Aim for No. 1 RAG

Deploy

• v1.0.3: 東京/大阪/シカゴリージョンのみをサポートしています。（デフォルト：大阪リージョン） Click

Setup

conda create -n no.1-rag python=3.11 -y
conda activate no.1-rag

pip install -r requirements.txt -i https://pypi.tuna.tsinghua.edu.cn/simple
# pip list --format=freeze > requirements.txt

# Test cuda
python -c "import torch;print(torch.cuda.is_available())"

# Expected
True

Full Text Search

BEGIN
  CTX_DDL.CREATE_PREFERENCE('world_lexer','WORLD_LEXER');
END;
/

CREATE INDEX embed_data_idx ON langchain_oracle_embedding(embed_data) INDEXTYPE IS CTXSYS.CONTEXT PARAMETERS ('LEXER world_lexer sync (on commit)');

Using Unstructured Open Source

sudo apt install build-essential libmagic-dev poppler-utils tesseract-ocr libreoffice pandoc -y

Generate LangFuse Secret and Salt

echo <your_secret> | openssl rand -base64 32
echo <your_salt> | openssl rand -base64 32

Change Region Image ID

https://docs.oracle.com/en-us/iaas/images/image/50cf60da-4374-44e2-ab38-70185991f833/index.htm