Make the number of background thread configurable.

Add a FTPClientFactory with minimal capabilities for reusing
connections based on commons-pool2.

Make credentials for NCBI, GEO and ArrayExpress FTP servers
configurable.

…-cell

Eliminate nano context-related work now that CLIs are lazy-loaded.

…-cell

…-cell

…-cell

…-cell

…-cell

Separate MPSS from SRA sample types.

Add a null-check for LibrarySource.

…ackages are installed in tests

…ials

Fix missing awaitBatchTaskExecutorService() after submitting batch
tasks, but awaitng them so that they can write their progress to the
summary output file.

Include the filename of all additional supplementary materials in the
summary output file.

…ound

…ples

Add a new utility for retrying that is simpler to use than
RetryTemplate.

The toString() of URL does not include user information, so it's safe to
use for logging and error messages.

When checking if a sample is single-cell or retrieving additional
supplementary materials, consider the files that the series provide.

This allow resolving cases of MEX datasets where the genes or features
are shared across all samples.

This allow for graceful shutdown if SIGTERM is sent to the Java process.

…yed on any exception

…nning

…5 1.12

Looking up the size and opening a stream at the same time can result in
a deadlock.

Fix TarInputStream not closing the base stream by explicitly adding the
base stream to the try-with-resources block.

Fix lastAttempt check.

Use nulls for cell, gene, sample counts so that they do not end-up as zeroes
for unsupported datasets.

Allow overriding the filename pattern for detecting MEX files from the
CLI. This can be used as a workaround instead of waiting for GEO to fix
the submission.

Add support for detecting Loom files at the series and sample level, but
not for downloading or loading them.

…ed datasets

Retrieve additional files before downloading so that if the download
fails, the files are still listed in the summary output.

Ignore userinfo from the URL.

Add a test case for duplicated cell IDs.

Add an option to process specific samples.

Introduce a hierarchy of BioMaterial such that a sample can be derived
from another sample and inherit its characteristics, treatments and
factor values.

Prevent splitting an experiment that uses a hierarchy of biomaterial for
now. This might be implemented in the future.

Add a separate field in QuantitationType to identify preferred
single-cell QT.

Preserve sparsity of categorical columns as much as possible.

…er logic

Add support for a simple tabular format for importing cell type
annotations.

Fix size-checking of compressed file and buffer FTP streams

Add support for looking at transposed AnnData file and detecting sample
and cell type columns.

Redownload truncated H5 files and handle non-string categorical arrays

…-single-cell

…tegory

Use ByteArrayUtils in ByteArrayType

There's some optimization for converting byte[] to various data type
that are beneficial for loading single-cell data.