Align OpenAI API for FaqGen, DocSum

Algin all the inputs to OpenAI API format for FaqGen, DocSum, related with GenAIComps PR opea-project/GenAIComps#1161

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

.github/workflows/_run-docker-compose.yml

@@ -134,6 +134,7 @@ jobs:
           SERVING_TOKEN: ${{ secrets.SERVING_TOKEN }}
           IMAGE_REPO: ${{ inputs.registry }}
           IMAGE_TAG: ${{ inputs.tag }}
+          opea_branch: "llm_openai_api"
           example: ${{ inputs.example }}
           hardware: ${{ inputs.hardware }}
           test_case: ${{ matrix.test_case }}

.github/workflows/pr-docker-compose-e2e.yml

@@ -4,7 +4,7 @@
 name: E2E test with docker compose
 
 on:
-  pull_request_target:
+  pull_request:
     branches: ["main", "*rc"]
     types: [opened, reopened, ready_for_review, synchronize] # added `ready_for_review` since draft is skipped
     paths:

.github/workflows/pr-dockerfile-path-and-build-yaml-scan.yml

@@ -22,6 +22,7 @@ jobs:
         run: |
           cd ..
           git clone https://github.com/opea-project/GenAIComps.git
+          cd GenAIComps && git checkout llm_openai_api
 
       - name: Check for Missing Dockerfile Paths in GenAIComps
         run: |

DocSum/docsum.py

@@ -17,7 +17,7 @@
     ChatMessage,
     UsageInfo,
 )
-from comps.cores.proto.docarray import DocSumLLMParams
+from comps.cores.proto.api_protocol import DocSumChatCompletionRequest
 from fastapi import File, Request, UploadFile
 from fastapi.responses import StreamingResponse
 
@@ -34,13 +34,13 @@ def align_inputs(self, inputs, cur_node, runtime_graph, llm_parameters_dict, **k
     if self.services[cur_node].service_type == ServiceType.LLM:
         for key_to_replace in ["text", "asr_result"]:
             if key_to_replace in inputs:
-                inputs["query"] = inputs[key_to_replace]
+                inputs["messages"] = inputs[key_to_replace]
                 del inputs[key_to_replace]
 
         docsum_parameters = kwargs.get("docsum_parameters", None)
         if docsum_parameters:
             docsum_parameters = docsum_parameters.model_dump()
-            del docsum_parameters["query"]
+            del docsum_parameters["messages"]
             inputs.update(docsum_parameters)
     elif self.services[cur_node].service_type == ServiceType.ASR:
         if "video" in inputs:
@@ -217,13 +217,13 @@ async def handle_request(self, request: Request, files: List[UploadFile] = File(
                 initial_inputs_data = {}
                 initial_inputs_data[data_type] = prompt
             else:
-                initial_inputs_data = {"query": prompt}
+                initial_inputs_data = {"messages": prompt}
 
         else:
             raise ValueError(f"Unknown request type: {request.headers.get('content-type')}")
 
-        docsum_parameters = DocSumLLMParams(
-            query="",
+        docsum_parameters = DocSumChatCompletionRequest(
+            messages="",
             max_tokens=chat_request.max_tokens if chat_request.max_tokens else 1024,
             top_k=chat_request.top_k if chat_request.top_k else 10,
             top_p=chat_request.top_p if chat_request.top_p else 0.95,

DocSum/tests/test_compose_on_gaudi.sh

@@ -161,7 +161,7 @@ function validate_microservices() {
         "data: " \
         "llm-docsum-tgi" \
         "llm-docsum-gaudi-server" \
-        '{"query":"Text Embeddings Inference (TEI) is a toolkit for deploying and serving open source text embeddings and sequence classification models. TEI enables high-performance extraction for the most popular models, including FlagEmbedding, Ember, GTE and E5."}'
+        '{"messages":"Text Embeddings Inference (TEI) is a toolkit for deploying and serving open source text embeddings and sequence classification models. TEI enables high-performance extraction for the most popular models, including FlagEmbedding, Ember, GTE and E5."}'
 
     # whisper microservice
     ulimit -s 65536

DocSum/tests/test_compose_on_rocm.sh

@@ -141,7 +141,7 @@ function validate_microservices() {
         "data: " \
         "docsum-llm-server" \
         "docsum-llm-server" \
-        '{"query":"Text Embeddings Inference (TEI) is a toolkit for deploying and serving open source text embeddings and sequence classification models. TEI enables high-performance extraction for the most popular models, including FlagEmbedding, Ember, GTE and E5."}'
+        '{"messages":"Text Embeddings Inference (TEI) is a toolkit for deploying and serving open source text embeddings and sequence classification models. TEI enables high-performance extraction for the most popular models, including FlagEmbedding, Ember, GTE and E5."}'
 
 }
 

DocSum/tests/test_compose_on_xeon.sh

@@ -163,7 +163,7 @@ function validate_microservices() {
         "data: " \
         "llm-docsum-tgi" \
         "llm-docsum-server" \
-        '{"query":"Text Embeddings Inference (TEI) is a toolkit for deploying and serving open source text embeddings and sequence classification models. TEI enables high-performance extraction for the most popular models, including FlagEmbedding, Ember, GTE and E5."}'
+        '{"messages":"Text Embeddings Inference (TEI) is a toolkit for deploying and serving open source text embeddings and sequence classification models. TEI enables high-performance extraction for the most popular models, including FlagEmbedding, Ember, GTE and E5."}'
 
     # whisper microservice
     ulimit -s 65536

FaqGen/faqgen.py

@@ -113,7 +113,7 @@ async def handle_request(self, request: Request, files: List[UploadFile] = File(
             model=chat_request.model if chat_request.model else None,
         )
         result_dict, runtime_graph = await self.megaservice.schedule(
-            initial_inputs={"query": prompt}, llm_parameters=parameters
+            initial_inputs={"messages": prompt}, llm_parameters=parameters
         )
         for node, response in result_dict.items():
             # Here it suppose the last microservice in the megaservice is LLM.

FaqGen/tests/test_compose_on_gaudi.sh

@@ -94,7 +94,7 @@ function validate_microservices() {
         "data: " \
         "llm" \
         "llm-faqgen-server" \
-        '{"query":"Text Embeddings Inference (TEI) is a toolkit for deploying and serving open source text embeddings and sequence classification models. TEI enables high-performance extraction for the most popular models, including FlagEmbedding, Ember, GTE and E5."}'
+        '{"messages":"Text Embeddings Inference (TEI) is a toolkit for deploying and serving open source text embeddings and sequence classification models. TEI enables high-performance extraction for the most popular models, including FlagEmbedding, Ember, GTE and E5."}'
 }
 
 function validate_megaservice() {

FaqGen/tests/test_compose_on_rocm.sh

@@ -98,7 +98,7 @@ function validate_microservices() {
         "data: " \
         "llm" \
         "faqgen-llm-server" \
-        '{"query":"Text Embeddings Inference (TEI) is a toolkit for deploying and serving open source text embeddings and sequence classification models. TEI enables high-performance extraction for the most popular models, including FlagEmbedding, Ember, GTE and E5."}'
+        '{"messages":"Text Embeddings Inference (TEI) is a toolkit for deploying and serving open source text embeddings and sequence classification models. TEI enables high-performance extraction for the most popular models, including FlagEmbedding, Ember, GTE and E5."}'
 }
 
 function validate_megaservice() {

FaqGen/tests/test_compose_on_xeon.sh

@@ -94,7 +94,7 @@ function validate_microservices() {
         "data: " \
         "llm" \
         "llm-faqgen-server" \
-        '{"query":"Text Embeddings Inference (TEI) is a toolkit for deploying and serving open source text embeddings and sequence classification models. TEI enables high-performance extraction for the most popular models, including FlagEmbedding, Ember, GTE and E5."}'
+        '{"messages":"Text Embeddings Inference (TEI) is a toolkit for deploying and serving open source text embeddings and sequence classification models. TEI enables high-performance extraction for the most popular models, including FlagEmbedding, Ember, GTE and E5."}'
 }
 
 function validate_megaservice() {