feat(rerankers): Add new backend, support jina rerankers API

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

.github/workflows/test-extra.yml

@@ -74,6 +74,37 @@ jobs:
            make --jobs=5 --output-sync=target -C backend/python/sentencetransformers
            make --jobs=5 --output-sync=target -C backend/python/sentencetransformers test
 
+
+  tests-rerankers:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Clone
+        uses: actions/checkout@v4
+        with: 
+          submodules: true
+      - name: Dependencies
+        run: |
+          sudo apt-get update
+          sudo apt-get install build-essential ffmpeg
+          curl https://repo.anaconda.com/pkgs/misc/gpgkeys/anaconda.asc | gpg --dearmor > conda.gpg && \
+             sudo install -o root -g root -m 644 conda.gpg /usr/share/keyrings/conda-archive-keyring.gpg && \
+              gpg --keyring /usr/share/keyrings/conda-archive-keyring.gpg --no-default-keyring --fingerprint 34161F5BF5EB1D4BFBBB8F0A8AEB4F8B29D82806 && \
+             sudo /bin/bash -c 'echo "deb [arch=amd64 signed-by=/usr/share/keyrings/conda-archive-keyring.gpg] https://repo.anaconda.com/pkgs/misc/debrepo/conda stable main" > /etc/apt/sources.list.d/conda.list' && \
+             sudo /bin/bash -c 'echo "deb [arch=amd64 signed-by=/usr/share/keyrings/conda-archive-keyring.gpg] https://repo.anaconda.com/pkgs/misc/debrepo/conda stable main" | tee -a /etc/apt/sources.list.d/conda.list' && \
+             sudo apt-get update && \
+             sudo apt-get install -y conda
+          sudo apt-get install -y ca-certificates cmake curl patch python3-pip
+          sudo apt-get install -y libopencv-dev
+          pip install --user grpcio-tools
+          
+          sudo rm -rfv /usr/bin/conda || true
+
+      - name: Test rerankers
+        run: |
+           export PATH=$PATH:/opt/conda/bin
+           make --jobs=5 --output-sync=target -C backend/python/rerankers
+           make --jobs=5 --output-sync=target -C backend/python/rerankers test
+
   tests-diffusers:
     runs-on: ubuntu-latest
     steps:

Dockerfile

@@ -16,7 +16,7 @@ ARG TARGETVARIANT
 
 ENV BUILD_TYPE=${BUILD_TYPE}
 ENV DEBIAN_FRONTEND=noninteractive
-ENV EXTERNAL_GRPC_BACKENDS="coqui:/build/backend/python/coqui/run.sh,huggingface-embeddings:/build/backend/python/sentencetransformers/run.sh,petals:/build/backend/python/petals/run.sh,transformers:/build/backend/python/transformers/run.sh,sentencetransformers:/build/backend/python/sentencetransformers/run.sh,autogptq:/build/backend/python/autogptq/run.sh,bark:/build/backend/python/bark/run.sh,diffusers:/build/backend/python/diffusers/run.sh,exllama:/build/backend/python/exllama/run.sh,vall-e-x:/build/backend/python/vall-e-x/run.sh,vllm:/build/backend/python/vllm/run.sh,mamba:/build/backend/python/mamba/run.sh,exllama2:/build/backend/python/exllama2/run.sh,transformers-musicgen:/build/backend/python/transformers-musicgen/run.sh,parler-tts:/build/backend/python/parler-tts/run.sh"
+ENV EXTERNAL_GRPC_BACKENDS="coqui:/build/backend/python/coqui/run.sh,huggingface-embeddings:/build/backend/python/sentencetransformers/run.sh,petals:/build/backend/python/petals/run.sh,transformers:/build/backend/python/transformers/run.sh,sentencetransformers:/build/backend/python/sentencetransformers/run.sh,rerankers:/build/backend/python/rerankers/run.sh,autogptq:/build/backend/python/autogptq/run.sh,bark:/build/backend/python/bark/run.sh,diffusers:/build/backend/python/diffusers/run.sh,exllama:/build/backend/python/exllama/run.sh,vall-e-x:/build/backend/python/vall-e-x/run.sh,vllm:/build/backend/python/vllm/run.sh,mamba:/build/backend/python/mamba/run.sh,exllama2:/build/backend/python/exllama2/run.sh,transformers-musicgen:/build/backend/python/transformers-musicgen/run.sh,parler-tts:/build/backend/python/parler-tts/run.sh"
 
 ARG GO_TAGS="stablediffusion tinydream tts"
 
@@ -259,6 +259,9 @@ RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
 RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
     make -C backend/python/sentencetransformers \
     ; fi
+RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
+    make -C backend/python/rerankers \
+    ; fi
 RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
     make -C backend/python/transformers \
     ; fi

Makefile

@@ -437,10 +437,10 @@ protogen-go-clean:
 	$(RM) bin/*
 
 .PHONY: protogen-python
-protogen-python: autogptq-protogen bark-protogen coqui-protogen diffusers-protogen exllama-protogen exllama2-protogen mamba-protogen petals-protogen sentencetransformers-protogen transformers-protogen parler-tts-protogen transformers-musicgen-protogen vall-e-x-protogen vllm-protogen
+protogen-python: autogptq-protogen bark-protogen coqui-protogen diffusers-protogen exllama-protogen exllama2-protogen mamba-protogen petals-protogen rerankers-protogen sentencetransformers-protogen transformers-protogen parler-tts-protogen transformers-musicgen-protogen vall-e-x-protogen vllm-protogen
 
 .PHONY: protogen-python-clean
-protogen-python-clean: autogptq-protogen-clean bark-protogen-clean coqui-protogen-clean diffusers-protogen-clean exllama-protogen-clean exllama2-protogen-clean mamba-protogen-clean petals-protogen-clean sentencetransformers-protogen-clean transformers-protogen-clean transformers-musicgen-protogen-clean parler-tts-protogen-clean vall-e-x-protogen-clean vllm-protogen-clean
+protogen-python-clean: autogptq-protogen-clean bark-protogen-clean coqui-protogen-clean diffusers-protogen-clean exllama-protogen-clean exllama2-protogen-clean mamba-protogen-clean petals-protogen-clean sentencetransformers-protogen-clean rerankers-protogen-clean transformers-protogen-clean transformers-musicgen-protogen-clean parler-tts-protogen-clean vall-e-x-protogen-clean vllm-protogen-clean
 
 .PHONY: autogptq-protogen
 autogptq-protogen:
@@ -506,6 +506,14 @@ petals-protogen:
 petals-protogen-clean:
 	$(MAKE) -C backend/python/petals protogen-clean
 
+.PHONY: rerankers-protogen
+rerankers-protogen:
+	$(MAKE) -C backend/python/rerankers protogen
+
+.PHONY: rerankers-protogen-clean
+rerankers-protogen-clean:
+	$(MAKE) -C backend/python/rerankers protogen-clean
+
 .PHONY: sentencetransformers-protogen
 sentencetransformers-protogen:
 	$(MAKE) -C backend/python/sentencetransformers protogen
@@ -564,6 +572,7 @@ prepare-extra-conda-environments: protogen-python
 	$(MAKE) -C backend/python/vllm
 	$(MAKE) -C backend/python/mamba
 	$(MAKE) -C backend/python/sentencetransformers
+	$(MAKE) -C backend/python/rerankers
 	$(MAKE) -C backend/python/transformers
 	$(MAKE) -C backend/python/transformers-musicgen
 	$(MAKE) -C backend/python/parler-tts

aio/cpu/rerank.yaml

@@ -0,0 +1,27 @@
+name: jina-reranker-v1-base-en
+backend: rerankers
+parameters:
+  model: cross-encoder
+
+usage: |
+    You can test this model with curl like this:
+
+    curl http://localhost:8080/v1/rerank \
+      -H "Content-Type: application/json" \
+      -d '{
+      "model": "jina-reranker-v1-base-en",
+      "query": "Organic skincare products for sensitive skin",
+      "documents": [
+        "Eco-friendly kitchenware for modern homes",
+        "Biodegradable cleaning supplies for eco-conscious consumers",
+        "Organic cotton baby clothes for sensitive skin",
+        "Natural organic skincare range for sensitive skin",
+        "Tech gadgets for smart homes: 2024 edition",
+        "Sustainable gardening tools and compost solutions",
+        "Sensitive skin-friendly facial cleansers and toners",
+        "Organic food wraps and storage solutions",
+        "All-natural pet food for dogs with allergies",
+        "Yoga mats made from recycled materials"
+      ],
+      "top_n": 3
+    }'

aio/entrypoint.sh

@@ -129,7 +129,7 @@ detect_gpu
 detect_gpu_size
 
 PROFILE="${PROFILE:-$GPU_SIZE}" # default to cpu
-export MODELS="${MODELS:-/aio/${PROFILE}/embeddings.yaml,/aio/${PROFILE}/text-to-speech.yaml,/aio/${PROFILE}/image-gen.yaml,/aio/${PROFILE}/text-to-text.yaml,/aio/${PROFILE}/speech-to-text.yaml,/aio/${PROFILE}/vision.yaml}"
+export MODELS="${MODELS:-/aio/${PROFILE}/embeddings.yaml,/aio/${PROFILE}/rerank.yaml,/aio/${PROFILE}/text-to-speech.yaml,/aio/${PROFILE}/image-gen.yaml,/aio/${PROFILE}/text-to-text.yaml,/aio/${PROFILE}/speech-to-text.yaml,/aio/${PROFILE}/vision.yaml}"
 
 check_vars
 

aio/gpu-8g/rerank.yaml

@@ -0,0 +1,27 @@
+name: jina-reranker-v1-base-en
+backend: rerankers
+parameters:
+  model: cross-encoder
+
+usage: |
+    You can test this model with curl like this:
+
+    curl http://localhost:8080/v1/rerank \
+      -H "Content-Type: application/json" \
+      -d '{
+      "model": "jina-reranker-v1-base-en",
+      "query": "Organic skincare products for sensitive skin",
+      "documents": [
+        "Eco-friendly kitchenware for modern homes",
+        "Biodegradable cleaning supplies for eco-conscious consumers",
+        "Organic cotton baby clothes for sensitive skin",
+        "Natural organic skincare range for sensitive skin",
+        "Tech gadgets for smart homes: 2024 edition",
+        "Sustainable gardening tools and compost solutions",
+        "Sensitive skin-friendly facial cleansers and toners",
+        "Organic food wraps and storage solutions",
+        "All-natural pet food for dogs with allergies",
+        "Yoga mats made from recycled materials"
+      ],
+      "top_n": 3
+    }'

aio/intel/rerank.yaml

@@ -0,0 +1,27 @@
+name: jina-reranker-v1-base-en
+backend: rerankers
+parameters:
+  model: cross-encoder
+
+usage: |
+    You can test this model with curl like this:
+
+    curl http://localhost:8080/v1/rerank \
+      -H "Content-Type: application/json" \
+      -d '{
+      "model": "jina-reranker-v1-base-en",
+      "query": "Organic skincare products for sensitive skin",
+      "documents": [
+        "Eco-friendly kitchenware for modern homes",
+        "Biodegradable cleaning supplies for eco-conscious consumers",
+        "Organic cotton baby clothes for sensitive skin",
+        "Natural organic skincare range for sensitive skin",
+        "Tech gadgets for smart homes: 2024 edition",
+        "Sustainable gardening tools and compost solutions",
+        "Sensitive skin-friendly facial cleansers and toners",
+        "Organic food wraps and storage solutions",
+        "All-natural pet food for dogs with allergies",
+        "Yoga mats made from recycled materials"
+      ],
+      "top_n": 3
+    }'

backend/backend.proto

@@ -23,6 +23,30 @@ service Backend {
   rpc StoresDelete(StoresDeleteOptions) returns (Result) {}
   rpc StoresGet(StoresGetOptions) returns (StoresGetResult) {}
   rpc StoresFind(StoresFindOptions) returns (StoresFindResult) {}
+
+  rpc Rerank(RerankRequest) returns (RerankResult) {}
+}
+
+message RerankRequest {
+  string query = 1;
+  repeated string documents = 2;
+  int32 top_n = 3;
+}
+
+message RerankResult {
+  Usage usage = 1;
+  repeated DocumentResult results = 2;
+}
+
+message Usage {
+  int32 total_tokens = 1;
+  int32 prompt_tokens = 2;
+}
+
+message DocumentResult {
+  int32 index = 1;
+  string text = 2;
+  float relevance_score = 3;
 }
 
 message StoresKey {

backend/python/common-env/transformers/transformers-nvidia.yml

@@ -120,4 +120,6 @@ dependencies:
       - transformers>=4.38.2  # Updated Version
       - transformers_stream_generator==0.0.5
       - xformers==0.0.23.post1  
+      - rerankers[transformers]
+      - pydantic
 prefix: /opt/conda/envs/transformers

backend/python/common-env/transformers/transformers-rocm.yml

@@ -108,4 +108,6 @@ dependencies:
       - transformers>=4.38.2  # Updated Version
       - transformers_stream_generator==0.0.5
       - xformers==0.0.23.post1
+      - rerankers[transformers]
+      - pydantic
 prefix: /opt/conda/envs/transformers

backend/python/common-env/transformers/transformers.yml

@@ -111,5 +111,7 @@ dependencies:
       - vllm>=0.4.0
       - transformers>=4.38.2  # Updated Version
       - transformers_stream_generator==0.0.5
-      - xformers==0.0.23.post1  
+      - xformers==0.0.23.post1
+      - rerankers[transformers]
+      - pydantic
 prefix: /opt/conda/envs/transformers

backend/python/rerankers/Makefile

@@ -0,0 +1,27 @@
+.PHONY: rerankers
+rerankers: protogen
+	$(MAKE) -C ../common-env/transformers
+
+
+.PHONY: run
+run: protogen
+	@echo "Running rerankers..."
+	bash run.sh
+	@echo "rerankers run."
+
+# It is not working well by using command line. It only6 works with IDE like VSCode.
+.PHONY: test
+test: protogen
+	@echo "Testing rerankers..."
+	bash test.sh
+	@echo "rerankers tested."
+
+.PHONY: protogen
+protogen: backend_pb2_grpc.py backend_pb2.py
+
+.PHONY: protogen-clean
+protogen-clean:
+	$(RM) backend_pb2_grpc.py backend_pb2.py
+
+backend_pb2_grpc.py backend_pb2.py:
+	python3 -m grpc_tools.protoc -I../.. --python_out=. --grpc_python_out=. backend.proto

backend/python/rerankers/README.md

@@ -0,0 +1,5 @@
+# Creating a separate environment for the reranker project
+
+```
+make reranker
+```