def test_build_hashCounter():
mh1 = MinHash(0, 21, scaled=1, track_abundance=True)
mh2 = MinHash(0, 21, scaled=1, track_abundance=True)
mh1.add_many((1, 2, 3, 4))
mh2.add_many((1, 2, 5))
true_res = Counter({1: 2, 2: 2, 3: 1, 4: 1, 5: 1})
ss1 = SourmashSignature(mh1)
ss2 = SourmashSignature(mh2)
counts = Counter()
hc = build_hashCounter([ss1, ss2], counts)
print("Hash Counter: ", hc)
assert hc == true_res
def test_drop_below_mincount_threshold():
mh1 = MinHash(0, 21, scaled=1, track_abundance=True)
mh2 = MinHash(0, 21, scaled=1, track_abundance=True)
mh1.add_many((1, 2, 3, 4))
mh2.add_many((1, 1, 2, 5))
ss1 = SourmashSignature(mh1)
ss2 = SourmashSignature(mh2)
counts = Counter()
hc = build_hashCounter([ss1, ss2], counts)
kept_hashes = drop_below_mincount(hc, 3)
true_kept = Counter({1: 3})
print("kept hashes: ", kept_hashes)
assert kept_hashes == true_kept
print(f"{hashval}:{ct}")
if ct < min_count:
counts.pop(hashval)
# write out hashes
# let's try building a sig. we will use this sig later to intersect with sample-specific sigs
new_mins = set(counts.keys())
print(len(new_mins))
with open(outhashes, "w") as out:
for hsh in new_mins:
out.write(str(hsh) + '\n')
if len(new_mins) > 0:
minhash = MinHash(
n=0, ksize=ksize, scaled=scaled
) # scaled=1 so we keep all (though these were previously at some other scaled val)
minhash.add_many(set(counts.keys()))
# write sig to file
sigobj = sourmash.SourmashSignature(
minhash,
name=f"aggregated_hashvals_above_{min_count}",
filename=f"generated with drop_unique_hashes.py")
sigobjs += [sigobj]
## this part only handles one output file -- doesn't take care of case with many ksizes/moltypes
with open(outsig, 'wt') as sigout:
sourmash.save_signatures(sigobjs, sigout)
#notify('wrote signature to {}', args.output)
# write out hashes to a text file
# this part is from
def main():
p = argparse.ArgumentParser()
p.add_argument('hashfile') # file that contains hashes
p.add_argument('-o', '--output', default=None,
help='file to output signature to')
p.add_argument('-k', '--ksize', default=None, type=int)
p.add_argument('--scaled', default=None, type=int)
p.add_argument('--num', default=None, type=int)
p.add_argument('--name', default='', help='signature name')
p.add_argument('--filename', default='',
help='filename to add to signature')
args = p.parse_args()
# check arguments.
if args.scaled and args.num:
error('cannot specify both --num and --scaled! exiting.')
return -1
if not args.ksize:
error('must specify --ksize')
return -1
if not args.output:
error('must specify --output')
return -1
# first, load in all the hashes
hashes = set()
for line in open(args.hashfile, 'rt'):
hashval = int(line.strip())
hashes.add(hashval)
if not hashes:
error("ERROR, no hashes loaded from {}!", args.hashfile)
return -1
notify('loaded {} distinct hashes from {}', len(hashes), args.hashfile)
# now, create the MinHash object that we'll use.
scaled = 0
num = 0
if args.scaled:
scaled = args.scaled
elif args.num:
num = args.num
else:
notify('setting --num automatically from the number of hashes.')
num = len(hashes)
# construct empty MinHash object according to args
minhash = MinHash(n=num, ksize=args.ksize, scaled=scaled)
# add hashes into!
minhash.add_many(hashes)
if len(minhash) < len(hashes):
notify("WARNING: loaded {} hashes, but only {} made it into MinHash.",
len(hashes), len(minhash))
if scaled:
notify("This is probably because of the scaled argument.")
elif args.num:
notify("This is probably because your --num is set to {}",
args.num)
if num > len(minhash):
notify("WARNING: --num set to {}, but only {} hashes in signature.",
num, len(minhash))
sigobj = sourmash.SourmashSignature(minhash, name=args.name,
filename=args.filename)
with open(args.output, 'wt') as fp:
sourmash.save_signatures([sigobj], fp)
notify('wrote signature to {}', args.output)