def __read_xml_by_slide__(xml_path,
slide_path,
slide_batch,
pro_method,
image_method,
zoom,
format_trans=None,
is_positive=True,
sub_class=None,
is_hard=False):
# get slide format
slide_format = slide_path[slide_path.rfind('.') + 1:]
slide_proxy = srf.get_proxy(slide_format)
# construct slide info
try:
slide_proxy.open(slide_path)
seg = '\\' # 默认路径分割符
if slide_path.find(seg) == -1:
seg = '/'
slide_name = slide_path[slide_path.rfind(seg) + 1:]
slide_path = slide_path[:slide_path.rfind(seg)]
slide_path = slide_path.replace('/', '\\')
mpp = slide_proxy.mpp()
# modify person's name and last modify time
m_name = 'csh'
m_time = time.strftime('"%Y-%m-%d %I:%M:%S"',
time.localtime(os.stat(xml_path).st_ctime))
modify_info = ModifyInfo(m_name, m_time)
file_permission = None # Tentatively None
md5 = mmd5(xml_path)
slide_info = SlideInfo(slide_path,
slide_name,
slide_batch,
pro_method,
image_method,
mpp,
zoom,
slide_format,
format_trans,
is_positive,
is_hard,
width=slide_proxy.width(),
height=slide_proxy.height(),
sub_class=sub_class,
bounds_x=slide_proxy.boundsx(),
bounds_y=slide_proxy.boundsy(),
modify_info=modify_info,
file_permission=file_permission,
md5_info=md5)
slide_proxy.close()
except:
raise SlideError("slide read error!")
annos, _ = __read_xml_by_path__(xml_path)
return slide_info, annos
def get_slide_info(slide_path, slide_format):
slide_path = slide_path.replace('\\', '/')
slide_name = slide_path[slide_path.rfind('/') + 1:]
slide_path = slide_path[:slide_path.rfind('/')]
pro_method = 'Non-BD'
image_method = slide_path.split('/')[2]
slide_group = slide_path.split('/')[3]
if slide_path.find('Not_BD') != -1:
pro_method = 'Non-BD'
zoom = '20x'
is_positive = 'Yes'
if slide_path.find('Neg') != -1 or slide_path.find('neg') != -1:
is_positive = 'No'
mpp, width, height, bounds_x, bounds_y = -1, -1, -1, -1, -1
try:
sr = srf.get_proxy(slide_format)
sr.open(os.path.join(slide_path, slide_name))
mpp = sr.mpp()
if float(mpp) < 0.2:
zoom = '40x'
width = sr.width()
height = sr.height()
bounds_x = sr.boundsx()
bounds_y = sr.boundsy()
sr.close()
except:
read_slide_log.write(
os.path.join(slide_path, slide_name) + '\t read exception\n')
print(os.path.join(slide_path, slide_name), ' read exception')
zoom = ''
print(slide_path, '---', slide_name, ' ', mpp)
slide = SlideInfo(slide_path, slide_name, slide_group, pro_method,
image_method, mpp, zoom, slide_format, 'NULL',
is_positive, '', width, height, '', bounds_x, bounds_y)
# slide.show_info()
return slide
xml_items = os.listdir(detect_path)
update_sql_flag = True
save_img_flag = True
# WNLO's Shengfuyou_3th and Shengfuyou_5th
for xml_item in xml_items:
xmls = os.listdir(os.path.join(detect_path, xml_item))
xmls = [x for x in xmls if x.find('.json') != -1]
pro_method = 'BD'
image_method = 'BD'
slide_group = xml_item
is_positive = '%'
zoom = '20x'
slide_format = 'srp'
slide_read = srf.get_proxy(slide_format)
for xk, xml in enumerate(xmls):
name = xml[:xml.find('.json')]
# handle detection xml
c_xml_path = os.path.join(detect_path, xml_item, xml)
json_annos = read_detection_json(c_xml_path)
# generate anno by json
print('(%d/%d)' % (xk + 1, len(xmls)), c_xml_path, len(json_annos))
# query slide info
slide_list = query_slide_info(pro_method, image_method, slide_group,
is_positive, slide_format,
name + '.' + slide_format, zoom)
assert len(slide_list) == 1
slide_info = slide_list[0]
pro_method, image_method, slide_group;'
slide_batches = sql_proxy.execute_query(sql)
log = open(os.path.join(anno_imgs_path, 'log.txt'), 'w')
for slide_batch in slide_batches[2:4]:
slides = query_slide_info(pro_method=slide_batch[0],
image_method=slide_batch[1],
slide_group=slide_batch[2],
is_positive=slide_batch[3],
zoom=slide_batch[4])
print(slide_batch, len(slides))
if len(slides) <= 0:
continue
slide_reader = srf.get_proxy(slides[0].slide_format())
item = '%s__%s__%s__%s__%s' % (slide_batch[0], slide_batch[1],
slide_batch[2], 'Pos' if slide_batch[3]
== 'Yes' else 'Neg', slide_batch[4])
log.write(item + '\t' + str(len(slides)) + '\n')
cpath = os.path.join(anno_imgs_path, item)
if not os.path.exists(cpath):
os.makedirs(cpath)
for slide in slides:
annos = query_annos(slide.sid())
if len(annos) <= 0:
continue
# slide_reader = srf.get_proxy(slide.slide_format())
# slide_reader.open(os.path.join(slide.slide_path(), slide.slide_name()))
recheck_list = read_recheck()
xml_path = 'O:/TransSrp/SZSQ_originaldata/Model3Labelfiles/xml'
xml_items = os.listdir(xml_path)
print(xml_items)
for k, xml_item in enumerate(xml_items):
xmls = os.listdir(os.path.join(xml_path, xml_item))
xmls = [x for x in xmls if x.find('.xml') != -1]
# print(xml_item)
us = xml_item.split('_')
im = 'SZSQ'
sf = 'sdpc'
zoom = '40x'
slide_group = us[0] + '_' + us[1]
slide_read = srf.get_proxy(sf)
# set model3_top path
model3_top50_path = 'H:/model3/reco_top50/' + slide_group
model2_top20_path = 'L:/GXB/model2_recon/SZSQ_originaldata/Shengfuyou_8th/positive/all'
for xk, xml in enumerate(xmls):
# print('%d/%d' % (xk, len(xmls)), im, sf, zoom, slide_group, xml)
annos, _ = read_xml_by_path(os.path.join(xml_path, xml_item, xml))
print(xml_item, xml, len(annos))
if xk > 3:
break
continue
name = xml[: xml.find('.')]
slide_list = query_slide_info('Non-BD', im, slide_group, 'Yes', sf,
name + '.' + sf, zoom)
assert len(slide_list) == 1