Adding git to Dockerfile, tested dockerfile and dockercompose. Also p…

…arametrized variables in prepare_doc_arango.py

comps/dataprep/arango/langchain/Dockerfile

@@ -11,7 +11,8 @@ RUN apt-get update -y && apt-get install -y --no-install-recommends --fix-missin
     build-essential \
     default-jre \
     libgl1-mesa-glx \
-    libjemalloc-dev
+    libjemalloc-dev \
+    git
 
 RUN useradd -m -s /bin/bash user && \
     mkdir -p /home/user && \
@@ -35,4 +36,4 @@ USER user
 
 WORKDIR /home/user/comps/dataprep/arango/langchain
 
-ENTRYPOINT ["python", "prepare_doc_arango.py"]
+ENTRYPOINT ["python", "prepare_doc_arango.py"]

comps/dataprep/arango/langchain/README.md

@@ -89,7 +89,7 @@ curl -X POST \
     http://localhost:6007/v1/dataprep
 ```
 
-Additional options that can be specified from the config.py file are as follows:
+Additional options that can be specified from the environment variables are as follows (default values are in the config.py file):
 
 OpenAI Configuration:
 - `OPENAI_EMBED_MODEL`: The embedding model to use for the OpenAI service.

comps/dataprep/arango/langchain/config.py

@@ -10,10 +10,10 @@
 ARANGO_DB_NAME = os.getenv("ARANGO_DB_NAME", "_system")
 
 # ArangoDB graph configuration
-USE_ONE_ENTITY_COLLECTION = True
-INSERT_ASYNC = False
-ARANGO_BATCH_SIZE = 1000
-INCLUDE_SOURCE = True
+USE_ONE_ENTITY_COLLECTION = os.getenv("USE_ONE_ENTITY_COLLECTION", True)
+INSERT_ASYNC = os.getenv("INSERT_ASYNC", False)
+ARANGO_BATCH_SIZE = os.getenv("ARANGO_BATCH_SIZE", 1000)
+INCLUDE_SOURCE = os.getenv("INCLUDE_SOURCE", True)
 
 # Text Generation Inference configuration
 TGI_LLM_ENDPOINT = os.getenv("TGI_LLM_ENDPOINT", "http://localhost:8080")
@@ -27,11 +27,13 @@
 OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")
 OPENAI_EMBED_MODEL = os.getenv("OPENAI_EMBED_MODEL", "text-embedding-3-small")
 OPENAI_EMBED_DIMENSIONS = os.getenv("OPENAI_EMBED_DIMENSIONS", 512)
+OPENAI_CHAT_MODEL = os.getenv("OPENAI_CHAT_MODEL", "gpt-4o")
+OPENAI_CHAT_TEMPERATURE = os.getenv("OPENAI_CHAT_TEMPERATURE", 0)
 
 # LLMGraphTransformer configuration
-ALLOWED_NODES = [] # ["Person", "Organization"]
-ALLOWED_RELATIONSHIPS = [] # [("Person", "knows", "Person"), ("Person", "works_at", "Organization")]
-NODE_PROPERTIES = False # ["description"]
-RELATIONSHIP_PROPERTIES = False # ["description"]
+ALLOWED_NODES = os.getenv("ALLOWED_NODES", []) # ["Person", "Organization"]
+ALLOWED_RELATIONSHIPS = os.getenv("ALLOWED_RELATIONSHIPS", []) # [("Person", "knows", "Person"), ("Person", "works_at", "Organization")]
+NODE_PROPERTIES = os.getenv("NODE_PROPERTIES", ['description'])
+RELATIONSHIP_PROPERTIES = os.getenv("RELATIONSHIP_PROPERTIES", ['description'])
 
 SYSTEM_PROMPT_PATH = os.getenv("SYSTEM_PROMPT_PATH", "./prompt.txt")

comps/dataprep/arango/langchain/docker-compose-dataprep-arango.yaml

@@ -8,6 +8,8 @@ services:
     container_name: arango-graph-db
     ports:
       - "8529:8529"
+    environment:
+      ARANGO_ROOT_PASSWORD: ${ARANGO_PASSWORD}
   tgi_gaudi_service:
     image: ghcr.io/huggingface/tgi-gaudi:2.0.5
     container_name: tgi-service
@@ -23,7 +25,7 @@ services:
       HF_TOKEN: ${HF_TOKEN}
     command: --model-id ${LLM_MODEL_ID} --auto-truncate --max-input-tokens 1024 --max-total-tokens 2048
   dataprep-arango:
-    image: opea/gen-ai-comps:dataprep-arango-xeon-server
+    image: opea/dataprep-arango:latest
     container_name: dataprep-arango-server
     depends_on:
       - arango-vector-db
@@ -35,7 +37,7 @@ services:
       no_proxy: ${no_proxy}
       http_proxy: ${http_proxy}
       https_proxy: ${https_proxy}
-      ARANGO_URL: ${ARANGO_URL}
+      ARANGO_URL: http://arango-graph-db:8529
       ARANGO_USERNAME: ${ARANGO_USERNAME}
       ARANGO_PASSWORD: ${ARANGO_PASSWORD}
       ARANGO_DB_NAME: ${ARANGO_DB_NAME}