hello, I am trying to execute code with model cnn, I have the following error
File “demo.py”, line 49, in demo
imageio.imwrite(args.outputPath+args.inputPath("/")[-1])
TypeError: ‘str’ object is not callable
if you have a solution, thank you
I guess args.inputPath is set to a string containing the path, so “calling” it won’t work:
inputPath = 'PATH'
inputPath("/")
> TypeError: 'str' object is not callable
What are you trying to do with the args.inputPath there?
I’m trying to make the project work:
I want to annotate an image, so I have two directory input and output
from pipeline import Pipeline
from config import Config
import os
import glob
import imageio
import cv2
import argparse
import numpy as np
def get_parser():
parser = argparse.ArgumentParser('demo')
parser.add_argument('--inputPath', '-i', required=True, default=r"C:\Users\ETTAQI\Documents\STAGE\imagecode\img2.png", help='path to input image')
parser.add_argument('--outputPath', '-o', required=True, default=r"C:\Users\ETTAQI\Documents\STAGE\imagecode", help='path to input image')
parser.add_argument('--configPath', '-c', required=True, default=r"C:\Users\ETTAQI\Documents\STAGE\PathoNet-master1\configs\demo.json", help='path to input image')
return parser
def visualizer(img,points):
r=1
colors=[
(255,0,0),
(0,255,0),
(0,0,255)
]
image=np.copy(img)
for p in points:
x,y,c=p[0],p[1],p[2]
cv2.circle(image, (int(x), int(y)), int(r), colors[int(c)], 2)
return image
def demo(args=None):
parser = get_parser()
args = parser.parse_args(args)
conf=Config()
conf.load(args.configPath)
pipeline=Pipeline(conf)
if os.path.isdir(args.inputPath):
data = [args.inputPath+"/"+f for f in os.listdir(args.inputPath) if ‘.jpg’ in f]
for d in data:
print(d)
img=imageio.imread(d)
pred=pipeline.predict(img)
output=visualizer(img,pred)
imageio.imwrite(args.outputPath+d.split("/")[-1],output)
else:
img=imageio.imread(args.inputPath)
#imageio.imwrite(args.outputPath+args.inutPath("/")[-1])
if name == “main”:
demo()
You are still trying to call inputPath, which won’t work.
Maybe you want to use inputPath.split('/') instead, as you cannot use inputPath as a function?
yes it works with that but i still have an error
File “demo.py”, line 52, in demo
imageio.imwrite(args.outputPath+args.inputPath.split("/")[-1])
TypeError: imwrite() missing 1 required positional argument: ‘im’
après j’ai essayé ca : imageio.imwrite(args.outputPath+args.inputPath.split("/")[-1], 1024*1024, format=‘png’)
et j’ai comme erreur : ValueError: Image must be 2D (grayscale, RGB, or RGBA).
et je sais pas comment on peut changer l’image en RGB par exemple avec juste le répertoire
imageio.imwrite expects at least two input arguments: the path where the image should be stored, and the image itself (passed as a numpy array).
In your current code snippet you are calling imwrite with the path only without providing the actual image array.
thanks a lot for your answer
else:
img=imageio.imread(args.inputPath)
image0=img0.open(r"C:\Users\ETTAQI\Documents\STAGE\imagecode\img2.png")
image0_array=np.array(image0)
imageio.imwrite(args.inputPath.split("/")[-1], image0_array)
I used this, I opened the image and defined it as an array, does that convert the image to an array?
because I have an error in another file:
File “C:\Users\ETTAQI\Documents\STAGE\PathoNet-master1\utils.py”, line 15, in generator
batch=np.random.choice(self.dataList,size=self.batchSize,replace=False)
File “mtrand.pyx”, line 908, in numpy.random.mtrand.RandomState.choice
ValueError: ‘a’ cannot be empty unless no samples are taken
the code is 
class DataLoader:
def init(self,batchSize,inputShape,dataList,guaMaxValue):
self.inputShape=inputShape
self.batchSize=batchSize
self.dataList=dataList
self.guaMaxValue=guaMaxValue
def generator(self):
while(1):
batch=np.random.choice(self.dataList,size=self.batchSize,replace=False)
images=[]
labels=[]
for b in batch:
img=imread(b)
images.append(img)
temp=np.load(b.replace(".jpg",".npy")).astype(int)
np.place(temp,temp==255,self.guaMaxValue)
labels.append(temp)
images=np.array(images)
yield (np.array(images)/255.).astype(np.float32),np.array(labels)
No, I don’t think you can open the png directly and convert it to a numpy array.
I also don’t know what img0 is, but wouldn’t imread work?
it’s the library: from PIL import Image as img0,
thank you very much for your answer
now i am trying the following code:
class DataLoader:
def init(self,batchSize,inputShape,dataList,guaMaxValue):
self.inputShape=inputShape
self.batchSize=batchSize
self.dataList=dataList
self.guaMaxValue=guaMaxValue
def generator(self):
while(1):
batch=np.random.choice(self.dataList,size=self.batchSize,replace=False) #Génère un échantillon aléatoire à partir d’un tableau 1D donné
images=[]
labels=[]
for b in batch:
img=imread(b)
images.append(img)
temp=np.load(b.replace(".jpg",".npy")).astype(int)
np.place(temp,temp==255,self.guaMaxValue)
labels.append(temp)
images=np.array(images)
yield (np.array(images)/255.).astype(np.float32),np.array(labels)
but I have as an error:
batch=np.random.choice(self.dataList,size=self.batchSize,replace=True)
File “mtrand.pyx”, line 908, in numpy.random.mtrand.RandomState.choice
ValueError: ‘a’ cannot be empty unless no samples are taken
do you have a solution for me please?
thank you in advance
Based on the error message self.dataList is empty and cannot be used in np.random.choice:
a = np.random.randn(0)
np.random.choice(a, size=10, replace=True)
> ValueError: 'a' cannot be empty unless no samples are taken
Thank you for your answer