RhyDB

High-performance analytical database for sequence alignment data.

Check out rhydb.org to interactively try out this database engine on a few sample instances.

For information on how to build, test, and contribute to RhyDB, see Contributing.

Python Bindings

RhyDB provides Python bindings via Cython. The bindings wrap the core C++ Database and are installable by pip install rhydb.

See Contributing for build instructions.

Usage

from rhydb import Database

# Create a new database
db = Database()

# Or load from a saved state
db = Database("/path/to/saved/database")

# Create a table with a nucleotide sequence column, whose reference is taken
# from the built-in reference_genomes table
db.append_data_from_string(
    "reference_genomes", '{"name": "main", "type": "nucleotide", "sequence": "ACGT..."}'
)
db.query(
    "createTable(sequences, {id := string, country := string, main := nucleotideSequence(reference := main)}, primaryKey := id)"
)

# Append data from file
db.append_data_from_file("sequences", "/path/to/data.ndjson")

# Query data through the SaneQL front-end
table = db.query("sequences.filter(true)")

# Update and delete rows in place
db.query("update(sequences, set := {country := 'Switzerland'}, where := country = 'CH')")
db.query("delete(sequences, where := country.isNull())")

# Save database state
db.save_checkpoint("/path/to/save/directory")

Configuration Files

For RhyDB, there are three different configuration files:

• DatabaseConfig described in file database_config.h
• PreprocessingConfig used when started with preprocessing and described in file preprocessing_config.h. For details see rhydb preprocessing --help.
• RuntimeConfig used when started with api and described in file runtime_config.h For details see rhydb api --help.

The database config contains the schema of the database and is always required when preprocessing data. The database config will be saved together with the output of the preprocessing and is therefore not required when starting RhyDB as an API.

An example configuration file can be seen in testBaseData/exampleDataset/database_config.yaml.

By default, the config files are expected to be YAML files in the current working directory in snake_case (database_config.yaml, preprocessing_config.yaml, runtime_config.yaml), but their location can be overridden using the options --database-config=X, --preprocessing-config=X, and --runtime-config=X.

Preprocessing and Runtime configurations contain default values for all fields and are thus only optional. Their parameters can also be provided as command-line arguments in snake_case and as environment variables prefixed with RHYDB_ in capital SNAKE_CASE. (e.g. RHYDB_INPUT_DIRECTORY).

The precendence is CLI argument > Environment Variable > Configuration File > Default Value

Run The Preprocessing

The preprocessing acts as a program that takes an input directory that contains the to-be-processed data and an output directory where the processed data will be stored. Both need to be mounted to the container.

RhyDB expects a preprocessing config that can be mounted to the default location /app/preprocessing_config.yaml.

Additionally, a database config and a ndjson file containing the data are required. They should typically be mounted in /preprocessing/input.

docker run \
  -v your/input/directory:/preprocessing/input \
  -v your/preprocessing/output:/preprocessing/output \
  -v your/preprocessing_config.yaml:/app/preprocessing_config.yaml \
  ghcr.io/rhyorg/rhydb preprocessing

Both config files can also be provided in custom locations:

rhydb preprocessing --preprocessing-config=./custom/preprocessing_config.yaml --database-config=./custom/database_config.yaml

The Docker image contains a default preprocessing config that sets defaults specific for running RhyDB in Docker. Apart from that, there are default values if neither user-provided nor default config specify fields. The user-provided preprocessing config can be used to overwrite the default values. For a full reference, see the help text.

Run docker container (api)

After preprocessing the data, the api can be started with the following command:

docker run \
  -p 8081:8081 \
  -v your/preprocessing/output:/data \
  ghcr.io/rhyorg/rhydb api

The directory where RhyDB expects the preprocessing output can be overwritten via rhydb api --data-directory=/custom/data/directory or in a corresponding configuration file.

Acknowledgments

Original genome indexing logic with roaring bitmaps by Prof. Neumann: https://db.in.tum.de/~neumann/gi/