Initial commit, first public version.

src/aichat/Dockerfile

@@ -0,0 +1,7 @@
+FROM rust:latest
+
+RUN rustup target add x86_64-unknown-linux-musl
+RUN apt update && apt install -y musl-tools musl-dev
+RUN update-ca-certificates
+
+RUN cargo install --target x86_64-unknown-linux-musl aichat

src/aichat/config.yaml

@@ -0,0 +1,8 @@
+# see https://github.com/sigoden/aichat/blob/main/config.example.yaml
+
+model: ollama
+clients:
+- type: openai-compatible
+  name: ollama
+  api_base: http://localhost:11434/v1
+  api_key: __LLM_API_KEY__

src/llm_provision/Dockerfile

@@ -0,0 +1,12 @@
+FROM debian:bookworm-slim
+
+ENV DEBIAN_FRONTEND=noninteractive
+RUN apt-get update
+RUN apt-get --yes -o Dpkg::Options::="--force-confdef" -o Dpkg::Options::="--force-confnew" install bash curl jq
+
+ADD ./src/llm_provision/init_models.sh /init_models.sh
+ADD ./src/llm_provision/entrypoint.sh /entrypoint.sh
+RUN chmod 755 /entrypoint.sh
+
+ENTRYPOINT ["/entrypoint.sh"]
+#ENTRYPOINT ["tail", "-f", "/dev/null"] # to debug

src/llm_provision/entrypoint.sh

@@ -0,0 +1,4 @@
+#!/usr/bin/env bash
+
+echo "pull models into ollama volumes"
+bash /init_models.sh

src/llm_provision/init_models.sh

@@ -0,0 +1,17 @@
+#!/usr/bin/env bash
+
+OLLAMA_HOST="http://ollama:11434"
+
+IFS=',' read -r -a models_arr <<< "${MODELS}"
+
+## now loop through the above array
+for m in "${models_arr[@]}"
+do
+  curl -s "${OLLAMA_HOST}/api/tags" | jq '.models[].name' | grep ${m} > /dev/null
+  if [[ $? -ne 0 ]]
+  then
+    curl -s "${OLLAMA_HOST}/api/pull" -d "{\"model\": \"${m}\"}"
+  else
+    echo "${m} already installed"
+  fi
+done

src/nginx/nginx.conf

@@ -0,0 +1,61 @@
+events{}
+http {
+    server_tokens off;
+    client_max_body_size 200m;
+
+    server {
+        listen 11434;
+        set $deny 1;
+        if ($http_authorization = "Bearer $API_KEY") {
+            set $deny 0;
+        }
+        if ($deny) {
+            return 403;
+        }
+        location / {
+            proxy_pass http://ollama:11434;
+            proxy_set_header Host $host;
+            proxy_set_header X-Real-IP $remote_addr;
+            proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
+            proxy_set_header X-Forwarded-Proto $scheme;
+        }
+    }
+    server {
+        listen 8000;
+        set $deny 1;
+        if ($http_authorization = "Bearer $API_KEY") {
+            set $deny 0;
+        }
+        if ($deny) {
+            return 403;
+        }
+        location / {
+            proxy_pass http://openedai-speech:8000;
+            proxy_set_header Host $host;
+            proxy_set_header X-Real-IP $remote_addr;
+            proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
+            proxy_set_header X-Forwarded-Proto $scheme;
+        }
+    }
+    server {
+        listen 8001;
+        set $deny 1;
+        if ($http_authorization = "Bearer $API_KEY") {
+            set $deny 0;
+        }
+        if ($deny) {
+            return 403;
+        }
+        location / {
+            proxy_pass http://faster-whisper-server:8000;
+            proxy_set_header Host $host;
+            proxy_set_header X-Real-IP $remote_addr;
+            proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
+            proxy_set_header X-Forwarded-Proto $scheme;
+            proxy_read_timeout 180;
+            proxy_http_version 1.1;
+            proxy_set_header Upgrade $http_upgrade;
+            proxy_set_header Connection "upgrade";
+        }
+    }
+}

src/tts/Dockerfile

@@ -0,0 +1,47 @@
+FROM python:3.11-slim
+
+RUN --mount=type=cache,target=/root/.cache/pip pip install -U pip
+
+ARG TARGETPLATFORM
+RUN <<EOF
+apt-get update
+apt-get install --no-install-recommends -y curl ffmpeg git
+if [ "$TARGETPLATFORM" != "linux/amd64" ]; then
+	apt-get install --no-install-recommends -y build-essential
+	curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- -y
+fi
+
+# for deepspeed support - image +7.5GB, over the 10GB ghcr.io limit, and no noticable gain in speed or VRAM usage?
+#curl -O https://developer.download.nvidia.com/compute/cuda/repos/debian11/x86_64/cuda-keyring_1.1-1_all.deb
+#dpkg -i cuda-keyring_1.1-1_all.deb
+#rm cuda-keyring_1.1-1_all.deb
+#apt-get install --no-install-recommends -y libaio-dev build-essential cuda-toolkit
+
+apt-get clean
+rm -rf /var/lib/apt/lists/*
+EOF
+#ENV CUDA_HOME=/usr/local/cuda
+ENV PATH="/root/.cargo/bin:${PATH}"
+
+WORKDIR /app
+RUN mkdir -p voices config
+
+ARG USE_ROCM
+ENV USE_ROCM=${USE_ROCM}
+
+RUN git clone https://github.com/matatonic/openedai-speech.git /tmp/app
+RUN mv /tmp/app/* /app/
+ADD src/tts/download_voices_tts-1.sh /app/download_voices_tts-1.sh
+ADD src/tts/voice_to_speaker.default.yaml /app/voice_to_speaker.default.yaml
+RUN if [ "${USE_ROCM}" = "1" ]; then mv /app/requirements-rocm.txt /app/requirements.txt; fi
+RUN --mount=type=cache,target=/root/.cache/pip pip install -r requirements.txt
+
+
+ARG PRELOAD_MODEL
+ENV PRELOAD_MODEL=${PRELOAD_MODEL}
+ENV TTS_HOME=voices
+ENV HF_HOME=voices
+ENV COQUI_TOS_AGREED=1
+
+CMD bash startup.sh
+

src/tts/download_voices_tts-1.sh

@@ -0,0 +1,8 @@
+#!/bin/sh
+# cat voice_to_speaker.default.yaml | yq '.tts-1 ' | grep mode | cut -d'/' -f2 | cut -d'.' -f1 | sort -u | xargs
+models=${*:-"en_GB-alba-medium en_GB-northern_english_male-medium en_US-bryce-medium en_US-john-medium en_US-libritts_r-medium en_US-ryan-high fr_FR-siwis-medium fr_FR-tom-medium fr_FR-upmc-medium"}
+piper --update-voices --data-dir voices --download-dir voices --model x 2> /dev/null
+for i in $models ; do
+    [ ! -e "voices/$i.onnx" ] && piper --data-dir voices --download-dir voices --model $i < /dev/null > /dev/null
+done
+

src/tts/voice_to_speaker.default.yaml

@@ -0,0 +1,36 @@
+# Use https://rhasspy.github.io/piper-samples/ to configure
+tts-1:
+  alloy:
+    model: voices/en_US-libritts_r-medium.onnx
+    speaker: 79 
+  siwis:
+    model: voices/fr_FR-siwis-medium.onnx
+    speaker: 0
+  tom:
+    model: voices/fr_FR-tom-medium.onnx
+    speaker: 0
+  pierre:
+    model: voices/fr_FR-upmc-medium.onnx
+    speaker: 1
+  jessica:
+    model: voices/fr_FR-upmc-medium.onnx
+    speaker: 0
+  alba:
+    model: voices/en_GB-alba-medium.onnx
+    speaker: 0
+  jack:
+    model: voices/en_GB-northern_english_male-medium.onnx
+    speaker: 0
+  john:
+    model: voices/en_US-john-medium.onnx
+    speaker: 0
+  bryce:
+    model: voices/en_US-bryce-medium.onnx
+    speaker: 0
+  ryan:
+    model: voices/en_US-ryan-high.onnx
+    speaker: 0
+  echo:
+    model: voices/en_US-libritts_r-medium.onnx
+    speaker: 134 
+

src/whisper/Dockerfile

@@ -0,0 +1,13 @@
+FROM debian:bookworm-slim
+
+RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y \
+    sudo \
+    python3 \
+    python3-distutils \
+    python3-pip \
+    ffmpeg
+
+RUN pip install -U openai-whisper --break-system-packages
+WORKDIR /app
+
+CMD ["whisper"]