Package index • dafr

Core data model

DafReader(): Base abstract reader class.
DafReadOnly(): Abstract read-only reader class.
DafWriter(): Abstract writer class.
MemoryDaf(): Concrete DafWriter backed by R environments (no I/O).
FilesDaf(): File-backed Daf writer class.
FilesDafReadOnly(): File-backed read-only Daf class.
ZarrDaf(): Zarr-backed Daf class.
ZarrDafReadOnly(): Read-only Zarr-backed Daf class.
HttpDaf() http_daf(): Read-only HTTP-served FilesDaf.
memory_daf(): In-memory Daf store.
files_daf(): File-backed Daf store.
zarr_daf(): Open a Zarr-backed Daf store.
pack_files_daf_metadata(): Pack a FilesDaf directory's JSON metadata into metadata.zip.
is_daf(): Test whether an object is a DafReader.
is_leaf(): Is daf a leaf storage format?
daf_name(): Return the name of a DafReader.
description(): Human-readable summary of a Daf store.
complete_path(): Canonical disk path of a (possibly chained) daf.
read_only(): Wrap a writer into a read-only view via a 1-element chain.
empty_cache(): Empty caches on a Daf object.

Readers

get_scalar(): Get a scalar, optionally with a default when missing.
get_vector(): Get a vector, returning it as an axis-named R vector.
get_matrix(): Get a matrix, returning it with axis-entry dimnames.
has_scalar(): Test whether a scalar exists.
has_vector(): Test whether a vector exists on an axis.
has_matrix(): Test whether a matrix exists for an axis pair.
has_axis(): Test whether an axis exists.
scalars_set(): Names of all scalars, sorted.
axes_set(): Names of all axes, sorted.
axis_entries(): Entry names of an axis (full or by index).
axis_dict(): Entry-name to 1-based-index hash for an axis.
axis_indices(): Look up 1-based positions of entries in an axis.
axis_length(): Length (entry count) of an axis.
axis_vector(): Entry-name vector for an axis.
vectors_set(): Names of vectors on an axis, sorted.
matrices_set(): Names of matrices for an axis pair, sorted.

Writers

add_axis(): Add a new axis.
delete_axis(): Delete an axis (and all vectors / matrices that depend on it).
set_scalar(): Set a scalar.
delete_scalar(): Delete a scalar.
set_vector(): Set a vector on an axis.
delete_vector(): Delete a vector on an axis.
set_matrix(): Set a matrix indexed by a pair of axes.
delete_matrix(): Delete a matrix.
relayout_matrix(): Physically store the transposed layout of a matrix.
reorder_axes(): Reorder one or more axis entries in place.
reset_reorder_axes(): Reset any in-progress axis reorder.
empty_data(): Build a flat-keyed empty (or types) list for copy_all().

Query DSL (string form)

get_query(): Evaluate a query against a daf reader.
has_query(): Check whether a query can be evaluated against a daf without error.
parse_query(): Parse a query string into an AST (list of qop nodes).
canonical_query(): Canonicalise a query string.
is_axis_query(): Test whether a query yields an axis entry vector.
query_axis_name(): Return the axis name implied by a query, if any.
query_result_dimensions(): Return the dimensionality of a query's result.
query_requires_relayout(): Does evaluating this query require a matrix relayout (transpose)?
escape_value(): Escape a value for use as a query literal.
unescape_value(): Inverse of escape_value().

Query builders (DafrQuery)

DafrQuery(): Pipe-composable query object.
Abs(): Element-wise query operation: absolute value.
Clamp(): Element-wise query operation: clamp values to a range.
Convert(): Element-wise query operation: convert values to a given type.
Fraction(): Element-wise query operation: convert each entry to its fraction of the total.
Log(): Element-wise query operation: logarithm.
Round(): Element-wise query operation: round to nearest integer.
Significant(): Element-wise query operation: keep only "significant" entries.
Count(): Reduction query operation: count non-zero entries.
CountBy(): Count-by query operation.
GeoMean(): Reduction query operation: geometric mean.
GroupBy(): Group-by query operation.
GroupColumnsBy(): Group-columns-by query operation.
GroupRowsBy(): Group-rows-by query operation.
Max(): Reduction query operation: maximum.
Mean(): Reduction query operation: mean.
Median(): Reduction query operation: median.
Min(): Reduction query operation: minimum.
Mode(): Reduction query operation: most-frequent value.
Quantile(): Reduction query operation: quantile.
ReduceToColumn(): Reduce-to-column query operation.
ReduceToRow(): Reduce-to-row query operation.
Std(): Reduction query operation: standard deviation.
StdN(): Reduction query operation: standard deviation normalised by the mean.
Sum(): Reduction query operation: sum.
Var(): Reduction query operation: variance.
VarN(): Reduction query operation: variance normalised by the mean.
Axis(): Axis selection query operation.
AsAxis(): As-axis query operation.
BeginMask(): Begin-mask query operation.
BeginNegatedMask(): Negated begin-mask query operation.
EndMask(): End-mask query operation.
IfMissing(): If-missing query operation.
IfNot(): If-not query operation.
LookupMatrix(): Lookup-matrix query operation.
LookupScalar(): Lookup-scalar query operation.
LookupVector(): Lookup-vector query operation.
Names(): Names query operation.
SquareColumnIs(): Square-column-is query operation.
SquareRowIs(): Square-row-is query operation.
AndMask(): And-mask query operation.
AndNegatedMask(): And-negated-mask query operation.
OrMask(): Or-mask query operation.
OrNegatedMask(): Or-negated-mask query operation.
XorMask(): Xor-mask query operation.
XorNegatedMask(): Xor-negated-mask query operation.
IsEqual(): Equal-to comparison query operation.
IsGreater(): Greater-than comparison query operation.
IsGreaterEqual(): Greater-than-or-equal comparison query operation.
IsLess(): Less-than comparison query operation.
IsLessEqual(): Less-than-or-equal comparison query operation.
IsMatch(): Regex-match query operation.
IsNotEqual(): Not-equal-to comparison query operation.
IsNotMatch(): Negated regex-match query operation.

DataFrames

get_dataframe(): Extract vectors on an axis as a data.frame.
get_dataframe_query(): Extract an axis-resolving query's result as a data.frame.
get_tidy(): Pivot axis vectors into a tidy long-format tibble.

AnnData interop

DafAnnData: Read-only AnnData-shaped facade over a Daf.
as_anndata(): One-shot factory for a DafAnnData facade.
h5ad_as_daf(): Load a Muon-style h5ad file into a memory_daf.
daf_as_h5ad(): Write a Daf to a Muon-style h5ad file.

Chains & views

chain_reader(): Create a read-only chain of DafReaders.
chain_writer(): Create a chain of DafReaders with a final DafWriter.
viewer(): Construct a ViewDaf over a base daf.
ViewDaf(): A read-only view over a base daf.
ReadOnlyChainDaf(): Read-only chain of DafReaders.
WriteChainDaf(): Write chain of DafReaders with a final DafWriter.
VIEW_ALL_AXES: Default view spec: expose all axes as-is.
VIEW_ALL_DATA: Default view spec: expose all data (scalars + vectors + matrices) as-is.
VIEW_ALL_MATRICES: Default view spec: expose all matrices as-is.
VIEW_ALL_SCALARS: Default view spec: expose all scalars as-is.
VIEW_ALL_VECTORS: Default view spec: expose all vectors as-is.

Contracts

Contract(): A contract describing a computation's inputs and outputs.
ContractDaf(): Contract-enforcing daf wrapper
contract_scalar() contract_vector() contract_matrix(): Contract entry constructors
tensor_contract(): Tensor-contract record.
axis_contract(): Axis-contract record.
function_contract(): Retrieve the contract bound to a wrapped computation.
create_contract(): Construct a Contract() from typed per-category argument lists.
contract_description(): Render a contract as a human-readable multi-line string.
contract_docs(): Render a Contract() as human-readable documentation.
contractor(): Wrap a daf with a contract for a computation
verify_input() verify_output(): Verify contract inputs / outputs
verify_contract(): Single-pass contract verification (input + output).
merge_contracts(): Merge two contracts.
RequiredInput OptionalInput CreatedOutput GuaranteedOutput OptionalOutput: Contract expectation constants

Computations & adapters

computation(): Wrap a function so a contract is enforced on every call.
adapter(): Apply a computation to a renaming view of a DafWriter.

Mmap readers

mmap_dgCMatrix(): Construct a dgCMatrix whose x, i, p slots are ALTREP mmap views.
mmap_real() mmap_int() mmap_lgl(): Open a binary file as a read-only mmap-backed vector.

Op registry

register_eltwise(): Register an eltwise operation.
register_reduction(): Register a reduction operation.
registered_eltwise(): List registered eltwise operations.
registered_reductions(): List registered reduction operations.
get_eltwise(): Retrieve a registered eltwise operation by name.
get_reduction(): Retrieve a registered reduction operation by name.

Fast kernels

OpenMP-parallel kernels callable directly on in-memory data, bypassing the DSL overhead. Intended for downstream packages with their own session-level caches.

fast_log(): Parallel log(x + eps) / log(base) on dense numeric input.
top_k_per_col(): Per-column bounded top-K with OpenMP.

Handlers

register_dafr_handler(): Register a handler for an action category.
inefficient_action_handler(): Register a handler for the "inefficient" action category.
ERROR_HANDLER WARN_HANDLER IGNORE_HANDLER: Inefficient-action handler constants.

Threads

OpenMP thread-count control. Kernels honour the value set via set_num_threads() (or the dafr.num_threads option). CRAN-like environments are auto-capped to two cores on package load.

set_num_threads() get_num_threads(): Control OpenMP thread count for dafr kernels.

Version counters

axis_version_counter(): Per-axis version counter.
vector_version_counter(): Per-vector version counter.
matrix_version_counter(): Per-matrix version counter.
MEMORY_DATA MAPPED_DATA QUERY_DATA: Cache-group constants.

Group helpers

compact_groups(): Renumber group indices to be dense in 1..N.
collect_group_members(): Invert group-index assignment into per-group entry lists.
group_names(): Deterministic names for groups of axis entries.

Example data

example_cells_daf(): Load cells example data into a MemoryDaf.
example_metacells_daf(): Load metacells example data into a MemoryDaf.
example_chain_daf(): Load a chain of cells and metacells example data.

Copy & concat

copy_all(): Copy everything from one daf to another.
copy_axis(): Copy an axis (its entries) from one daf to another.
copy_vector(): Copy a vector from one daf to another.
copy_matrix(): Copy a matrix from one daf to another.
copy_scalar(): Copy a scalar from one daf to another.
copy_tensor(): Copy a tensor (set of per-main-axis-entry matrices) between dafs.
concatenate(): Concatenate multiple dafs along one or more axes.
files_to_zarr(): Convert a files_daf directory to a zarr_daf directory.
zarr_to_files(): Convert a zarr_daf directory to a files_daf directory.
ALL_AXES: Wildcard for all axes.
ALL_MATRICES: Wildcard for all matrices.
ALL_SCALARS: Wildcard for all scalars.
ALL_VECTORS: Wildcard for all vectors.
MERGE_SKIP MERGE_LAST_VALUE MERGE_COLLECT_AXIS: Merge action constants for concatenate().

Zarr store backend

ZarrStore() DirStore() DictStore() MmapZipStore(): Zarr v2 Store abstract base class.
new_dict_store(): Create an in-memory Zarr v2 store.
new_dir_store(): Create a directory-backed Zarr v2 store.
new_mmap_zip_store(): Create a zip-archive-backed Zarr v2 store (mmap, append-only).
new_http_store(): Construct a read-only ZarrStore over HTTP.
HttpStore(): Read-only HTTP-served Zarr store.
store_get_bytes() store_set_bytes() store_delete() store_exists() store_list(): Zarr v2 store generics.

Reconstruction & completion

open_daf(): Open a Daf store by URI or path.
complete_daf(): Reopen a persistent chain from disk.
complete_chain(): Create a persistent chain by linking new_daf to a base_daf.
reconstruct_axis(): Promote an implicit property to an explicit axis.