def __init__()

in src/similarity/index.py [0:0]

• 23 lines of code
• 6 McCabe index (conditional complexity)

    def __init__(self, data_dir):
        
        self.idx = Zappos50kIndex.IDX_STRUCTURE        
        self.pdata = []
        self.data_dir = data_dir+"/"
            
        i= 0
        for category in os.listdir(data_dir):

            cid = int(hashlib.sha256(category.encode('utf-8')).hexdigest(), 16) % 10**9
            print(category+": id: "+str(cid)+" i: "+str(i))
            
            self.idx[category]["i"] = cid
            self.idx[category]["r"][0] = i

            for subcat in os.listdir(data_dir+"/"+category):
            
                scid = int(hashlib.sha256(subcat.encode('utf-8')).hexdigest(), 16) % 10**9
                print("  "+subcat+": id: "+str(scid)+" i: "+str(i))
        
                
                self.idx[category][subcat]["i"] = scid
                self.idx[category][subcat]["r"][0] = i

                for (root,dirs,files) in os.walk(data_dir+"/"+category+'/'+subcat):          
                    for f in files:

                        img_path = os.path.join(root.replace(data_dir+"/",''),f)
                        if img_path not in Zappos50kIndex.IMG_BLACK_LIST :
                            self.pdata.append(img_path) 
                            i= i+1

                            self.idx[category][subcat]["r"][1] = i-1                    
                        self.idx[category]["r"][1] = i-1