0.11.1

Reads the looms other tools write, as found converting 29 real looms from
Baysor, R's LoomExperiment and anndata/loompy: all now convert, and match
anndata.read_loom where it can read them.

Fixed

• A .loom holding the matrix alone is read
  (#59). Baysor writes some
  looms as /matrix with no attribute groups, which were not recognised.
  In a file named .loom that is now a loom, its cells and genes numbered;
  a 7,979,604 x 8,729 one converted in 10.5 minutes at a 211 MB peak.

• Baysor's gene names are used (#59).
  Baysor names genes in a Name attribute; its looms came out with
  numbered genes.

• R's LoomExperiment looms read cleanly (#59).
  The spec version, stored as a one-element array, showed as ['2.0.1'],
  and graph vertices stored as floats were truncated without a check; they
  must now be whole numbers. loompy, and so anndata.read_loom, refuses
  these files.

Benchmark vs anndata/scanpy

Results — 0.11.1

Tier	ci, 50,000 obs x 20,000 var, 5% dense CSR
Var-heavy shape	2,000 obs x 36,601 var
Page cache	dropped between runs
Address-space ceiling	12.9 GB per process
Repeats	3 (fastest shown)
Platform	Linux x86_64, Python 3.12.3
Versions	adata-cli 0.11.1, anndata 0.13.4, scanpy 1.12.4
Generated	2026-10-19T19:34:47Z

startup

What does each contender cost before it does any work?

Contender	Wall time	Peak RSS	Output	vs previous
adata-cli	0.52 s	68.5 MB	-	no change
anndata	1.30 s	133.6 MB	-	time -13%
scanpy	2.81 s	260.2 MB	-	time -13%

concat-inner

Concatenate two stores on the obs axis, inner join.

Contender	Wall time	Peak RSS	Output	vs previous
adata-cli	7.60 s	174.9 MB	583.3 MB	time -37%
anndata (concat_on_disk)	4.78 s	450.9 MB	831.9 MB	time -21%
anndata (in memory)	7.56 s	1.8 GB	580.3 MB	time -38%
scanpy	n/a	n/a	n/a	-
	scanpy has no concat of its own; it re-exports anndata's

concat-merge-same

Concatenate carrying var columns forward -- the REQ-71798 case.

Contender	Wall time	Peak RSS	Output	vs previous
adata-cli	1.24 s	193.5 MB	45.7 MB	time -29%
anndata (concat_on_disk)	1.82 s	245.3 MB	63.2 MB	time -18%
anndata (in memory)	2.03 s	289.1 MB	44.6 MB	time -23%

subset-query

Keep the obs rows matching a predicate on an existing column.

Contender	Wall time	Peak RSS	Output	vs previous
adata-cli	2.73 s	210.9 MB	147.1 MB	time -33%
anndata (backed)	3.48 s	675.4 MB	146.6 MB	time -30%
anndata (in memory)	3.74 s	985.0 MB	146.6 MB	time -29%
scanpy (filter_cells)	6.32 s	1.6 GB	291.2 MB	time -29%

h5ad-to-zarr

Convert a store to Zarr.

Contender	Wall time	Peak RSS	Output	vs previous
adata-cli	5.31 s	123.1 MB	151.6 MB in 4717 files	time -19%
anndata (in memory)	4.77 s	776.2 MB	136.8 MB in 55 files	time -20%
scanpy	n/a	n/a	n/a	-
	no streaming converter; scanpy defers to anndata

inspect

Report what is in the store, without reading the matrix.

Contender	Wall time	Peak RSS	Output	vs previous
adata-cli	0.53 s	70.3 MB	-	time -11%
anndata (read_elem on obs)	1.51 s	160.8 MB	-	time -11%
anndata (full load)	2.57 s	584.6 MB	-	time -18%

export-obs

Write the obs table out as CSV.

Contender	Wall time	Peak RSS	Output	vs previous
adata-cli	0.66 s	84.1 MB	1.8 MB	time -14%
anndata (read_elem on obs)	1.51 s	161.0 MB	1.8 MB	time -12%
anndata (full load)	2.57 s	584.7 MB	1.8 MB	time -21%

split-by-sample

Write one store per distinct value of an obs column.

Contender	Wall time	Peak RSS	Output	vs previous
adata-cli	4.20 s	197.5 MB	302.6 MB in 9 files	time -36%
anndata (hand-written loop)	5.39 s	685.3 MB	301.7 MB in 8 files	time -31%

ls

List everything in the store.

Contender	Wall time	Peak RSS	Output	vs previous
adata-cli	0.54 s	69.9 MB	-	time -12%
h5py (visit)	1.39 s	136.7 MB	-	time -15%
h5ls -r	n/a	n/a	n/a	-
	h5ls is not installed

create

Write an empty store with a given obs/var shape.

Contender	Wall time	Peak RSS	Output	vs previous
adata-cli	0.58 s	83.3 MB	3.4 MB	time -13%
anndata	6.28 s	228.9 MB	49.2 MB	time -28%

export-array

Write a dense element out as .npy.

Contender	Wall time	Peak RSS	Output	vs previous
adata-cli	0.55 s	82.2 MB	10.0 MB	time -11%
anndata (read_elem)	1.40 s	154.9 MB	10.0 MB	time -13%
anndata (full load)	2.59 s	584.3 MB	10.0 MB	time -20%

export-sparse

Write X out as Matrix Market text.

Contender	Wall time	Peak RSS	Output	vs previous
adata-cli	10.27 s	254.8 MB	661.3 MB	no change
anndata (sparse_dataset)	5.50 s	968.4 MB	661.6 MB	time -22%
anndata (full load)	5.45 s	775.6 MB	661.6 MB	time -24%

import-dataframe

Replace obs from a CSV.

Contender	Wall time	Peak RSS	Output	vs previous
adata-cli	1.20 s	113.3 MB	296.0 MB	no change
anndata	4.40 s	597.0 MB	291.2 MB	time -37%

import-10x-mtx

Read a gzipped Cell Ranger 3 MTX folder of the same counts into an h5ad. adata-cli streams it in two passes and keeps integer counts; scanpy loads it and converts them to float32.

Contender	Wall time	Peak RSS	Output	vs previous
adata-cli	15.46 s	351.9 MB	279.2 MB	no change
scanpy	12.54 s	3.3 GB	284.6 MB	time -11%

concat-outer

Concatenate two stores keeping the union of variables.

Contender	Wall time	Peak RSS	Output	vs previous
adata-cli	7.66 s	180.5 MB	583.3 MB	time -38%
anndata (concat_on_disk)	4.89 s	449.7 MB	831.9 MB	time -21%
anndata (in memory)	7.66 s	1.8 GB	580.3 MB	time -38%

recast-narrow

Narrow X from float64 values with int64 indices to float32 and int32, as issue #13 asked. X is uncompressed, as write_h5ad leaves it by default, so the output keeps that and the baseline writes it the same way.

Contender	Wall time	Peak RSS	Input → output	vs previous
adata-cli	2.28 s	118.5 MB	815.6 MB → 415.4 MB (-49%)	no change
anndata (in memory)	3.82 s	1.8 GB	815.6 MB → 417.3 MB (-49%)	no change
scanpy	n/a	n/a	n/a	-
	no dtype rewrite; scanpy defers to anndata

recast-narrow-lzf

The same narrowing on an lzf-compressed X. Integer counts in float64 are mostly zero bytes, which lzf already removes, so expect a small saving on disk; the halving is in memory, once the matrix is loaded.

Contender	Wall time	Peak RSS	Input → output	vs previous
adata-cli	5.21 s	126.4 MB	302.5 MB → 291.1 MB (-4%)	time -35%
anndata (in memory)	5.45 s	1.8 GB	302.5 MB → 291.2 MB (-4%)	time -33%

convert-layout

Transpose X from CSR to CSC. The output should be the size of the input: 0.7.0 fixed a --layout that wrote it three times larger.

Contender	Wall time	Peak RSS	Input → output	vs previous
adata-cli (streaming)	15.12 s	374.6 MB	291.2 MB → 297.0 MB (+2%)	no change
adata-cli (--in-memory)	7.68 s	2.3 GB	291.2 MB → 297.0 MB (+2%)	time -19%
anndata (in memory)	5.09 s	984.6 MB	291.2 MB → 297.0 MB (+2%)	time -28%

rss-vs-size

Does peak memory track input size, at a fixed --chunk? This is the README's claim, stated as a measurement.

Contender	Wall time	Peak RSS	Output	vs previous
adata-cli	7.54 s	174.3 MB	583.3 MB	time -37%
anndata (in memory)	7.45 s	1.8 GB	580.3 MB	time -39%

---

Rows marked n/a are operations the baseline does not offer; that is a result, not a gap in the measurement. Where a baseline is faster, the row stands as it is -- the trade this tool makes is memory for time, and hiding the cost would make the table useless.

Output size is the file as the filesystem reports it, which for HDF5 includes allocation slack. On a small tier that slack can be several times the stored bytes and the column says more about the writer's allocation strategy than about the data; at ci and above it is noise. Zarr stores report a file count alongside, because a directory of small chunks measures much larger than it holds. Where a case exists to change a file's size, the column reads input → output, with the change.