Standardize name for LLM comps (opea-project#1162)

Update all the names for classes and files in llm comps to follow the standard format

Related to issue opea-project#998

Signed-off-by: Xinyao Wang <[email protected]>

comps/llms/deployment/docker_compose/text-generation_tgi.yaml

@@ -41,7 +41,7 @@ services:
       LLM_ENDPOINT: ${LLM_ENDPOINT}
       HF_TOKEN: ${HF_TOKEN}
       LLM_MODEL_ID: ${LLM_MODEL_ID}
-      LLM_COMPONENT_NAME: "OPEA_LLM"
+      LLM_COMPONENT_NAME: "OpeaTextGenService"
     restart: unless-stopped
 
 networks:

comps/llms/deployment/docker_compose/text-generation_vllm_langchain.yaml

@@ -39,7 +39,7 @@ services:
       vLLM_ENDPOINT: ${vLLM_ENDPOINT}
       HF_TOKEN: ${HF_TOKEN}
       LLM_MODEL: ${LLM_MODEL}
-      LLM_COMPONENT_NAME: "OPEA_LLM"
+      LLM_COMPONENT_NAME: "OpeaTextGenService"
     restart: unless-stopped
 
 networks:

comps/llms/src/doc-summarization/README.md

@@ -17,7 +17,7 @@ export LLM_ENDPOINT="http://${host_ip}:${LLM_ENDPOINT_PORT}"
 export LLM_MODEL_ID=${your_hf_llm_model}
 export MAX_INPUT_TOKENS=2048
 export MAX_TOTAL_TOKENS=4096
-export DocSum_COMPONENT_NAME="OPEADocSum_TGI" # or "OPEADocSum_vLLM"
+export DocSum_COMPONENT_NAME="OpeaDocSumTgi" # or "OpeaDocSumvLLM"
 ```
 
 Please make sure MAX_TOTAL_TOKENS should be larger than (MAX_INPUT_TOKENS + max_new_tokens + 50), 50 is reserved prompt length.

comps/llms/src/doc-summarization/integrations/common.py

@@ -56,7 +56,7 @@ def get_llm_endpoint():
             raise ConfigError(f"Input model {MODEL_NAME} not present in model_configs")
 
 
-class OPEADocSum(OpeaComponent):
+class OpeaDocSum(OpeaComponent):
     """A specialized OPEA DocSum component derived from OpeaComponent.
 
     Attributes:
@@ -72,7 +72,7 @@ def __init__(self, name: str, description: str, config: dict = None):
         self.tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
         health_status = self.check_health()
         if not health_status:
-            logger.error("OPEADocSum health check failed.")
+            logger.error("OpeaDocSum health check failed.")
 
     async def generate(self, input: DocSumLLMParams, client):
         """Invokes the TGI/vLLM LLM service to generate summarization for the provided input.

comps/llms/src/doc-summarization/integrations/tgi.py

@@ -14,9 +14,9 @@
 logflag = os.getenv("LOGFLAG", False)
 
 
-@OpeaComponentRegistry.register("OPEADocSum_TGI")
-class OPEADocSum_TGI(OPEADocSum):
-    """A specialized OPEA DocSum TGI component derived from OPEADocSum for interacting with TGI services based on Lanchain HuggingFaceEndpoint API.
+@OpeaComponentRegistry.register("OpeaDocSumTgi")
+class OpeaDocSumTgi(OpeaDocSum):
+    """A specialized OPEA DocSum TGI component derived from OpeaDocSum for interacting with TGI services based on Lanchain HuggingFaceEndpoint API.
 
     Attributes:
         client (TGI): An instance of the TGI client for text generation.

comps/llms/src/doc-summarization/integrations/vllm.py

@@ -14,9 +14,9 @@
 logflag = os.getenv("LOGFLAG", False)
 
 
-@OpeaComponentRegistry.register("OPEADocSum_vLLM")
-class OPEADocSum_vLLM(OPEADocSum):
-    """A specialized OPEA DocSum vLLM component derived from OPEADocSum for interacting with vLLM services based on Lanchain VLLMOpenAI API.
+@OpeaComponentRegistry.register("OpeaDocSumvLLM")
+class OpeaDocSumvLLM(OpeaDocSum):
+    """A specialized OPEA DocSum vLLM component derived from OpeaDocSum for interacting with vLLM services based on Lanchain VLLMOpenAI API.
 
     Attributes:
         client (vLLM): An instance of the vLLM client for text generation.

comps/llms/src/doc-summarization/opea_docsum_microservice.py

@@ -4,8 +4,8 @@
 import os
 import time
 
-from integrations.tgi import OPEADocSum_TGI
-from integrations.vllm import OPEADocSum_vLLM
+from integrations.tgi import OpeaDocSumTgi
+from integrations.vllm import OpeaDocSumvLLM
 
 from comps import (
     CustomLogger,
@@ -21,7 +21,7 @@
 logger = CustomLogger("llm_docsum")
 logflag = os.getenv("LOGFLAG", False)
 
-llm_component_name = os.getenv("DocSum_COMPONENT_NAME", "OPEADocSum_TGI")
+llm_component_name = os.getenv("DocSum_COMPONENT_NAME", "OpeaDocSumTgi")
 # Initialize OpeaComponentLoader
 loader = OpeaComponentLoader(llm_component_name, description=f"OPEA LLM DocSum Component: {llm_component_name}")
 

comps/llms/src/faq-generation/README.md

@@ -15,7 +15,7 @@ export FAQ_PORT=9000
 export HUGGINGFACEHUB_API_TOKEN=${your_hf_api_token}
 export LLM_ENDPOINT="http://${host_ip}:${LLM_ENDPOINT_PORT}"
 export LLM_MODEL_ID=${your_hf_llm_model}
-export FAQGen_COMPONENT_NAME="OPEAFAQGen_TGI" # or "vllm"
+export FAQGen_COMPONENT_NAME="OpeaFaqGenTgi" # or "vllm"
 ```
 
 ### 1.2 Build Docker Image

comps/llms/src/faq-generation/integrations/common.py

@@ -56,7 +56,7 @@ def get_llm_endpoint():
             raise ConfigError(f"Input model {MODEL_NAME} not present in model_configs")
 
 
-class OPEAFAQGen(OpeaComponent):
+class OpeaFaqGen(OpeaComponent):
     """A specialized OPEA FAQGen component derived from OpeaComponent.
 
     Attributes:
@@ -72,7 +72,7 @@ def __init__(self, name: str, description: str, config: dict = None):
         self.llm_endpoint = get_llm_endpoint()
         health_status = self.check_health()
         if not health_status:
-            logger.error("OPEAFAQGen health check failed.")
+            logger.error("OpeaFaqGen health check failed.")
 
     async def generate(self, input: LLMParamsDoc, client):
         """Invokes the TGI/vLLM LLM service to generate FAQ output for the provided input.

comps/llms/src/faq-generation/integrations/tgi.py

@@ -14,9 +14,9 @@
 logflag = os.getenv("LOGFLAG", False)
 
 
-@OpeaComponentRegistry.register("OPEAFAQGen_TGI")
-class OPEAFAQGen_TGI(OPEAFAQGen):
-    """A specialized OPEA FAQGen TGI component derived from OPEAFAQGen for interacting with TGI services based on Lanchain HuggingFaceEndpoint API.
+@OpeaComponentRegistry.register("OpeaFaqGenTgi")
+class OpeaFaqGenTgi(OpeaFaqGen):
+    """A specialized OPEA FAQGen TGI component derived from OpeaFaqGen for interacting with TGI services based on Lanchain HuggingFaceEndpoint API.
 
     Attributes:
         client (TGI): An instance of the TGI client for text generation.

comps/llms/src/faq-generation/integrations/vllm.py

@@ -14,9 +14,9 @@
 logflag = os.getenv("LOGFLAG", False)
 
 
-@OpeaComponentRegistry.register("OPEAFAQGen_vLLM")
-class OPEAFAQGen_vLLM(OPEAFAQGen):
-    """A specialized OPEA FAQGen vLLM component derived from OPEAFAQGen for interacting with vLLM services based on Lanchain VLLMOpenAI API.
+@OpeaComponentRegistry.register("OpeaFaqGenvLLM")
+class OpeaFaqGenvLLM(OpeaFaqGen):
+    """A specialized OPEA FAQGen vLLM component derived from OpeaFaqGen for interacting with vLLM services based on Lanchain VLLMOpenAI API.
 
     Attributes:
         client (vLLM): An instance of the vLLM client for text generation.

comps/llms/src/faq-generation/opea_faqgen_microservice.py

@@ -4,8 +4,8 @@
 import os
 import time
 
-from integrations.tgi import OPEAFAQGen_TGI
-from integrations.vllm import OPEAFAQGen_vLLM
+from integrations.tgi import OpeaFaqGenTgi
+from integrations.vllm import OpeaFaqGenvLLM
 
 from comps import (
     CustomLogger,
@@ -21,7 +21,7 @@
 logger = CustomLogger("llm_faqgen")
 logflag = os.getenv("LOGFLAG", False)
 
-llm_component_name = os.getenv("FAQGen_COMPONENT_NAME", "OPEAFAQGen_TGI")
+llm_component_name = os.getenv("FAQGen_COMPONENT_NAME", "OpeaFaqGenTgi")
 # Initialize OpeaComponentLoader
 loader = OpeaComponentLoader(llm_component_name, description=f"OPEA LLM FAQGen Component: {llm_component_name}")
 

comps/llms/src/text-generation/integrations/native.py

@@ -200,18 +200,18 @@ def initialize():
             initialized = True
 
 
-@OpeaComponentRegistry.register("OPEATextGen_Native")
-class OPEATextGen_Native(OpeaComponent):
+@OpeaComponentRegistry.register("OpeaTextGenNative")
+class OpeaTextGenNative(OpeaComponent):
     """A specialized OPEA TextGen component derived from OpeaComponent for interacting with LLM services based on native optimum habana."""
 
     def __init__(self, name: str, description: str, config: dict = None):
         super().__init__(name, ServiceType.LLM.name.lower(), description, config)
         initialize()
         health_status = self.check_health()
         if not health_status:
-            logger.error("OPEATextGen_Native health check failed.")
+            logger.error("OpeaTextGenNative health check failed.")
         else:
-            logger.info("OPEATextGen_Native health check success.")
+            logger.info("OpeaTextGenNative health check success.")
 
     def check_health(self) -> bool:
         """Checks the health of the LLM service.

comps/llms/src/text-generation/integrations/predictionguard.py

@@ -15,8 +15,8 @@
 logflag = os.getenv("LOGFLAG", False)
 
 
-@OpeaComponentRegistry.register("OPEATextGen_Predictionguard")
-class OPEATextGen_Predictionguard(OpeaComponent):
+@OpeaComponentRegistry.register("OpeaTextGenPredictionguard")
+class OpeaTextGenPredictionguard(OpeaComponent):
     """A specialized OPEA TextGen component derived from OpeaComponent for interacting with Predictionguard services.
 
     Attributes:
@@ -28,9 +28,9 @@ def __init__(self, name: str, description: str, config: dict = None):
         self.client = PredictionGuard()
         health_status = self.check_health()
         if not health_status:
-            logger.error("OPEATextGen_Predictionguard health check failed.")
+            logger.error("OpeaTextGenPredictionguard health check failed.")
         else:
-            logger.info("OPEATextGen_Predictionguard health check success.")
+            logger.info("OpeaTextGenPredictionguard health check success.")
 
     def check_health(self) -> bool:
         """Checks the health of the Predictionguard LLM service.

comps/llms/src/text-generation/integrations/opea.py → comps/llms/src/text-generation/integrations/service.py

@@ -47,8 +47,8 @@ def get_llm_endpoint():
         raise ConfigError(f"Input model {MODEL_NAME} not present in model_configs")
 
 
-@OpeaComponentRegistry.register("OPEA_LLM")
-class OPEALLM(OpeaComponent):
+@OpeaComponentRegistry.register("OpeaTextGenService")
+class OpeaTextGenService(OpeaComponent):
     """A specialized OPEA LLM component derived from OpeaComponent for interacting with TGI/vLLM services based on OpenAI API.
 
     Attributes:
@@ -60,7 +60,7 @@ def __init__(self, name: str, description: str, config: dict = None):
         self.client = self._initialize_client()
         health_status = self.check_health()
         if not health_status:
-            logger.error("OPEALLM health check failed.")
+            logger.error("OpeaTextGenService health check failed.")
 
     def _initialize_client(self) -> AsyncOpenAI:
         """Initializes the AsyncOpenAI."""

comps/llms/src/text-generation/opea_llm_microservice.py

@@ -21,15 +21,15 @@
 logger = CustomLogger("llm")
 logflag = os.getenv("LOGFLAG", False)
 
-llm_component_name = os.getenv("LLM_COMPONENT_NAME", "OPEA_LLM")
+llm_component_name = os.getenv("LLM_COMPONENT_NAME", "OpeaTextGenService")
 if logflag:
     logger.info(f"Get llm_component_name {llm_component_name}")
 
-if llm_component_name == "OPEATextGen_Native":
-    from integrations.native import OPEATextGen_Native
+if llm_component_name == "OpeaTextGenNative":
+    from integrations.native import OpeaTextGenNative
 else:
-    from integrations.opea import OPEALLM
-    from integrations.predictionguard import OPEATextGen_Predictionguard
+    from integrations.predictionguard import OpeaTextGenPredictionguard
+    from integrations.service import OpeaTextGenService
 
 # Initialize OpeaComponentLoader
 loader = OpeaComponentLoader(llm_component_name, description=f"OPEA LLM Component: {llm_component_name}")

tests/llms/test_llms_doc-summarization_langchain_tgi.sh → tests/llms/test_llms_doc-summarization_tgi.sh

@@ -28,7 +28,7 @@ function start_service() {
     export LLM_MODEL_ID="Intel/neural-chat-7b-v3-3"
     export MAX_INPUT_TOKENS=2048
     export MAX_TOTAL_TOKENS=4096
-    export DocSum_COMPONENT_NAME="OPEADocSum_TGI" # or "vllm"
+    export DocSum_COMPONENT_NAME="OpeaDocSumTgi" # or "vllm"
     export LOGFLAG=True
 
     cd $WORKPATH/comps/llms/deployment/docker_compose

tests/llms/test_llms_doc-summarization_langchain_tgi_on_intel_hpu.sh → tests/llms/test_llms_doc-summarization_tgi_on_intel_hpu.sh

@@ -28,7 +28,7 @@ function start_service() {
     export LLM_MODEL_ID="Intel/neural-chat-7b-v3-3"
     export MAX_INPUT_TOKENS=2048
     export MAX_TOTAL_TOKENS=4096
-    export DocSum_COMPONENT_NAME="OPEADocSum_TGI" # or "vllm"
+    export DocSum_COMPONENT_NAME="OpeaDocSumTgi" # or "vllm"
     export LOGFLAG=True
 
     cd $WORKPATH/comps/llms/deployment/docker_compose

tests/llms/test_llms_doc-summarization_langchain_vllm_on_intel_hpu.sh → tests/llms/test_llms_doc-summarization_vllm_on_intel_hpu.sh

@@ -40,7 +40,7 @@ function start_service() {
     export LLM_MODEL_ID="Intel/neural-chat-7b-v3-3"
     export MAX_INPUT_TOKENS=2048
     export MAX_TOTAL_TOKENS=4096
-    export DocSum_COMPONENT_NAME="OPEADocSum_vLLM" # or "vllm"
+    export DocSum_COMPONENT_NAME="OpeaDocSumvLLM" # or "vllm"
     export VLLM_SKIP_WARMUP=true
     export LOGFLAG=True
 

tests/llms/test_llms_faq-generation_langchain_tgi.sh → tests/llms/test_llms_faq-generation_tgi.sh

@@ -26,7 +26,7 @@ function start_service() {
     export HUGGINGFACEHUB_API_TOKEN=${HF_TOKEN} # Remember to set HF_TOKEN before invoking this test!
     export LLM_ENDPOINT="http://${host_ip}:${LLM_ENDPOINT_PORT}"
     export LLM_MODEL_ID="Intel/neural-chat-7b-v3-3"
-    export FAQGen_COMPONENT_NAME="OPEAFAQGen_TGI"
+    export FAQGen_COMPONENT_NAME="OpeaFaqGenTgi"
     export LOGFLAG=True
 
     cd $WORKPATH/comps/llms/deployment/docker_compose

tests/llms/test_llms_faq-generation_langchain_tgi_on_intel_hpu.sh → tests/llms/test_llms_faq-generation_tgi_on_intel_hpu.sh

@@ -26,7 +26,7 @@ function start_service() {
     export HUGGINGFACEHUB_API_TOKEN=${HF_TOKEN} # Remember to set HF_TOKEN before invoking this test!
     export LLM_ENDPOINT="http://${host_ip}:${LLM_ENDPOINT_PORT}"
     export LLM_MODEL_ID="Intel/neural-chat-7b-v3-3"
-    export FAQGen_COMPONENT_NAME="OPEAFAQGen_TGI"
+    export FAQGen_COMPONENT_NAME="OpeaFaqGenTgi"
     export LOGFLAG=True
 
     cd $WORKPATH/comps/llms/deployment/docker_compose

tests/llms/test_llms_faq-generation_langchain_vllm_on_intel_hpu.sh → tests/llms/test_llms_faq-generation_vllm_on_intel_hpu.sh

@@ -39,7 +39,7 @@ function start_service() {
     export HUGGINGFACEHUB_API_TOKEN=${HF_TOKEN} # Remember to set HF_TOKEN before invoking this test!
     export LLM_ENDPOINT="http://${host_ip}:${LLM_ENDPOINT_PORT}"
     export LLM_MODEL_ID="Intel/neural-chat-7b-v3-3"
-    export FAQGen_COMPONENT_NAME="OPEAFAQGen_vLLM"
+    export FAQGen_COMPONENT_NAME="OpeaFaqGenvLLM"
     export VLLM_SKIP_WARMUP=true
     export LOGFLAG=True
 

tests/llms/test_llms_text-generation_native_on_intel_hpu.sh

@@ -38,7 +38,7 @@ function start_service() {
         -e HABANA_VISIBLE_DEVICES=all \
         -e OMPI_MCA_btl_vader_single_copy_mechanism=none \
         -e TOKENIZERS_PARALLELISM=false \
-        -e LLM_COMPONENT_NAME="OPEATextGen_Native" \
+        -e LLM_COMPONENT_NAME="OpeaTextGenNative" \
         --restart unless-stopped \
         --network bridge \
         opea/llm-textgen-gaudi:comps

tests/llms/test_llms_text-generation_predictionguard.sh

@@ -28,7 +28,7 @@ function start_service() {
     docker run -d --name=test-comps-llm-textgen-pg-server \
         -e http_proxy= -e https_proxy= \
         -e PREDICTIONGUARD_API_KEY=${PREDICTIONGUARD_API_KEY} \
-        -e LLM_COMPONENT_NAME="OPEATextGen_Predictionguard" \
+        -e LLM_COMPONENT_NAME="OpeaTextGenPredictionguard" \
         -e LOGFLAG=True \
         -p 9000:9000 --ipc=host opea/llm-textgen:comps
     sleep 60  # Sleep for 1 minute to allow the service to start