
+COPY --from=build /app/CHANGELOG.md /app/CHANGELOG.md
+COPY --from=build /app/package.json /app/package.json
+
+# copy backend files
+COPY ./backend .
+
+CMD ["sh", "-c", "ollama serve & bash start.sh"]
\ No newline at end of file