ParangonadaLoader

loader.alignment.parangonada.ParangonadaLoader()

Load a parangonada CSV export as one multimodal AlignmentBundle.

The loader discovers every performance under the dataset’s match/match_transkun/ directory, parses the shared score part.csv once, each performance’s ppart.csv and align.csv, and binds each recording’s .wav for its sample rate. It then assembles a single bundle with one shared score group and one group per performance, linked by cross-group :class:MatchClaim instances.

Usage follows the standard loader two-phase pattern:

loader.load(dataset_dir) — ingest the whole dataset directory.
loader.create_bundle() — assemble the AlignmentBundle.
loader.create_timeline(uid) / create_timelines() — retrieve individual timelines.

The loader reads existing alignments; it never runs an aligner.

Attributes

Name	Description
performer_keys	The discovered performer keys, in sorted (chronological) order.

Methods

Name	Description
create_bundle	Assemble the AlignmentBundle from the loaded dataset.
create_timeline	Return a single timeline by its uid.
create_timelines	Return all timelines: the two score timelines then each performer’s.
load	Ingest a whole parangonada dataset directory.

create_bundle

loader.alignment.parangonada.ParangonadaLoader.create_bundle()

Assemble the AlignmentBundle from the loaded dataset.

Returns

Name	Type	Description
	'AlignmentBundle'	An `AlignmentBundle` with one shared `"score"` group (two
	'AlignmentBundle'	logical timelines), one `perf:<key>` group per performance
	'AlignmentBundle'	(two physical timelines), and all cross-group MatchClaims.

Raises

Name	Type	Description
	`RuntimeError`	If `load()` has not been called yet.

create_timeline

loader.alignment.parangonada.ParangonadaLoader.create_timeline(
    uid=None,
    **kwargs,
)

Return a single timeline by its uid.

Parameters

Name	Type	Description	Default
uid	`str` \| None	A timeline uid — `"score:clt1"`, `"score:dlt1"`, or a performer timeline uid (`"perf:<key>:cpt1"` / `"perf:<key>:dpt1"`).	`None`

Raises

Name	Type	Description
	`KeyError`	If no timeline matches.
	`RuntimeError`	If `load()` has not been called yet.

create_timelines

loader.alignment.parangonada.ParangonadaLoader.create_timelines(id_pattern=None)

Return all timelines: the two score timelines then each performer’s.

Parameters

Name	Type	Description	Default
id_pattern	`str` \| None	Optional regex pattern to filter timeline IDs.	`None`

load

loader.alignment.parangonada.ParangonadaLoader.load(dataset_dir)

Ingest a whole parangonada dataset directory.

Parameters

Name	Type	Description	Default
dataset_dir	`str` \| `Path`	Path to the dataset directory (the one that contains the `match/` and `audio/` subdirectories).	required

Returns

Name	Type	Description
	`Self`	Self, for method chaining.

Raises

Name	Type	Description
	`FileNotFoundError`	If the directory or a required subdirectory is missing.