Grag ex1.1 (#1123)

Signed-off-by: Rita Brugarolas <[email protected]> Signed-off-by: theresa <[email protected]> Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com> Co-authored-by: theresa <[email protected]>
opea-project · Nov 15, 2024 · 6f8fa6a · 6f8fa6a
1 parent 39f68d5
commit 6f8fa6a
Show file tree

Hide file tree

Showing 8 changed files with 29 additions and 16 deletions.
diff --git a/GraphRAG/README.md b/GraphRAG/README.md
@@ -53,7 +53,7 @@ To set up environment variables for deploying GraphRAG services, follow these st
 
 ### Quick Start: 2.Run Docker Compose
 
-If the microservice images are available in Docker Hub they will be pulled, otherwise you will need to build the container images manually. Please refer to the 'Build Docker Images' in [Guide](../ChatQnA/docker_compose/intel/cpu/xeon/README.md). [test_compose.sh](tests/test_compose.sh) can be a good resource as it shows how to do image build, starting services, validated each microservices and megaservices. This is what is used in CI/CD.
+If the microservice images are available in Docker Hub they will be pulled, otherwise you will need to build the container images manually. Please refer to the 'Build Docker Images' in [Guide](../ChatQnA/docker_compose/intel/cpu/xeon/README.md). [test_compose_on_gaudi.sh](tests/test_compose_on_gaudi.sh) can be a good resource as it shows how to do image build, starting services, validated each microservices and megaservices. This is what is used in CI/CD.
 
 Docker compose will start 8 services: ![8 servicesi in GraphRAG](assets/8microservices.png)
 
@@ -191,7 +191,7 @@ Working on a table comparison of various model sizes vs. naive RAG with a datase
 
 ### Check Service Status
 
-Before consuming GraphRAG Service, make sure each microservice is ready by checking the docker logs of each microservice. [test_compose.sh](tests/test_compose.sh) can be a good resource as it shows how CI/CD validated each microservices based on returned HTTP status and response body.
+Before consuming GraphRAG Service, make sure each microservice is ready by checking the docker logs of each microservice. [test_compose_on_gaudi.sh](tests/test_compose_on_gaudi.sh) can be a good resource as it shows how CI/CD validated each microservices based on returned HTTP status and response body.
 
 ```bash
 docker logs container_name

diff --git a/GraphRAG/docker_compose/intel/hpu/gaudi/compose.yaml b/GraphRAG/docker_compose/intel/hpu/gaudi/compose.yaml
@@ -1,7 +1,6 @@
 # Copyright (C) 2024 Intel Corporation
 # SPDX-License-Identifier: Apache-2.0
 
-version: "3.8"
 services:
   neo4j-apoc:
     image: neo4j:latest
@@ -52,6 +51,7 @@ services:
       http_proxy: ${http_proxy}
       https_proxy: ${https_proxy}
       HUGGING_FACE_HUB_TOKEN: ${HUGGINGFACEHUB_API_TOKEN}
+      HF_TOKEN: ${HF_TOKEN}
       HF_HUB_DISABLE_PROGRESS_BARS: 1
       HF_HUB_ENABLE_HF_TRANSFER: 0
       HABANA_VISIBLE_DEVICES: all
@@ -64,9 +64,9 @@ services:
     cap_add:
       - SYS_NICE
     ipc: host
-    command: --model-id ${LLM_MODEL_ID} --max-input-length 2048 --max-total-tokens 4096
+    command: --model-id ${LLM_MODEL_ID} --max-input-length 6000 --max-total-tokens 8192
   dataprep-neo4j-llamaindex:
-    image: opea/dataprep-neo4j-llamaindex:latest
+    image: ${REGISTRY:-opea}/dataprep-neo4j-llamaindex:${TAG:-latest}
     container_name: dataprep-neo4j-server
     depends_on:
       - neo4j-apoc
@@ -80,6 +80,8 @@ services:
       http_proxy: ${http_proxy}
       https_proxy: ${https_proxy}
       host_ip: ${host_ip}
+      HUGGING_FACE_HUB_TOKEN: ${HUGGINGFACEHUB_API_TOKEN}
+      HF_TOKEN: ${HF_TOKEN}
       NEO4J_URL: ${NEO4J_URL}
       NEO4J_USERNAME: ${NEO4J_USERNAME}
       NEO4J_PASSWORD: ${NEO4J_PASSWORD}
@@ -93,7 +95,7 @@ services:
       LOGFLAG: ${LOGFLAG}
     restart: unless-stopped
   retriever-neo4j-llamaindex:
-    image: opea/retriever-neo4j-llamaindex:latest
+    image: ${REGISTRY:-opea}/retriever-neo4j-llamaindex:${TAG:-latest}
     container_name: retriever-neo4j-server
     depends_on:
       - neo4j-apoc
@@ -107,6 +109,8 @@ services:
       http_proxy: ${http_proxy}
       https_proxy: ${https_proxy}
       host_ip: ${host_ip}
+      HUGGING_FACE_HUB_TOKEN: ${HUGGINGFACEHUB_API_TOKEN}
+      HF_TOKEN: ${HF_TOKEN}
       NEO4J_URL: ${NEO4J_URL}
       NEO4J_USERNAME: ${NEO4J_USERNAME}
       NEO4J_PASSWORD: ${NEO4J_PASSWORD}
@@ -136,7 +140,6 @@ services:
       - MEGA_SERVICE_HOST_IP=graphrag-gaudi-backend-server
       - RETRIEVER_SERVICE_HOST_IP=retriever-neo4j-llamaindex
       - RETRIEVER_SERVICE_PORT=6009
-      - RETRIEVER_SERVICE_PORT=${RETRIEVER_SERVICE_PORT:-80}
       - LLM_SERVER_HOST_IP=tgi-gaudi-service
       - LLM_SERVER_PORT=${LLM_SERVER_PORT:-80}
       - LOGFLAG=${LOGFLAG}

diff --git a/GraphRAG/docker_compose/intel/hpu/gaudi/set_env.sh b/GraphRAG/docker_compose/intel/hpu/gaudi/set_env.sh
@@ -17,4 +17,4 @@ export NEO4J_URL="bolt://${host_ip}:7687"
 export NEO4J_USERNAME=neo4j
 export DATAPREP_SERVICE_ENDPOINT="http://${host_ip}:6004/v1/dataprep"
 export LOGFLAG=True
-export RETRIEVER_SERVICE_PORT=6009
+export RETRIEVER_SERVICE_PORT=80
diff --git a/GraphRAG/graphrag.py b/GraphRAG/graphrag.py
@@ -70,7 +70,11 @@ def align_outputs(self, data, cur_node, inputs, runtime_graph, llm_parameters_di
         # handle template
         # if user provides template, then format the prompt with it
         # otherwise, use the default template
-        prompt = inputs.messages[0]["content"]
+        print("outputs before align:\n", inputs)
+        if isinstance(inputs.messages, str):
+            prompt = inputs.messages
+        else:
+            prompt = inputs.messages[0]["content"]
         chat_template = llm_parameters_dict["chat_template"]
         if chat_template:
             prompt_template = PromptTemplate.from_template(chat_template)

diff --git a/GraphRAG/tests/test_compose.sh → GraphRAG/tests/test_compose_on_gaudi.sh b/GraphRAG/tests/test_compose.sh → GraphRAG/tests/test_compose_on_gaudi.sh
@@ -2,7 +2,7 @@
 # Copyright (C) 2024 Intel Corporation
 # SPDX-License-Identifier: Apache-2.0
 
-set -e
+set -xe
 IMAGE_REPO=${IMAGE_REPO:-"opea"}
 IMAGE_TAG=${IMAGE_TAG:-"latest"}
 echo "REGISTRY=IMAGE_REPO=${IMAGE_REPO}"
@@ -11,7 +11,6 @@ export REGISTRY=${IMAGE_REPO}
 export TAG=${IMAGE_TAG}
 
 WORKPATH=$(dirname "$PWD")
-WORKPATH=/home/rbrugaro/GenAIExamples/GraphRAG
 LOG_PATH="$WORKPATH/tests"
 ip_address=$(hostname -I | awk '{print $1}')
 
@@ -34,6 +33,13 @@ function start_services() {
     export EMBEDDING_MODEL_ID="BAAI/bge-base-en-v1.5"
     export LLM_MODEL_ID="meta-llama/Meta-Llama-3-8B-Instruct"
     export HUGGINGFACEHUB_API_TOKEN=${HUGGINGFACEHUB_API_TOKEN}
+    export HF_TOKEN=${HUGGINGFACEHUB_API_TOKEN}
+    export NEO4J_USERNAME="neo4j"
+    export NEO4J_PASSWORD="neo4jtest"
+    export NEO4J_URL="bolt://${ip_address}:7687"
+    export TEI_EMBEDDING_ENDPOINT="http://${ip_address}:6006"
+    export TGI_LLM_ENDPOINT="http://${ip_address}:6005"
+    export host_ip=${ip_address}
 
     # Start Docker Containers
     docker compose -f compose.yaml up -d > ${LOG_PATH}/start_services_with_compose.log

diff --git a/GraphRAG/ui/svelte/.env b/GraphRAG/ui/svelte/.env
@@ -1,4 +1,4 @@
-CHAT_BASE_URL = '/v1/chatqna'
+CHAT_BASE_URL = '/v1/graphrag'
 
 UPLOAD_FILE_BASE_URL = '/v1/dataprep'
 

diff --git a/GraphRAG/ui/svelte/playwright.config.ts b/GraphRAG/ui/svelte/playwright.config.ts
@@ -15,13 +15,13 @@ import { defineConfig, devices } from "@playwright/test";
 export default defineConfig({
 	testDir: "./tests",
 	/* Maximum time one test can run for. */
-	timeout: 30 * 1000,
+	timeout: 300 * 1000,
 	expect: {
 		/**
 		 * Maximum time expect() should wait for the condition to be met.
 		 * For example in `await expect(locator).toHaveText();`
 		 */
-		timeout: 5000,
+		timeout: 30000,
 	},
 	/* Run tests in files in parallel */
 	fullyParallel: true,
@@ -38,7 +38,7 @@ export default defineConfig({
 		/* Maximum time each action such as `click()` can take. Defaults to 0 (no limit). */
 		actionTimeout: 0,
 		/* Base URL to use in actions like `await page.goto('/')`. */
-		baseURL: "http://10.7.4.57:80",
+		baseURL: "http://localhost:80",
 
 		/* Collect trace when retrying the failed test. See https://playwright.dev/docs/trace-viewer */
 		trace: "on-first-retry",

diff --git a/GraphRAG/ui/svelte/src/routes/+page.svelte b/GraphRAG/ui/svelte/src/routes/+page.svelte
@@ -197,7 +197,7 @@
 >
 	<div class="mx-auto flex h-full w-full flex-col sm:mt-0 sm:w-[72%]">
 		<div class="flex justify-between p-2">
-			<p class="text-[1.7rem] font-bold tracking-tight">ChatQnA</p>
+			<p class="text-[1.7rem] font-bold tracking-tight">GraphRAG</p>
 			<UploadFile />
 		</div>
 		<div