### Install AutoRAG

Source: https://marker-inc-korea.github.io/AutoRAG/_sources/install.md

Installs the core AutoRAG package using pip. This is the primary method for getting started with AutoRAG.

```bash
pip install AutoRAG
```

--------------------------------

### Start vLLM API Server

Source: https://marker-inc-korea.github.io/AutoRAG/nodes/generator/vllm_api

Instructions on how to start the vLLM API server. It references the official vLLM documentation for detailed setup and usage of the OpenAI-compatible API server.

```shell
# Start vLLM API server
# Refer to vLLM documentation for the exact command and options.
# Example: python -m vllm.entrypoints.openai.api_server --host 0.0.0.0 --port 8012
# The port 8012 is commonly used for the vLLM API server.
```

--------------------------------

### Install Recommended Libraries

Source: https://marker-inc-korea.github.io/AutoRAG/_sources/install.md

Installs PyOpenSSL and nltk, which are recommended for full AutoRAG features. It also includes commands to download necessary nltk data like 'punkt_tab' and 'averaged_perceptron_tagger_eng'.

```bash
pip install --upgrade pyOpenSSL
pip install nltk
python3 -c "import nltk; nltk.download('punkt_tab')"
python3 -c "import nltk; nltk.download('averaged_perceptron_tagger_eng')"
```

--------------------------------

### Install Requirements for Testing and Docs

Source: https://marker-inc-korea.github.io/AutoRAG/_sources/install.md

Installs additional packages required for running tests and building documentation when building AutoRAG from source.

```bash
pip install -r tests/requirements.txt
pip install -r docs/requirements.txt
```

--------------------------------

### Prompt Maker Configuration Example

Source: https://marker-inc-korea.github.io/AutoRAG/_sources/nodes/prompt_maker/prompt_maker.md

Example of a config.yaml file demonstrating the setup for the prompt_maker node, including strategy parameters and module definitions.

```yaml
node_lines:
- node_line_name: post_retrieve_node_line  # Arbitrary node line name
  nodes:
    - node_type: prompt_maker
      strategy:
        metrics: [bleu, meteor, rouge, sem_score]
        speed_threshold: 10
        token_threshold: 1000
        tokenizer: gpt-3.5-turbo
        generator_modules:
          - module_type: llama_index_llm
            llm: openai
            model: [gpt-3.5-turbo-16k, gpt-3.5-turbo-1106]
      modules:
        - module_type: fstring
          prompt: ["Tell me something about the question: {query} \n\n {retrieved_contents}",
                   "Question: {query} \n Something to read: {retrieved_contents} \n What's your answer?"]
```

--------------------------------

### Install AutoRAG

Source: https://marker-inc-korea.github.io/AutoRAG/_sources/index.rst

Installs the AutoRAG package using pip. This is the primary method to get started with AutoRAG on your local machine.

```bash
pip install AutoRAG
```

--------------------------------

### Run AutoRAG GUI from Source

Source: https://marker-inc-korea.github.io/AutoRAG/_sources/gui/gui.md

This sequence of commands guides users through building and running the AutoRAG GUI application directly from the source code. It involves navigating to the frontend directory, configuring environment variables, installing dependencies, building the application, and starting the server.

```bash
cd autorag-frontend
yarn install
npm run build
npm run start
```

--------------------------------

### Install AutoRAG from Source

Source: https://marker-inc-korea.github.io/AutoRAG/_sources/install.md

Installs AutoRAG in editable mode after cloning the repository. This is useful for development or contributing to the project.

```bash
cd AutoRAG
pip install -e .
```

--------------------------------

### Clone AutoRAG Repository

Source: https://marker-inc-korea.github.io/AutoRAG/_sources/install.md

Clones the AutoRAG GitHub repository to build the project from source.

```bash
git clone https://github.com/Marker-Inc-Korea/AutoRAG.git
```

--------------------------------

### Install AutoRAG with GPU Support

Source: https://marker-inc-korea.github.io/AutoRAG/_sources/install.md

Installs AutoRAG with additional dependencies required for GPU acceleration, enabling faster processing for compatible hardware.

```bash
pip install "AutoRAG[gpu]"
```

--------------------------------

### Install AutoRAG with Parsing Support

Source: https://marker-inc-korea.github.io/AutoRAG/_sources/install.md

Installs AutoRAG with dependencies necessary for document parsing. This requires external system packages like libmagic, tesseract, and poppler to be installed first.

```bash
pip install "AutoRAG[parse]"
```

--------------------------------

### Run Autorag GPU Version

Source: https://marker-inc-korea.github.io/AutoRAG/_sources/install.md

Shows how to execute the autorag:gpu Docker image, enabling GPU acceleration. It requires CUDA and cuDNN on the host and uses the --gpus all flag for GPU access.

```bash
docker run --rm -it \
  -v ~/.cache/huggingface:/cache/huggingface \
  -v $(pwd)/sample_config:/usr/src/app/sample_config \
  -v $(pwd)/projects:/usr/src/app/projects \
  -e HF_HOME=/cache/huggingface \
  --gpus all \
  autoraghq/autorag:gpu evaluate \
  --config /usr/src/app/sample_config/rag/simple/simple_openai.yaml \
  --qa_data_path /usr/src/app/projects/test01/qa_validation.parquet \
  --corpus_data_path /usr/src/app/projects/test01/corpus.parquet \
  --project_dir /usr/src/app/projects/test01
```

--------------------------------

### Install AutoRAG GUI with Docker Compose

Source: https://marker-inc-korea.github.io/AutoRAG/_sources/gui/gui.md

This snippet demonstrates the commands to clone the AutoRAG repository, including submodules, and then run the application using Docker Compose. It's the recommended method for a quick setup.

```bash
git clone --recurse-submodules https://github.com/Marker-Inc-Korea/AutoRAG.git
cd AutoRAG
docker compose up
```

--------------------------------

### Run Autorag Evaluation with Custom HF Cache

Source: https://marker-inc-korea.github.io/AutoRAG/_sources/install.md

Demonstrates how to run the autorag:api Docker image for evaluation. It shows mounting Hugging Face cache to a custom location and setting the HF_HOME environment variable for efficient model loading.

```bash
docker run --rm -it \
  -v ~/.cache/huggingface:/cache/huggingface \
  -v $(pwd)/sample_config:/usr/src/app/sample_config \
  -v $(pwd)/projects:/usr/src/app/projects \
  -e HF_HOME=/cache/huggingface \
  autoraghq/autorag:api evaluate \
  --config /usr/src/app/sample_config/rag/simple/simple_openai.yaml \
  --qa_data_path /usr/src/app/projects/test01/qa_validation.parquet \
  --corpus_data_path /usr/src/app/projects/test01/corpus.parquet \
  --project_dir /usr/src/app/projects/test01
```

--------------------------------

### vllm Module Example Configuration

Source: https://marker-inc-korea.github.io/AutoRAG/nodes/generator/vllm

An example configuration snippet demonstrating how to set up the vllm module in a config.yaml file for AutoRAG.

```yaml
modules:
- module_type: vllm
  llm: mistralai/Mistral-7B-Instruct-v0.2
  temperature: [0.1, 1.0]
  max_tokens: 512
```

--------------------------------

### Window Replacement Example Config

Source: https://marker-inc-korea.github.io/AutoRAG/nodes/prompt_maker/window_replacement

Example configuration for the Window Replacement prompt maker module, specifying the prompt templates to be used. The prompt string must contain `{query}` and `{retrieved_contents}` placeholders.

```YAML
modules:
-module_type:window_replacement
prompt:["Tellmesomethingaboutthequestion:{query}\n\n{retrieved_contents}",
"Question:{query}\nSomethingtoread:{retrieved_contents}\nWhat'syouranswer?"]
```

--------------------------------

### Install AutoRAG

Source: https://marker-inc-korea.github.io/AutoRAG/index

Installs the AutoRAG library using pip. This is the primary method to get started with the AutoRAG tool for building and optimizing RAG pipelines.

```bash
pip install AutoRAG
```

--------------------------------

### Create .env File for API Key

Source: https://marker-inc-korea.github.io/AutoRAG/_sources/install.md

Example content for a .env file to store sensitive API keys, such as the OPENAI_API_KEY. This file should be placed in the root folder.

```dotenv
OPENAI_API_KEY=sk-...your-api-key...
```

--------------------------------

### Qdrant Configuration Example

Source: https://marker-inc-korea.github.io/AutoRAG/_sources/integration/vectordb/qdrant.md

Provides an example YAML configuration for setting up the Qdrant vector database with specific embedding models and client types.

```yaml
- name: openai_qdrant
  db_type: qdrant
  embedding_model: openai_embed_3_large
  collection_name: openai_embed_3_large
  client_type: docker
  embedding_batch: 50
  similarity_metric: cosine
  dimension: 1536
```

```yaml
vectordb:
  - name: openai_qdrant
    db_type: qdrant
    embedding_model: openai_embed_3_large
    collection_name: openai_embed_3_large
    client_type: docker
    embedding_batch: 50
    similarity_metric: cosine
    dimension: 1536
node_lines:
- node_line_name: retrieve_node_line  # Arbitrary node line name
  nodes:
    - node_type: semantic_retrieval
      strategy:
        metrics: [retrieval_f1, retrieval_recall, retrieval_precision]
      top_k: 3
      modules:
        - module_type: vectordb
          vectordb: openai_qdrant
- node_line_name: post_retrieve_node_line  # Arbitrary node line name
  nodes:
    - node_type: prompt_maker
      strategy:
        metrics: [bleu, meteor, rouge]
      modules:
        - module_type: fstring
          prompt: "Read the passages and answer the given question. 
 Question: {query} 
 Passage: {retrieved_contents} 
 Answer : "
    - node_type: generator
      strategy:
        metrics: [bleu, rouge]
      modules:
        - module_type: llama_index_llm
          llm: openai
          model: [ gpt-4o-mini ]
```

--------------------------------

### Access Autorag Container for Debugging

Source: https://marker-inc-korea.github.io/AutoRAG/_sources/install.md

Provides a command to start a Bash shell inside the autorag:api Docker container. This is useful for manual inspection, running commands, or troubleshooting issues within the container environment.

```bash
docker run --rm -it --entrypoint /bin/bash autoraghq/autorag:api
```

--------------------------------

### Install AutoRAG for Japanese Language Support

Source: https://marker-inc-korea.github.io/AutoRAG/_sources/install.md

Installs optional dependencies for Japanese language processing.

```bash
pip install "AutoRAG[ja]"
```

--------------------------------

### Run AutoRAG Web Interface with Runner (from Trial Folder)

Source: https://marker-inc-korea.github.io/AutoRAG/_sources/deploy/web.md

Initializes the AutoRAG Runner from a trial folder and starts the web interface, with options to configure server binding and public sharing. The `share=True` option generates a public link.

```python
from autorag.deploy import Runner

runner = Runner.from_trial_folder('your/path/to/trial_folder')
runner.run_web(server_name="0.0.0.0", server_port=7680, share=True)
```

--------------------------------

### Chroma Initialization via YAML

Source: https://marker-inc-korea.github.io/AutoRAG/_sources/integration/vectordb/chroma.md

Example YAML structure for initializing the Chroma vector store, specifying database type, client type, and connection details.

```yaml
vectordb:
  - name: chroma_default
    db_type: chroma
    client_type: persistent
    embedding_model: mock
    collection_name: openai
    path: ${PROJECT_DIR}/resources/chroma
```

--------------------------------

### Install AutoRAG for Korean Language Support

Source: https://marker-inc-korea.github.io/AutoRAG/_sources/install.md

Installs optional dependencies for Korean language processing. Requires JDK 17 to be installed and configured in the environment path for konlpy.

```bash
pip install "AutoRAG[ko]"
```

--------------------------------

### Start vLLM API Server

Source: https://marker-inc-korea.github.io/AutoRAG/_sources/nodes/generator/vllm_api.md

Command to launch the vLLM API server. It specifies the model to serve and the port for the API endpoint.

```bash
vllm serve Qwen/Qwen2.5-14B-Instruct-AWQ -q awq --port 8012
```

--------------------------------

### Milvus Configuration Examples

Source: https://marker-inc-korea.github.io/AutoRAG/_sources/integration/vectordb/milvus.md

Provides example YAML configurations for setting up the Milvus vector database, including connection details, embedding model specifications, and retrieval node parameters.

```YAML
- name: openai_milvus
  db_type: milvus
  embedding_model: openai_embed_3_large
  collection_name: openai_embed_3_large
  uri: ${MILVUS_URI}
  token: ${MILVUS_TOKEN}
  embedding_batch: 50
  similarity_metric: cosine
  index_type: IVF_FLAT
  params:
    nlist: 16384


```

```YAML
vectordb:
  - name: openai_milvus
    db_type: milvus
    embedding_model: openai_embed_3_large
    collection_name: openai_embed_3_large
    uri: ${MILVUS_URI}
    token: ${MILVUS_TOKEN}
    embedding_batch: 50
    similarity_metric: cosine
    index_type: IVF_FLAT
    params:
      nlist: 16384
node_lines:
- node_line_name: retrieve_node_line  # Arbitrary node line name
  nodes:
    - node_type: semantic_retrieval
      strategy:
        metrics: [retrieval_f1, retrieval_recall, retrieval_precision]
      top_k: 3
      modules:
        - module_type: vectordb
          vectordb: openai_milvus
- node_line_name: post_retrieve_node_line  # Arbitrary node line name
  nodes:
    - node_type: prompt_maker
      strategy:
        metrics: [bleu, meteor, rouge]
      modules:
        - module_type: fstring
          prompt: "Read the passages and answer the given question. 
 Question: {query} 
 Passage: {retrieved_contents} 
 Answer : "
    - node_type: generator
      strategy:
        metrics: [bleu, rouge]
      modules:
        - module_type: llama_index_llm
          llm: openai
          model: [ gpt-4o-mini ]

```

--------------------------------

### FlashRank Reranker Configuration Example

Source: https://marker-inc-korea.github.io/AutoRAG/nodes/passage_reranker/flashrank_reranker

An example configuration file demonstrating how to set up the FlashRank reranker module, including batch size and the specific model to be used.

```yaml
-module_type:flashrank_reranker
batch:32
model:"ms-marco-MiniLM-L-12-v2"

```

--------------------------------

### Example config.yaml for Query Expansion

Source: https://marker-inc-korea.github.io/AutoRAG/nodes/query_expansion/query_expansion

An example configuration file demonstrating how to set up the Query Expansion node within AutoRAG. It includes strategy parameters, retrieval modules, and specific module configurations.

```yaml
node_lines:
-node_line_name:pre_retrieve_node_line# Arbitrary node line name
nodes:
-node_type:query_expansion
strategy:
metrics:[retrieval_f1,retrieval_recall,retrieval_precision]
speed_threshold:10
top_k:10
retrieval_modules:
-module_type:bm25
-module_type:vectordb
vectordb:default
modules:
-module_type:pass_query_expansion
-module_type:query_decompose
llm:openai
temperature:[0.2,1.0]
-module_type:hyde
llm:openai
max_token:64

```

--------------------------------

### Run AutoRAG Web Interface from Trial Folder

Source: https://marker-inc-korea.github.io/AutoRAG/tutorial

Initializes the GradioRunner using the configuration and results from a previous AutoRAG trial directory, then launches the web interface. This allows for easy deployment of optimized pipelines.

```python
from autorag.deploy import GradioRunner

runner = GradioRunner.from_trial_folder('/your/path/to/trial_dir')
runner.run_web()
```

--------------------------------

### VoyageAI Reranker Setup

Source: https://marker-inc-korea.github.io/AutoRAG/nodes/passage_reranker/voyageai_reranker

Instructions for setting up the VoyageAI API key, either via environment variables or directly in the configuration file.

```shell
export VOYAGE_API_KEY=your_voyageai_api_key
```

--------------------------------

### Query Decompose Configuration Example

Source: https://marker-inc-korea.github.io/AutoRAG/nodes/query_expansion/query_decompose

An example configuration file (config.yaml) demonstrating how to set up the query_decompose module, specifying the generator module type and LLM.

```YAML
modules:
-module_type:query_decompose
generator_module_type:llama_index_llm
llm:openai
model:[gpt-3.5-turbo-16k,gpt-3.5-turbo-1106]
```

--------------------------------

### Get Start and End Index (Python)

Source: https://marker-inc-korea.github.io/AutoRAG/_modules/autorag/data/utils/util

Finds the starting and ending indices of a substring within a larger string. If the substring is not found, it returns (0, 0). This utility is useful for text processing and locating specific patterns.

```python
defget_start_end_idx(original_text: str, search_str: str) -> Tuple[int, int]:
	"""Get the start and end index of a search string in the original text.

	Args:
	    original_text: str: The text to search within.
	    search_str: str: The string to search for.

	Returns:
	    Tuple[int, int]: A tuple containing the start and end index (inclusive) of the search string.
	"""
	start_idx = original_text.find(search_str)
	if start_idx == -1:
		return 0, 0
	end_idx = start_idx + len(search_str)
	return start_idx, end_idx - 1
```

--------------------------------

### AutoRAG Passage Filter Configuration Example

Source: https://marker-inc-korea.github.io/AutoRAG/nodes/passage_filter/passage_filter

An example `config.yaml` file illustrating the setup for a passage filter node in AutoRAG. It defines node lines, the passage filter node type, retrieval metrics, speed threshold, and the specific modules like `pass_passage_filter` and `similarity_threshold_cutoff` with their respective parameters.

```yaml
node_lines:
-node_line_name:retrieve_node_line# Arbitrary node line name
nodes:
-node_type:passage_filter
strategy:
metrics:[retrieval_f1,retrieval_recall,retrieval_precision]
speed_threshold:5
modules:
-module_type:pass_passage_filter
-module_type:similarity_threshold_cutoff
threshold:0.85

```

--------------------------------

### Run Pytest

Source: https://marker-inc-korea.github.io/AutoRAG/_sources/install.md

Executes the test suite for AutoRAG using pytest. The '-n auto' flag enables parallel test execution.

```bash
python -m pytest -n auto
```

--------------------------------

### Full AutoRAG YAML Configuration with Qdrant

Source: https://marker-inc-korea.github.io/AutoRAG/integration/vectordb/qdrant

This example demonstrates a more comprehensive AutoRAG configuration file, specifying Qdrant as the vector database and detailing retrieval and generation node setups.

```yaml
vectordb:
-name:openai_qdrant
db_type:qdrant
embedding_model:openai_embed_3_large
collection_name:openai_embed_3_large
client_type:docker
embedding_batch:50
similarity_metric:cosine
dimension:1536
node_lines:
-node_line_name:retrieve_node_line# Arbitrary node line name
nodes:
-node_type:semantic_retrieval
strategy:
metrics:[retrieval_f1,retrieval_recall,retrieval_precision]
top_k:3
modules:
-module_type:vectordb
vectordb:openai_qdrant
-node_line_name:post_retrieve_node_line# Arbitrary node line name
nodes:
-node_type:prompt_maker
strategy:
metrics:[bleu,meteor,rouge]
modules:
-module_type:fstring
prompt:"Readthepassagesandanswerthegivenquestion.\nQuestion:{query}\nPassage:{retrieved_contents}\nAnswer:"
-node_type:generator
strategy:
metrics:[bleu,rouge]
modules:
-module_type:llama_index_llm
llm:openai
model:[gpt-4o-mini]

```

--------------------------------

### Start Parsing Documents

Source: https://marker-inc-korea.github.io/AutoRAG/data_creation/tutorial

Initiates the parsing process for raw documents using a specified configuration file. The results are stored in a designated project directory.

```python
from autorag.parser import Parser

parser = Parser(data_path_glob="your/data/path/*", project_dir="./parse_project_dir")
parser.start_parsing("your/path/to/parse_config.yaml")

```

--------------------------------

### Configure Pytest for Environment Variables

Source: https://marker-inc-korea.github.io/AutoRAG/_sources/install.md

Sets up a pytest.ini file to define environment variables, such as OPENAI_API_KEY, for testing purposes. This ensures tests can access necessary credentials.

```ini
[pytest]
env =
    OPENAI_API_KEY=sk-...your-api-key...

log_cli=true
log_cli_level=INFO
```

--------------------------------

### Run AutoRAG API Server via CLI

Source: https://marker-inc-korea.github.io/AutoRAG/_sources/tutorial.md

Launch the AutoRAG pipeline as an API server using the command-line interface. This method is convenient for quick deployments and allows specifying the trial directory, host, and port.

```bash
autorag run_api --trial_dir /trial/dir/0 --host 0.0.0.0 --port 8000
```

--------------------------------

### Generator Node Configuration Example

Source: https://marker-inc-korea.github.io/AutoRAG/nodes/generator/generator

An example configuration file (config.yaml) demonstrating how to set up the Generator node. It specifies metrics for evaluation, LLM models to use, and parameters like batch size, speed threshold, and token threshold.

```yaml
-node_line_name:post_retrieve_node_line# Arbitrary node line name
nodes:
-node_type:generator
strategy:
metrics:
-metric_name:bleu
-metric_name:meteor
-metric_name:sem_score
embedding_model:openai
-metric_name:g_eval
metrics:[consistency,fluency,relevance,coherence]
model:gpt-4
-metric_name:bert_score
lang:en
batch:64
speed_threshold:10
token_threshold:4000
modules:
-module_type:llama_index_llm
llm:[openai]
model:[gpt-3.5-turbo-16k,gpt-3.5-turbo-1106]
temperature:[0.5,1.0,1.5]

```

--------------------------------

### Build Docker Image for AutoRAG

Source: https://marker-inc-korea.github.io/AutoRAG/_sources/install.md

Builds a production-ready Docker image for AutoRAG. The `--target production` flag ensures only the production stage from the Dockerfile is used.

```docker
docker build --target production -t autorag:prod .
```

--------------------------------

### Start Parsing Process

Source: https://marker-inc-korea.github.io/AutoRAG/data_creation/parse/parse

Initiates the document parsing pipeline by calling the `start_parsing` method on the parser instance, passing the path to the configuration YAML file.

```python
parser.start_parsing("your/path/to/parse_config.yaml")
```

--------------------------------

### Set OpenAI API Key Environment Variable

Source: https://marker-inc-korea.github.io/AutoRAG/_sources/install.md

Configures the OPENAI_API_KEY environment variable, which is essential for using OpenAI models with AutoRAG. This can be done directly in the terminal.

```bash
export OPENAI_API_KEY="sk-...your-api-key..."
```