def upload():
if request.method == 'POST':
f = request.files['file']
if not (f and allowed_file(f.filename)):
return jsonify({
"error":
1001,
"error_msg":
"format error：请检查上传的图片类型，仅限于png、PNG、jpg、JPG、bmp"
})
# user_input = request.form.get("name")
basepath = os.path.dirname(__file__) # 当前文件所在路径
upload_path = os.path.join(
basepath, 'static/images',
secure_filename(f.filename)) # 注意：没有的文件夹一定要先创建，不然会提示没有该路径
# upload_path = os.path.join(basepath, 'static/images','test.jpg') #注意：没有的文件夹一定要先创建，不然会提示没有该路径
f.save(upload_path)
# 使用Opencv转换一下图片格式和名称
src = cv2.imread(upload_path)
cv2.imwrite(os.path.join(basepath, 'static/images', 'src.jpg'), src)
# 调用enhance()处理输出保存
dst = enhance(src)
cv2.imwrite(os.path.join(basepath, 'static/images', 'result.jpg'), dst)
# send_json()
return render_template('upload_ok.html', val1=time.time())
return render_template('upload.html')
def wraper_func(filename):
outfile = os.path.basename(filename).replace(search_criterium,
search_criterium + sufix)
ofile = os.path.join(data_output_dir, outfile)
out = enhance(inputFile=filename,
depth=depth,
model=model,
activation=activation,
ntype=ntype,
output=ofile)
out.define_network()
out.predict()
def select_image(func):
# grab a reference to the image panels
global panelA, panelB
maxsize = (512, 512)
# open a file chooser dialog and allow the user to select an input
# image
path = tkFileDialog.askopenfilename()
# ensure a file path was selected
if (len(path) > 0):
# os.system("script2.py "+str(path))
# subprocess.Popen("script2.py "+str(path), shell=True)
if (func == "redeye"):
input, output = redeye.redeye(path)
if (func == "enhance"):
input, output = enhance.enhance(path)
# convert the images to PIL format...
if (func == "redeye"):
input_pil = Image.fromarray(cv2.cvtColor(input, cv2.COLOR_BGR2RGB))
output_pil = Image.fromarray(
cv2.cvtColor(output, cv2.COLOR_BGR2RGB))
if (func == "enhance"):
input_pil = Image.fromarray(input)
output_pil = Image.fromarray(output)
input_pil.thumbnail(maxsize, Image.ANTIALIAS)
output_pil.thumbnail(maxsize, Image.ANTIALIAS)
# ...and then to ImageTk format
input_tk = ImageTk.PhotoImage(input_pil)
output_tk = ImageTk.PhotoImage(output_pil)
# if the panels are None, initialize them
if panelA is None or panelB is None:
# the first panel will store our original image
panelA = Label(image=input_tk)
panelA.image = input_tk
panelA.pack(side="left", padx=10, pady=10)
# while the second panel will store the edge map
panelB = Label(image=output_tk)
panelB.image = output_tk
panelB.pack(side="right", padx=10, pady=10)
# otherwise, update the image panels
else:
# update the pannels
panelA.configure(image=input_tk)
panelB.configure(image=output_tk)
panelA.image = input_tk
panelB.image = output_tk
def enhance_wrapper(sunpy_map,
depth=5,
model="keepsize",
activation="relu",
ntype="intensity"):
'''
This procedures run enhance https://github.com/cdiazbas/enhance (it works only from my fork https://github.com/lzivadinovic/enhance)
on input sunpy map
Check source code for explanation of code and input parameters
input: sunpy_map (sunpy.map) - input data set
output: sunpy.map - output data object (enhanced)
'''
# if rtype is spmap, there is no need for output, it will return sunpy.map object (lzivadinovic/enhance fork - master branch)
out = enhance(inputFile=sunpy_map,
depth=depth,
model=model,
activation=activation,
ntype=ntype,
output='1.fits',
rtype='spmap')
out.define_network()
return out.predict()
image = np.array(toimage(out))
images.append(image)
print('[info] done.\n')
torch.cuda.empty_cache()
return images
def save_images(images, folder):
import os
import cv2
dir = os.path.join(output_root, folder)
os.makedirs(dir, exist_ok=True)
for i, image in enumerate(images):
path = os.path.join(dir, f'{i}.jpg')
image_bgr = cv2.cvtColor(image, cv2.COLOR_RGB2BGR)
cv2.imwrite(path, image_bgr)
if __name__ == '__main__':
class_indices = [497]
number_images = 20
model = load_model()
model = model.to(device)
for index in class_indices:
images = generate_images(number_images, index, model, truncation=.9)
save_images(images, f'c{index}')
img_gen = enhance(images)
save_images(img_gen, f'c{index}-enhanced')
bbox = np.array([center-width/2, center+width/2])
ds = yt.load(filename,
midx_filename=midx,
bounding_box = bbox,
)
ds.domain_left_edge = ds.domain_left_edge.astype(np.float64)
ds.domain_right_edge = ds.domain_right_edge.astype(np.float64)
ds.domain_width = ds.domain_width.astype(np.float64)
ds.domain_center = ds.domain_center.astype(np.float64)
ad = ds.all_data()
Npix = 1024
image = np.zeros([Npix, Npix, 4], dtype='float64')
cbx = yt.visualization.color_maps.mcm.RdBu
col_field = ad['particle_velocity_z']
# Calculate image coordinates ix and iy based on what your view width is
#
ix = (ad['particle_position_x'] - ds.domain_left_edge[0])/ds.domain_width[0]
iy = (ad['particle_position_y'] - ds.domain_left_edge[1])/ds.domain_width[1]
#
col_field = (col_field - col_field.min()) / (col_field.mean() + 4*col_field.std() - col_field.min())
add_rgba_points_to_image(image, ix.astype('float64'), iy.astype('float64'), cbx(col_field))
#
yt.write_bitmap(enhance(image), 'zoom-output/zoom{:0>4d}.png'.format(rank))
print 'Splatted %i particles' % ad['particle_position_x'].size
if len(sys.argv) > 2:
nms= int(sys.argv[2])/2*3+1
nas= int(sys.argv[2])+1
else:
nms= 97
nas= 65
mv= linspace(-5,-2,nms)
mv= 10**mv
a= linspace(-3,-1,nas)
a= 10**a
S= zeros((nms,nas))
for ii in range(nms):
for jj in range(nas):
S[ii,jj]= log10(enhance(mv[ii],m,a[jj],b))
sys.stdout.write('\r'+str(ii*nas+jj+1)+'/'+str(nms*nas))
sys.stdout.flush()
sys.stdout.write('\n')
savefile=open(savefilename,'w')
data={'mv':mv,'a':a,'S':S}
pickle.dump(data,savefile)
#Plotting parameters
fig_width = 3.25 # width in inches
fig_height = 2.9 # height in inches
fig_size = [fig_width,fig_height]
center = np.array([-2505805.31114929, -3517306.7572399, -1639170.70554688]) + np.array([0, 0, offset])
width = 50.0e3 # 5 Mpc
bbox = np.array([center-width/2, center+width/2])
ds = yt.load(filename,
midx_filename=midx,
bounding_box = bbox,
)
ds.domain_left_edge = ds.domain_left_edge.astype(np.float64)
ds.domain_right_edge = ds.domain_right_edge.astype(np.float64)
ds.domain_width = ds.domain_width.astype(np.float64)
ds.domain_center = ds.domain_center.astype(np.float64)
ad = ds.all_data()
Npix = 1024
image = np.zeros([Npix, Npix, 4], dtype='float64')
cbx = yt.visualization.color_maps.mcm.RdBu
col_field = ad['particle_velocity_z']
# Calculate image coordinates ix and iy based on what your view width is
ix = (ad['particle_position_x'] - ds.domain_left_edge[0])/ds.domain_width[0]
iy = (ad['particle_position_y'] - ds.domain_left_edge[1])/ds.domain_width[1]
col_field = (col_field - col_field.min()) / (col_field.mean() + 4*col_field.std() - col_field.min())
add_rgba_points_to_image(image, ix.astype('float64'), iy.astype('float64'), cbx(col_field))
yt.write_bitmap(enhance(image), 'splat{:0>4d}.png'.format(rank))
print 'Splatted %i particles' % ad['particle_position_x'].size
savefilename='1d.sav'
if os.path.exists(savefilename):
savefile=open(savefilename,'r+')
data=pickle.load(savefile)
else:
mv=.09
a=1/30.0
m= linspace(0,2,nsamples)
m= 10**m
b= array([.1,.01,.001,.0001,.00001])
S= zeros((nsamples,nb))
for ii in range(nsamples):
for jj in range(nb):
S[ii,jj]= enhance(mv,m[ii],a,b[jj])
print str(ii*nb+jj)+'/'+str(nb*nsamples)
savefile=open(savefilename,'w')
data={'m':m,'S':S}
pickle.dump(data,savefile)
for jj in range(nb):
loglog(data['m'],data['S'][:,jj])
axis([1,100,1,10**6])
xlabel('m (TeV)')
ylabel('S')
savefig("enhancement.eps",format='eps')
ds = yt.load(
filename,
midx_filename=midx,
bounding_box=bbox,
)
ds.domain_left_edge = ds.domain_left_edge.astype(np.float64)
ds.domain_right_edge = ds.domain_right_edge.astype(np.float64)
ds.domain_width = ds.domain_width.astype(np.float64)
ds.domain_center = ds.domain_center.astype(np.float64)
ad = ds.all_data()
Npix = 1024
image = np.zeros([Npix, Npix, 4], dtype='float64')
cbx = yt.visualization.color_maps.mcm.RdBu
col_field = ad['particle_velocity_z']
# Calculate image coordinates ix and iy based on what your view width is
ix = (ad['particle_position_x'] - ds.domain_left_edge[0]) / ds.domain_width[0]
iy = (ad['particle_position_y'] - ds.domain_left_edge[1]) / ds.domain_width[1]
col_field = (col_field - col_field.min()) / (
col_field.mean() + 4 * col_field.std() - col_field.min())
add_rgba_points_to_image(image, ix.astype('float64'), iy.astype('float64'),
cbx(col_field))
yt.write_bitmap(enhance(image), 'splat{:0>4d}.png'.format(rank))
print 'Splatted %i particles' % ad['particle_position_x'].size
filename,
midx_filename=midx,
bounding_box=bbox,
)
ds.domain_left_edge = ds.domain_left_edge.astype(np.float64)
ds.domain_right_edge = ds.domain_right_edge.astype(np.float64)
ds.domain_width = ds.domain_width.astype(np.float64)
ds.domain_center = ds.domain_center.astype(np.float64)
ad = ds.all_data()
Npix = 1024
image = np.zeros([Npix, Npix, 4], dtype='float64')
cbx = yt.visualization.color_maps.mcm.RdBu
col_field = ad['particle_velocity_z']
# Calculate image coordinates ix and iy based on what your view width is
#
ix = (ad['particle_position_x'] - ds.domain_left_edge[0]) / ds.domain_width[0]
iy = (ad['particle_position_y'] - ds.domain_left_edge[1]) / ds.domain_width[1]
#
col_field = (col_field - col_field.min()) / (
col_field.mean() + 4 * col_field.std() - col_field.min())
add_rgba_points_to_image(image, ix.astype('float64'), iy.astype('float64'),
cbx(col_field))
#
yt.write_bitmap(enhance(image), 'zoom-output/zoom{:0>4d}.png'.format(rank))
print 'Splatted %i particles' % ad['particle_position_x'].size
