InferResult

InferResult#

The InferResult class represents the output of spatial or temporal inference on a tree sequence. It is returned by argscape.infer().

import argscape

result = argscape.infer(ts, method="fastgaia")
result.ts      # Tree sequence with inferred locations
result.info    # Metadata dictionary
result.method  # "fastgaia"

Class Definition#

@dataclass
class InferResult:
    ts: tskit.TreeSequence  # Tree sequence with inferred data
    info: Dict[str, Any]    # Inference metadata
    method: str             # Method name used

Attributes#

ts#

ts: tskit.TreeSequence

The tree sequence with inferred data. For spatial methods, internal nodes will have inferred locations stored in their individual metadata. For temporal methods (tsdate), node times will be updated.

Example:

result = argscape.infer(ts, method="fastgaia")

# Access inferred tree sequence
inferred_ts = result.ts

# Check a node's inferred location
for node_id in range(inferred_ts.num_nodes):
    node = inferred_ts.node(node_id)
    if node.individual != -1:
        individual = inferred_ts.individual(node.individual)
        print(f"Node {node_id}: location = {individual.location}")

info#

info: Dict[str, Any]

A dictionary containing method-specific metadata about the inference run. Contents vary by method but typically include:

  • method: The method name used

  • num_inferred_locations: Number of nodes with inferred locations (spatial methods)

  • inference_parameters: Parameters used for the inference

Example:

result = argscape.infer(ts, method="fastgaia")
print(result.info)
# {
#     'method': 'fastgaia',
#     'num_inferred_locations': 45,
#     'inference_parameters': {'weight_span': True, 'weight_branch_length': True},
#     ...
# }

method#

method: str

The name of the inference method that was used. One of:

  • Spatial: "midpoint", "fastgaia", "gaia-quadratic", "gaia-linear", "sparg", "spacetrees"

  • Temporal: "tsdate"

Usage Patterns#

Visualizing Results#

import argscape

# Run inference
result = argscape.infer(ts, method="fastgaia")

# Visualize the inferred tree sequence in 3D
viz = argscape.visualize(result.ts, mode="spatial_3d")
viz.show()

# Export visualization
viz.export("inferred_locations.png", dpi=300)

Chaining Inference and Visualization#

import argscape

# Infer spatial locations and visualize
result = argscape.infer(ts, method="fastgaia")
argscape.visualize(result.ts, mode="spatial_3d").show()

# Infer temporal information and visualize
result = argscape.infer(ts, method="tsdate", mutation_rate=1e-8)
argscape.visualize(result.ts, temporal_spacing="linear").show()

Saving Inferred Tree Sequences#

import argscape

# Run inference
result = argscape.infer(ts, method="fastgaia")

# Save the inferred tree sequence for later use
result.ts.dump("inferred_locations.trees")

Comparing Methods#

import argscape

methods = ["midpoint", "fastgaia", "sparg"]
results = {}

for method in methods:
    if argscape.available_methods().get(method, False):
        results[method] = argscape.infer(ts, method=method)
        print(f"{method}: {results[method].info.get('num_inferred_locations', 'N/A')} locations inferred")

See Also#

On this page