Installation¶

adaptiSeq is a Python package (Python ≥ 3.10) plus a small set of external command-line tools it shells out to.

1. Install with conda (recommended)¶

The conda package pulls the Python dependencies and every external tool (wget, axel, pigz, coreutils, sra-tools, and aspera-cli for ascp plus its ENA key), so all modes work without further setup:

conda create -n adaptiseq -c conda-forge -c bioconda adaptiseq
conda activate adaptiseq

Note The bioconda submission is pending. Until it is accepted, build the package locally from the recipe in this repository:
conda install -n base conda-build
conda build recipe -c conda-forge -c bioconda
conda create -n adaptiseq -c local -c conda-forge -c bioconda adaptiseq
aspera-cli ships a Linux x86_64 ascp; on other platforms -a is unavailable and the remaining modes work as usual.

2. Install with pip¶

pip install adaptiseq

pip installs only the Python package. Install the external tools (below) yourself, or use conda.

3. Install from source (development)¶

git clone https://github.com/swargo98/adaptiSeq.git
cd adaptiSeq
conda env create -f environment.yml   # external tools + editable install + pytest
conda activate adaptiseq
python -m pytest -q

Without conda, pip install -e '.[test]' gives the editable install and pytest; the external tools must then come from elsewhere.

Packaging notes¶

recipe/meta.yaml is the bioconda recipe. It builds the released PyPI sdist, and its test step runs the offline test suite shipped in that sdist.
environment.yml is the development environment for a checkout.
CI builds the recipe and creates the development environment on every push.

Python dependencies¶

Installed automatically by pip:

Package	Why
`aiohttp` (≥ 3.8)	segmented HTTP(S) engine
`aioftp` (≥ 0.21)	native segmented FTP transport
`numpy` (≥ 1.21)	gradient adaptive-concurrency controller

External command-line tools¶

adaptiSeq fetches metadata and runs integrity/conversion through a small set of external tools. Which ones you need depends on what you ask for:

Tool	Needed for
`wget`	metadata / discovery (always)
`sra-tools` (`srapath`, `fasterq-dump`, `vdb-validate`)	SRA-fallback resolution, FASTQ conversion, md5/validation
`pigz`	gzip compression after conversion (`-g`/`-q`)
`md5sum` (coreutils)	md5 integrity checks (skipped with `-k`)
`axel`	only the opt-in classic engine with `-p`
IBM `ascp` (Aspera)	only `-a`

adaptiSeq uses needs-based preflight: a pure ENA *.fastq.gz download needs only wget — sra-tools and the rest are required (with a clear message) only when a run actually takes a path that uses them. See the FAQ for the exact messages and how to resolve them.

Verifying the install¶

adaptiseq --version
adaptiseq -i SRR7706354 -m          # metadata-only smoke test