def run_verify(args):
setup_logging(args)
samples = tsinfer.SampleData.load(args.samples)
ts = tskit.load(args.tree_sequence)
progress_monitor = ProgressMonitor(enabled=args.progress, verify=True)
tsinfer.verify(samples, ts, progress_monitor=progress_monitor)
summarise_usage()
def verify_from_source(self, remove_leaves):
ts = msprime.simulate(15, recombination_rate=1, mutation_rate=2, random_seed=3)
samples = tsinfer.SampleData.from_tree_sequence(ts)
ancestors_ts = tsinfer.make_ancestors_ts(
samples, ts, remove_leaves=remove_leaves)
tsinfer.check_ancestors_ts(ancestors_ts)
for engine in [tsinfer.PY_ENGINE, tsinfer.C_ENGINE]:
final_ts = tsinfer.match_samples(samples, ancestors_ts, engine=engine)
tsinfer.verify(samples, final_ts)
def tsinfer_dev(
n,
L,
seed,
num_threads=1,
recombination_rate=1e-8,
error_rate=0,
engine="C",
log_level="WARNING",
precision=None,
debug=True,
progress=False,
path_compression=True,
):
np.random.seed(seed)
random.seed(seed)
L_megabases = int(L * 10**6)
# daiquiri.setup(level=log_level)
ts = msprime.simulate(
n,
Ne=10**4,
length=L_megabases,
recombination_rate=recombination_rate,
mutation_rate=1e-8,
random_seed=seed,
)
if debug:
print("num_sites = ", ts.num_sites)
assert ts.num_sites > 0
# ts = msprime.mutate(ts, rate=1e-8, random_seed=seed,
# model=msprime.InfiniteSites(msprime.NUCLEOTIDES))
samples = tsinfer.SampleData.from_tree_sequence(ts)
rho = recombination_rate
mu = 1e-3 # 1e-15
# num_alleles = samples.num_alleles(inference_sites=True)
# num_sites = samples.num_inference_sites
# with tsinfer.AncestorData(samples) as ancestor_data:
# t = np.sum(num_alleles) + 1
# for j in range(num_sites):
# for allele in range(num_alleles[j]):
# ancestor_data.add_ancestor(j, j + 1, t, [j], [allele])
# t -= 1
ancestor_data = tsinfer.generate_ancestors(samples,
engine=engine,
num_threads=num_threads)
ancestors_ts = tsinfer.match_ancestors(
samples,
ancestor_data,
engine=engine,
path_compression=True,
extended_checks=False,
precision=precision,
recombination_rate=rho,
mutation_rate=mu,
)
# print(ancestors_ts.tables)
# print("ancestors ts")
# for tree in ancestors_ts.trees():
# print(tree.draw_text())
# for site in tree.sites():
# if len(site.mutations) > 1:
# print(site.id)
# for mutation in site.mutations:
# print("\t", mutation.node, mutation.derived_state)
# for var in ancestors_ts.variants():
# print(var.genotypes)
# print(ancestors_ts.tables)
# ancestors_ts = tsinfer.augment_ancestors(samples, ancestors_ts,
# [5, 6, 7], engine=engine)
ts = tsinfer.match_samples(
samples,
ancestors_ts,
recombination_rate=rho,
mutation_rate=mu,
path_compression=False,
engine=engine,
precision=precision,
simplify=False,
)
print("num_edges = ", ts.num_edges)
# # print(ts.draw_text())
# for tree in ts.trees():
# print(tree.draw_text())
# for site in tree.sites():
# if len(site.mutations) > 1:
# print(site.id)
# for mutation in site.mutations:
# print("\t", mutation.node, mutation.derived_state)
# # print(ts.tables.edges)
# print(ts.dump_tables())
# simplified = ts.simplify()
# print("edges before = ", simplified.num_edges)
# new_ancestors_ts = insert_srb_ancestors(ts)
# ts = tsinfer.match_samples(samples, new_ancestors_ts,
# path_compression=False, engine=engine,
# simplify=True)
# for tree in ts.trees():
# print(tree.interval)
# print(tree.draw(format="unicode"))
# print(ts.tables.edges)
# for tree in ts.trees():
# print(tree.draw(format="unicode"))
tsinfer.verify(samples, ts)
def tsinfer_dev(n,
L,
seed,
num_threads=1,
recombination_rate=1e-8,
error_rate=0,
engine="C",
log_level="WARNING",
debug=True,
progress=False,
path_compression=True):
np.random.seed(seed)
random.seed(seed)
L_megabases = int(L * 10**6)
# daiquiri.setup(level=log_level)
ts = msprime.simulate(n,
Ne=10**4,
length=L_megabases,
recombination_rate=recombination_rate,
mutation_rate=1e-8,
random_seed=seed)
if debug:
print("num_sites = ", ts.num_sites)
assert ts.num_sites > 0
samples = tsinfer.SampleData.from_tree_sequence(ts)
ancestor_data = tsinfer.generate_ancestors(samples,
engine=engine,
num_threads=num_threads)
ancestors_ts = tsinfer.match_ancestors(samples,
ancestor_data,
engine=engine,
path_compression=True,
extended_checks=False)
ancestors_ts = tsinfer.augment_ancestors(samples,
ancestors_ts, [5, 6, 7],
engine=engine)
ts = tsinfer.match_samples(samples,
ancestors_ts,
path_compression=False,
engine=engine,
simplify=True)
# print(ts.tables.edges)
# print(ts.dump_tables())
# simplified = ts.simplify()
# print("edges before = ", simplified.num_edges)
# new_ancestors_ts = insert_srb_ancestors(ts)
# ts = tsinfer.match_samples(samples, new_ancestors_ts,
# path_compression=False, engine=engine,
# simplify=True)
# for tree in ts.trees():
# print(tree.interval)
# print(tree.draw(format="unicode"))
# print(ts.tables.edges)
# for tree in ts.trees():
# print(tree.draw(format="unicode"))
tsinfer.verify(samples, ts)