def auto_read_picture_move_to_standard_without_label(file_list,
width,
height,
outdir=r'./temp',
channel=3,
num_threads=2,
stay=1):
if (os.path.exists(outdir)) is False:
os.makedirs(outdir)
for one in file_list:
fh.image_change_to(
one, outdir + "/" + fh.get_path_file_basename(one) + ".png", width,
height)
train_listtest = fh.get_files_by_types(outdir, '')
filename_batch, image_batch = read_pic(train_listtest, width, height,
channel, num_threads,
len(train_listtest),
len(train_listtest))
filename, image = get_file_batchlist(filename_batch, image_batch)
filenamebatchlist = [
fh.get_path_file_basename(sh.getstringexep(r"'(.*?)'", one)[0])
for one in filename
]
if (stay == 0):
fh.finish_files_toproject()
return filenamebatchlist, image
def init_load(ignorePar=False):
if (testmode or onlytest or ignorePar):
fh.remove_file(outpredictcsv)
fileslist = None
if onlytest or ignorePar:
fileslist = fh.get_files_by_types(test_data, ".csv")
elif testmode:
fileslist = fh.get_files_by_types(
test_data, ".csv")[starttest:starttest + testnum]
df = pd.DataFrame()
filename = []
root_node = []
triggername = []
fileslist.sort(key=lambda elem: int(fh.get_path_file_basename(elem)))
for subfile in fileslist:
predict = predict_csv(subfile, 0.5)
# fh.appendfile(outpredictcsv,subfile+","+str(predict),startchar="\n")
filename.append(fh.get_path_file_basename(subfile))
try:
rootname = "node_" + str(predict[0][1])
root_node.append(rootname)
triggername.append("主机" + rootname + " " + predict[0][2])
except IndexError:
root_node.append("0")
triggername.append("0")
df['filename'] = filename
df['root_node'] = root_node
df['triggername'] = triggername
df.to_csv(outpredictcsv,
encoding="utf_8_sig",
header=True,
index=False)
def imagebatchs_enlarge_booleanimage_cnninfo(fileslist,
imagebatchs,
twidth,
theight,
width,
height,
outdir=None,
type=".png",
picmode="L",
gaussianblur=False,
mfilter=(5, 5),
mvalue=0.0):
if (outdir):
pout = outdir
else:
pout = fh.get_path_file_subpath(fileslist[0])
for one in range(len(fileslist)):
fh.array_to_image(enlarge_booleanimage_cnninfo(
fh.out_booleanimage_pre(np.multiply(imagebatchs[one], 255)),
twidth, theight),
pout + "/" +
fh.get_path_file_basename(fileslist[one]) + type,
picmode=picmode)
return fh.imagedir_to_arrays_tfable(pout,
type,
True,
width,
height,
gaussianblur=gaussianblur,
mfilter=mfilter,
mvalue=mvalue)
def bayesAlgorithm(trainPath, testPath, tfidfspace_out_arr_path,
tfidfspace_out_word_path, testspace_out_arr_path,
testspace_out_word_apth):
trainSet = readBunch(trainPath)
testSet = readBunch(testPath)
clf = MultinomialNB(alpha=0.001).fit(trainSet.tdm, trainSet.label)
# alpha:0.001 alpha 越小，迭代次数越多，精度越高
# print(shape(trainSet.tdm)) #输出单词矩阵的类型
# print(shape(testSet.tdm))
'''处理bat文件'''
tfidfspace_out_arr = str(trainSet.tdm) # 处理
tfidfspace_out_word = str(trainSet)
saveFile(tfidfspace_out_arr_path, tfidfspace_out_arr) # 矩阵形式的train_set.txt
saveFile(tfidfspace_out_word_path, tfidfspace_out_word) # 文本形式的train_set.txt
testspace_out_arr = str(testSet)
testspace_out_word = str(testSet.label)
saveFile(testspace_out_arr_path, testspace_out_arr)
saveFile(testspace_out_word_apth, testspace_out_word)
'''处理结束'''
predicted = clf.predict(testSet.tdm)
# total = len(predicted)
# rate = 0
numlist=[]
for flabel, fileName, expct_cate in zip(testSet.label, testSet.filenames, predicted):
# if flabel != expct_cate:
# rate += 1
#print(fileName, ":实际类别：", flabel, "-->预测类别：", expct_cate)
# print(fileName, "-->预测类别：", expct_cate)
numlist.append(int(fh.get_path_file_basename(fileName)))
# print("erroe rate:", float(rate) * 100 / float(total), "%")
return [predicted[one] for one in np.argsort(numlist)]
def auto_read_picture_move_to_standard(file_list,
width,
height,
labellist,
outdir=r'./temp',
channel=3,
num_threads=2,
stay=1,
type=".jpg",
default=None,
nodefault=False):
filelabelmap = dict(
zip([fh.get_path_file_basename(one) for one in file_list], labellist))
if (os.path.exists(outdir)) is False:
os.makedirs(outdir)
for one in file_list:
fh.image_change_to(
one, outdir + "/" + fh.get_path_file_basename(one) + type, width,
height)
train_listtest = fh.get_files_by_types(outdir, '')
# print(train_listtest)
# print(file_list)
filename_batch, image_batch = read_pic(train_listtest, width, height,
channel, num_threads,
len(train_listtest),
len(train_listtest))
filename, image = get_file_batchlist(filename_batch, image_batch)
filenamebatchlist = [
fh.get_path_file_basename(sh.getstringexep(r"'(.*?)'", one)[0])
for one in filename
]
finallabel = [
filelabelmap.get(one, default) for one in filenamebatchlist
if (not filelabelmap.get(one, default)) and nodefault
]
if (stay == 0):
fh.finish_files_toproject()
tf.reset_default_graph()
return filenamebatchlist, image, finallabel
def auto_read_picture(file_list,
width,
height,
labellist,
channel=3,
num_threads=2):
filelabelmap = dict(zip(file_list, labellist))
filename_batch, image_batch = read_pic(file_list, width, height, channel,
num_threads, len(file_list),
len(file_list))
filename, image = get_file_batchlist(filename_batch, image_batch)
filenamebatchlist = [
fh.get_path_file_basename(sh.getstringexep(r"'(.*?)'", one)[0])
for one in filename
]
finallabel = [filelabelmap[one] for one in filenamebatchlist]
return filenamebatchlist, image, finallabel