Datasets:
jblitzar
/
github-python

Dataset card Data Studio Files
xet
 Community
text
stringlengths
0
93.6k
cond3 = np.sum(sam_==idi)
if ol_rate * cond3 < cond1 and (cond1 > cond2 * ol_rate or cond1 < cond2 * 0.2):
partid1 = part_dicts[(idx, idf)]
vertices[partid1].append(partid0)
vertices[partid0].append(partid1)
for k,v in vertices.items():
vertices[k] = np.unique(v).tolist() # clear all edeges
visited = {}
for idx in range(len(vertices)):
visited[idx] = False
val_parts = {}
largest = [0,0]
find_connected_parts(vertices, visited, val_parts, mask_cents, largest)
### remove isolated parts with centers that are nearby to its isolated counterparts
cents = []
isolations, iso_cents = [], []
if isoproc:
for k,v in val_parts.items():
if len(vertices[v]) > 0:
cents.append(mask_cents[v][0].reshape(1,-1))
else:
isolations.append(v)
iso_cents.append(mask_cents[v][0].reshape(1,-1))
else:
for k,v in val_parts.items():
cents.append(mask_cents[v][0].reshape(1,-1))
# process isolated parts
if len(iso_cents) > 0:
val_parts_iso = {}
count_k = 0
for k in isolations:
val_parts_iso[count_k] = k
count_k += 1
for k,v in val_parts_iso.items():
cents.append(mask_cents[v][0].reshape(1,-1))
# record the number of non-isolated parts
solid_cents_num.append(np.max([len(cents) - len(val_parts_iso), 2]))
else:
solid_cents_num.append(len(cents))
cents = np.concatenate(cents, axis=0)
cents_lists.append(cents)
cents_tosave = {}
count = 0
for ol_code in overlap_lists.keys():
cent_ = cents_lists[count]
if cent_.shape[0] < 2 or cent_.shape[0] > parts_upbd or cent_.shape[0] < parts_lobd:
pass
elif count > 0:
cent_1 = cents_lists[count-1]
if cent_.shape[0]==cent_1.shape[0] and np.sum(np.abs(cent_1-cent_)) < 1e-6:
pass
else:
cents_tosave[ol_code] = cent_
cents_tosave[ol_code + '_sol'] = np.array([solid_cents_num[count]])
else:
cents_tosave[ol_code] = cent_
cents_tosave[ol_code + '_sol'] = np.array([solid_cents_num[count]])
count += 1
# in case no result meet the restrictions
if len(cents_tosave) < 1:
count = 0
for ol_code in overlap_lists.keys():
cent_ = cents_lists[count]
if cent_.shape[0] < 2:
pass
elif count > 0:
cent_1 = cents_lists[count-1]
if cent_.shape[0]==cent_1.shape[0] and np.sum(np.abs(cent_1-cent_)) < 1e-6:
pass
else:
cents_tosave[ol_code] = cent_
cents_tosave[ol_code + '_sol'] = np.array([solid_cents_num[count]])
else:
cents_tosave[ol_code] = cent_
cents_tosave[ol_code + '_sol'] = np.array([solid_cents_num[count]])
count += 1
np.savez(os.path.join('output', renderdir, name, 'sam_cents.npz'), **cents_tosave)
# <FILESEP>
#!/usr/bin/python
# -*- coding: utf-8 -*-
import tensorflow as tf
class TCNNConfig(object):
"""CNN配置参数"""