a RAG AI blog/email writer with Dolt as an optional vector store.
... more to come
cd robot-blogger
go install .
Required Flags
-host
, the vector store host to connect to.-port
, the vector store port to connect to.-user
, the vector store user to connect to.-model
, the LLM model to use.-store-name
, the name of the vector store to use.
One of:
-dolt
, uses dolt as vector store.-mariadb
, uses mariadb as vector store.-postgres
, uses postgres as vector store.
One of:
-ollama
, uses ollama llm runner.OLLAMA_HOST
environment variable must be set if not running on localhost.-openai
, uses openai llm runner.OPENAI_API_KEY
environment variable must be set.
Required for Store
DOCS_DIR
environment variable must be set. Specifies path to directory containing the docs to store.-doc-type
, the type of document you are storing.-include-file-ext
, the file extension of the docs to store.
Required for Generate
-prompt-file
, the path to a file containing the prompt to run.-topic
, the topic of the content to generate.-length
, the length of the content to generate.
Optional Flags
-output-format
, the format of the content to generate.-vector-dimensions
, the number of dimensions to use for the vector store. Required for MariaDB.
You must have a vector store running with the store name/database name already created. You may also need to pull the model you are trying to use, ie:
ollama pull llama3
export VECTOR_STORE_PASSWORD=mydbpass
export DOCS_DIR=/path/to/docs
./robot-blogger \
--ollama \
--model=llama3 \
--dolt \
--user=root \
--host=0.0.0.0 \
--port=3306 \
--store-name=robot_blogger_llama3_v1 \
--doc-type=blog_post \
--include-file-ext=".md"
export VECTOR_STORE_PASSWORD=mydbpass
./robot-blogger \
--ollama \
--model=llama3 \
--dolt \
--user=root \
--host=0.0.0.0 \
--port=3306 \
--topic="DoltHub Products" \
--length=100 \
--store-name=robot_blogger_llama3_v1 \
--prompt-file=/path/to/file/containing/prompt