sample_id.ann.ann module¶

class sample_id.ann.ann.Matcher(metadata)[source]¶

Bases: abc.ABC

Nearest neighbor matcher that may use one of various implementations under the hood.

add_fingerprint(fingerprint, dedupe=True)[source]¶

Add a Fingerprint to the matcher.

Return type: Matcher

add_fingerprints(fingerprints, **kwargs)[source]¶

Add Fingerprints to the matcher.

Return type: Matcher

can_add_fingerprint(fingerprint)[source]¶

Check if fingerprint can be added to matcher.

Return type: bool

classmethod create(sr=None, hop_length=None, **kwargs)[source]¶

Create an instance, pass any kwargs needed by the subclass.

Return type: Matcher

filter_matches(matches, abs_thresh=0.25, ratio_thresh=None, cluster_dist=4.0, cluster_size=2, match_orientation=True, ordered=False, cluster_filter=None)[source]¶

Return type: List[Cluster]

find_samples(fp, k=1, abs_thresh=0.25, ratio_thresh=None, cluster_dist=20.0, cluster_size=2, match_orientation=True, ordered=False, cluster_filter=None)[source]¶

Return type: Result

classmethod from_fingerprint(fp, **kwargs)[source]¶

Useful for determining metadata for the Matcher based on the data being added.

Return type: Matcher

classmethod from_fingerprints(fingerprints, **kwargs)[source]¶

My data is small, just create and train the entire matcher.

Return type: Matcher

abstract init_model()[source]¶

Initialize the model.

Return type: Any

classmethod load(filepath, blocksize=10485760, threads=None, **kwargs)[source]¶

Load a matcher from disk.

Return type: Matcher

abstract load_model(filepath, **kwargs)[source]¶

Load this matcher’s model from disk.

Return type: Any

abstract nearest_neighbors(fp, k=1)[source]¶

Fetch nearest neighbors to this fingerprint’s keypoints.

Return type: Iterable[Match]

save(filepath, compress=True, compress_level=9, blocksize=10485760, threads=None, **kwargs)[source]¶

Save this matcher to disk.

Return type: str

abstract save_model(filepath, **kwargs)[source]¶

Save this matcher’s model to disk.

Return type: str

unload()[source]¶

Unload things from memory and cleanup any temporary files.

Return type: None

class sample_id.ann.ann.MatcherMetadata(sr=None, hop_length=None, index_to_id=None, index_to_kp=None, **kwargs)[source]¶

Bases: object

Metadata for a Matcher object.

classmethod load(filepath)[source]¶

Load this matcher’s metadata from disk.

Return type: MatcherMetadata

save(filepath, compress=True)[source]¶

Save this matcher’s metadata to disk.

Return type: None