where is the starter.yaml referred in Tutorial doc?

[txsing]

In the tutorial document:

it said:

Let’s find an optimal RAG pipeline with AutoRAG! After you prepare your evaluation dataset, you need to have a config YAML file. There are few pre-made config YAML files at our GitHub repo sample_config folder. We highly recommend using pre-made config YAML files for starter. Download starter.yaml file to your local environment, and you are ready to go.

However, i could not find the starter.yaml file in the repo. Do you renamed it? Where i can find such config file to have a quick try on AutoRAG?
A lot of thx.

[e7217]

@txsing
I think it may have been changed.
If you need a simple example, how about referring to the link below?
The maintainers provide the example repository.

https://github.com/Marker-Inc-Korea/AutoRAG-tutorial

[txsing]

Thank you very much! I’ll take a look. It would still be great if the tutorial document could be updated to help more beginners avoid getting stuck at the very early stages.

[txsing]

@txsing I think it may have been changed. If you need a simple example, how about referring to the link below? The maintainers provide the example repository.

https://github.com/Marker-Inc-Korea/AutoRAG-tutorial

Seems like this tutorial repo is again not compatible to latest version of AutoRAG. I get stucked at very beginning step again. Is there a latest tutorial code for starters?

[e7217]

@txsing
The tutorial may not reflect the most recent version of the code.
If you just want to run the sample, I recommend installing AutoRAG[parse]==0.3.10 as the documentation may not be up-to-date.

If needed, please follow these steps:

1. Install the additional package:

pip install pycryptodome

2. Run the parser:

python run_parse.py

with this configuration.

modules:
  - module_type: langchain_parse
    parse_method: [ pdfminer, pypdf, pymupdf ]
  # - module_type: llamaparse
  #   result_type: markdown
  #   language: en
  # - module_type: table_hybrid_parse
  #   text_parse_module: langchain_parse
  #   text_params:
  #     parse_method: pdfplumber
  #   table_parse_module: llamaparse
  #   table_params:
  #     result_type: markdown
  #     language: en

3. Run the chunker:

python run_chunk.py --raw_path ./parsed_raw/0.parquet

4. Run the QA maker:

python make_qa.py --raw_path ./parsed_raw/0.parquet --corpus_path ./chunked_corpus/0.parquet --qa_size 5

remember this:

5. Run the evaluator:

python main.py --config ./config/tutorial.yaml

I hope this helps!