csv.ancestor_fit¶

Determine Closet Ancestor With Fit and Samples¶

Prototype¶

# at_cascade.csv.ancestor_fit
def ancestor_fit(
    fit_dir,
    job_table,
    predict_job_id,
    node_table,
    root_node_id,
    split_reference_table,
    root_split_reference_id,
    at_cascade_log_dict,
    allow_same_job,
) :
    assert type(fit_dir) == str
    assert type(job_table) == list
    assert type(predict_job_id) == int
    assert type(node_table) == list
    assert type( root_node_id ) == int
    assert type(split_reference_table) == list
    assert type( root_split_reference_id) == int
    assert type( at_cascade_log_dict ) == dict
    assert type( allow_same_job ) == bool
    assert type(predict_job_dir) == str
    if type(ancestor_job_dir) == str :
        assert predict_job_dir.startswith(ancestor_job_dir)
    else :
        assert ancestor_job_dir == None
    #
    return predict_job_dir, ancestor_job_dir

fit_dir¶

is the directory where the csv files are located.

job_table¶

is the job_table for this cascade.

predict_job_id¶

is the job_id for this prediction.

node_table¶

is the list of dict corresponding to the node table for this cascade.

root_node_id¶

is the node_id in the node table for the root node for this cascade. Note that csv version of the cascade does its sex split at this node.

root_split_reference_id¶

is the split_reference_id (sex id) for the root node of the cascade. The cascade can begin at female, both, or male.

at_cascade_log_dict¶

is a dictionary, with keys equal to job names, containing the log messages that have type at_cascade . The messages for each key are in the log table for the corresponding job.

allow_same_job¶

If this is true (false) the a job corresponding to the ancestor fit can be the same as the job for the prediction; i.e., the predict job is the closest ancestor job. Otherwise, the parent of the predict job is the closest ancestor job.

predict_job_dir¶

This is the directory, relative to the fit_dir, that corresponds to the predict_job_id . See get_database_dir .

ancestor_job_dir¶

This is the directory, relative to the fit_dir, that corresponds to the closest ancestor of predict_job_id that had a successful fit and posterior sampling. To be specific, sample: OK is in its log .