Refactor lvms

.github/workflows/docker/compose/lvms-compose.yaml

@@ -2,44 +2,40 @@
 # SPDX-License-Identifier: Apache-2.0
 
 # this file should be run in the root of the repo
+# Deprecated original wrappers opea/lvm-tgi, opea/lvm-llava-svc, opea/lvm-video-llama
+# and merged functionalities in opea/lvm
+# Original video-llama-lvm-server renamed as lvm-video-llama
+
 services:
-  lvm-tgi:
+  lvm:
     build:
-      dockerfile: comps/lvms/tgi-llava/Dockerfile
-    image: ${REGISTRY:-opea}/lvm-tgi:${TAG:-latest}
+      dockerfile: comps/lvms/src/Dockerfile
+    image: ${REGISTRY:-opea}/lvm:${TAG:-latest}
   lvm-video-llama:
     build:
-      dockerfile: comps/lvms/video-llama/Dockerfile
+      dockerfile: comps/lvms/src/integrations/dependency/video-llama/Dockerfile
     image: ${REGISTRY:-opea}/lvm-video-llama:${TAG:-latest}
-  video-llama-lvm-server:
-    build:
-      dockerfile: comps/lvms/video-llama/dependency/Dockerfile
-    image: ${REGISTRY:-opea}/video-llama-lvm-server:${TAG:-latest}
   lvm-llava:
     build:
-      dockerfile: comps/lvms/llava/dependency/Dockerfile
+      dockerfile: comps/lvms/src/integrations/dependency/llava/Dockerfile
     image: ${REGISTRY:-opea}/lvm-llava:${TAG:-latest}
-  lvm-llava-svc:
-    build:
-      dockerfile: comps/lvms/llava/Dockerfile
-    image: ${REGISTRY:-opea}/lvm-llava-svc:${TAG:-latest}
   llava-gaudi:
     build:
-      dockerfile: comps/lvms/llava/dependency/Dockerfile.intel_hpu
+      dockerfile: comps/lvms/src/integrations/dependency/llava/Dockerfile.intel_hpu
     image: ${REGISTRY:-opea}/llava-gaudi:${TAG:-latest}
   lvm-predictionguard:
     build:
-      dockerfile: comps/lvms/predictionguard/Dockerfile
+      dockerfile: comps/lvms/src/integrations/dependency/predictionguard/Dockerfile
     image: ${REGISTRY:-opea}/lvm-predictionguard:${TAG:-latest}
   lvm-llama-vision:
     build:
-      dockerfile: comps/lvms/llama-vision/Dockerfile
+      dockerfile: comps/lvms/src/integrations/dependency/llama-vision/Dockerfile
     image: ${REGISTRY:-opea}/lvm-llama-vision:${TAG:-latest}
   lvm-llama-vision-tp:
     build:
-      dockerfile: comps/lvms/llama-vision/Dockerfile_tp
+      dockerfile: comps/lvms/src/integrations/dependency/llama-vision/Dockerfile_tp
     image: ${REGISTRY:-opea}/lvm-llama-vision-tp:${TAG:-latest}
   lvm-llama-vision-guard:
     build:
-      dockerfile: comps/lvms/llama-vision/Dockerfile_guard
+      dockerfile: comps/lvms/src/integrations/dependency/llama-vision/Dockerfile_guard
     image: ${REGISTRY:-opea}/lvm-llama-vision-guard:${TAG:-latest}

comps/cores/mega/http_service.py

@@ -8,6 +8,7 @@
 from typing import Optional
 
 from fastapi import FastAPI
+from fastapi.responses import Response
 from prometheus_fastapi_instrumentator import Instrumentator
 from uvicorn import Config, Server
 
@@ -73,6 +74,11 @@
             """Get the health status of this GenAI microservice."""
             return {"Service Title": self.title, "Service Description": self.description}
 
+        @app.get("/health")
+        async def _health() -> Response:
+            """Health check."""
+            return Response(status_code=200)
+
         @app.get(
             path="/v1/statistics",
             summary="Get the statistics of GenAI services",

comps/dataprep/README.md

@@ -11,7 +11,7 @@ apt-get install libreoffice
 
 ## Use LVM (Large Vision Model) for Summarizing Image Data
 
-Occasionally unstructured data will contain image data, to convert the image data to the text data, LVM can be used to summarize the image. To leverage LVM, please refer to this [readme](../lvms/llava/README.md) to start the LVM microservice first and then set the below environment variable, before starting any dataprep microservice.
+Occasionally unstructured data will contain image data, to convert the image data to the text data, LVM can be used to summarize the image. To leverage LVM, please refer to this [readme](../lvms/src/README.md) to start the LVM microservice first and then set the below environment variable, before starting any dataprep microservice.
 
 ```bash
 export SUMMARIZE_IMAGE_VIA_LVM=1

comps/dataprep/multimodal/redis/langchain/README.md

@@ -38,7 +38,7 @@ export PYTHONPATH=${path_to_comps}
 
 This is required only if you are going to consume the _generate_captions_ API of this microservice as in [Section 4.3](#43-consume-generate_captions-api).
 
-Please refer to this [readme](../../../../lvms/llava/README.md) to start the LVM microservice.
+Please refer to this [readme](../../../../lvms/src/README.md) to start the LVM microservice.
 After LVM is up, set up environment variables.
 
 ```bash
@@ -64,7 +64,7 @@ Please refer to this [readme](../../../../vectorstores/redis/README.md).
 
 This is required only if you are going to consume the _generate_captions_ API of this microservice as described [here](#43-consume-generate_captions-api).
 
-Please refer to this [readme](../../../../lvms/llava/README.md) to start the LVM microservice.
+Please refer to this [readme](../../../../lvms/src/README.md) to start the LVM microservice.
 After LVM is up, set up environment variables.
 
 ```bash

comps/dataprep/src/README.md

@@ -11,7 +11,7 @@ apt-get install libreoffice
 
 ## Use LVM (Large Vision Model) for Summarizing Image Data
 
-Occasionally unstructured data will contain image data, to convert the image data to the text data, LVM can be used to summarize the image. To leverage LVM, please refer to this [readme](../../lvms/llava/README.md) to start the LVM microservice first and then set the below environment variable, before starting any dataprep microservice.
+Occasionally unstructured data will contain image data, to convert the image data to the text data, LVM can be used to summarize the image. To leverage LVM, please refer to this [readme](../../lvms/src/README.md) to start the LVM microservice first and then set the below environment variable, before starting any dataprep microservice.
 
 ```bash
 export SUMMARIZE_IMAGE_VIA_LVM=1

comps/lvms/src/Dockerfile

@@ -0,0 +1,23 @@
+# Copyright (C) 2024 Intel Corporation
+# SPDX-License-Identifier: Apache-2.0
+
+FROM python:3.11-slim
+
+RUN useradd -m -s /bin/bash user && \
+    mkdir -p /home/user && \
+    chown -R user /home/user/
+
+USER user
+
+ENV LANG=C.UTF-8
+
+COPY comps /home/user/comps
+
+RUN pip install --no-cache-dir --upgrade pip setuptools && \
+    pip install --no-cache-dir -r /home/user/comps/lvms/src/requirements.txt
+
+ENV PYTHONPATH=$PYTHONPATH:/home/user
+
+WORKDIR /home/user/comps/lvms/src
+
+ENTRYPOINT ["python", "opea_lvm_microservice.py"]

comps/lvms/src/README.md

@@ -0,0 +1,31 @@
+# LVM Microservice
+
+Visual Question and Answering is one of the multimodal tasks empowered by LVMs (Large Visual Models). This microservice supports visual Q&A by using LLaVA as the base large visual model. It accepts two inputs: a prompt and an image. It outputs the answer to the prompt about the image.
+
+## Build Image & Run
+
+Before this, you have to start the [dependency](./integrations/dependency/) service based on your demands.
+
+```bash
+docker build --no-cache -t opea/lvm:comps --build-arg https_proxy=$https_proxy --build-arg http_proxy=$http_proxy  -f comps/lvms/src/Dockerfile .
+# Change LVM_ENDPOINT to you dependency service endpoint
+docker run -d --name="test-comps-lvm" -e LVM_ENDPOINT=http://localhost:8399 -e http_proxy=$http_proxy -e https_proxy=$https_proxy -p 9399:9399 --ipc=host opea/lvm:comps
+```
+
+## Test
+
+- LLaVA & llama-vision & PredictionGuard & TGI LLaVA
+
+```bash
+# curl with an image and a prompt
+http_proxy="" curl http://localhost:9399/v1/lvm -XPOST -d '{"image": "iVBORw0KGgoAAAANSUhEUgAAAAoAAAAKCAYAAACNMs+9AAAAFUlEQVR42mP8/5+hnoEIwDiqkL4KAcT9GO0U4BxoAAAAAElFTkSuQmCC", "prompt":"What is this?"}' -H 'Content-Type: application/json'
+
+# curl with only the prompt
+http_proxy="" curl http://localhost:9399/v1/lvm --silent --write-out "HTTPSTATUS:%{http_code}" -XPOST -d '{"image": "", "prompt":"What is deep learning?"}' -H 'Content-Type: application/json'
+```
+
+- video-llama
+
+```bash
+http_proxy="" curl -X POST http://localhost:9399/v1/lvm -d '{"video_url":"https://github.com/DAMO-NLP-SG/Video-LLaMA/raw/main/examples/silence_girl.mp4","chunk_start": 0,"chunk_duration": 9,"prompt":"What is the person doing?","max_new_tokens": 150}' -H 'Content-Type: application/json'
+```

comps/lvms/llama-vision/Dockerfile → comps/lvms/src/integrations/dependency/llama-vision/Dockerfile

@@ -21,12 +21,12 @@ RUN git lfs install
 
 COPY comps /home/user/comps
 
-RUN cd /home/user/comps/lvms/llama-vision/ && \
+RUN cd /home/user/comps/lvms/src/integrations/dependency/llama-vision/ && \
     pip install --no-cache-dir -r requirements.txt  && \
     pip install --no-cache-dir --upgrade Pillow
 
 ENV PYTHONPATH=/root:/home/user
 
-WORKDIR /home/user/comps/lvms/llama-vision/
+WORKDIR /home/user/comps/lvms/src/integrations/dependency/llama-vision/
 
 ENTRYPOINT ["python", "lvm.py"]

comps/lvms/llama-vision/Dockerfile_guard → comps/lvms/src/integrations/dependency/llama-vision/Dockerfile_guard

@@ -21,12 +21,12 @@ RUN git lfs install
 
 COPY comps /home/user/comps
 
-RUN cd /home/user/comps/lvms/llama-vision/ && \
+RUN cd /home/user/comps/lvms/src/integrations/dependency/llama-vision/ && \
     pip install --no-cache-dir -r requirements.txt  && \
     pip install --no-cache-dir --upgrade Pillow
 
 ENV PYTHONPATH=/root:/home/user
 
-WORKDIR /home/user/comps/lvms/llama-vision/
+WORKDIR /home/user/comps/lvms/src/integrations/dependency/llama-vision/
 
 ENTRYPOINT ["python", "lvm_guard.py"]

comps/lvms/llama-vision/Dockerfile_tp → comps/lvms/src/integrations/dependency/llama-vision/Dockerfile_tp

@@ -22,13 +22,13 @@ COPY comps /home/user/comps
 RUN pip install --no-cache-dir git+https://github.com/HabanaAI/[email protected]
 RUN pip install --no-cache-dir git+https://github.com/huggingface/[email protected]
 
-RUN cd /home/user/comps/lvms/llama-vision/  \
+RUN cd /home/user/comps/lvms/src/integrations/dependency/llama-vision/  \
     pip install --no-cache-dir --upgrade pip && \
     bash update && \
-    pip install --no-cache-dir -r /home/user/comps/lvms/llama-vision/requirements_tp.txt
+    pip install --no-cache-dir -r /home/user/comps/lvms/src/integrations/dependency/llama-vision/requirements_tp.txt
 
 ENV PYTHONPATH=/root:/home/user
 
-WORKDIR /home/user/comps/lvms/llama-vision/
+WORKDIR /home/user/comps/lvms/src/integrations/dependency/llama-vision/
 
 ENTRYPOINT ["bash", "run_tp.sh"]