Overview
This is a Python implementation of the SOMA API specification for interacting with the Unified Single-cell Data Model.
Installation
TileDB-SOMA is available on PyPI and Conda, and can be installed via pip
or mamba
as indicated below.
python -m pip install tiledbsoma
mamba install -c conda-forge -c tiledb tiledbsoma-py
To install a specific version:
$ python -m pip install git+https://github.com/single-cell-data/TileDB-SOMA.git@0.0.6
To update to the latest version:
$ python -m pip install --upgrade tiledbsoma
In case of illegal instruction
errors when running on older architectures --- e.g. Opteron, non-AVX2 --- the issue is that the pre-compiled binaries available at Conda or PyPI aren't targeted for all processor variants over time. You can install from source, as shown below.
To see if this is the issue, on Linux:
grep avx2 /proc/cpuinfo
If this comes up empty for your system, you'll definitely need to build from source to run TileDB-SOMA on that system.
From source
Status
Please see https://github.com/single-cell-data/TileDB-SOMA/issues.
platform_config
format
When accessing SOMA APIs, TileDB-specific settings can be configured with the platform_config
parameter.
The options accepted by TileDB SOMA are described here, using TypeScript interface syntax:
interface PlatformConfig {
tiledb?: TDBConfig;
}
interface TDBConfig {
create?: TDBCreateOptions;
}
interface TDBCreateOptions {
dims?: { [dim: string]: TDBDimension };
attrs?: { [attr: string]: TDBAttr };
allows_duplicates?: bool;
offsets_filters?: TDBFilter[];
validity_filters?: TDBFilter[];
capacity?: number;
cell_order?: string;
tile_order?: string;
}
interface TDBDimension {
filters?: TDBFilter[];
tile?: number;
}
interface TDBAttr {
filters?: TDBFilter[];
}
type TDBFilter = string | TDBFilterSpec;
interface TDBFilterSpec {
_name: string;
[kwarg: string]: any;
}
Information for developers
Please see the TileDB-SOMA wiki.