ReIdentify / app.py
anon
Update app.py
8aeaacd verified
#!/usr/bin/env python
# coding: utf-8
# In[1]:
import numpy as np
import numpy as np
import gradio as gr
import pandas as pd
import numpy as np
import torch
from torch import nn
from torch.nn import init, MarginRankingLoss
from torch.optim import Adam
from distutils.version import LooseVersion
from torch.utils.data import Dataset, DataLoader
from torch.autograd import Variable
import math
from transformers import AutoConfig, AutoModel, AutoTokenizer
import nltk
import re
import torch.optim as optim
from transformers import AutoModelForMaskedLM
import torch.nn.functional as F
import random
# In[2]:
# eng_dict = []
# with open('eng_dict.txt', 'r') as file:
# # Read each line from the file and append it to the list
# for line in file:
# # Remove leading and trailing whitespace (e.g., newline characters)
# cleaned_line = line.strip()
# eng_dict.append(cleaned_line)
# In[14]:
def greet(X, ny):
global eng_dict
if ny == 0:
rand_no = random.random()
tok_map = {2: 0.4363429005892416,
1: 0.6672580202327398,
4: 0.7476060740459144,
3: 0.9618703668504087,
6: 0.9701028532809564,
7: 0.9729244545819342,
8: 0.9739508754144756,
5: 0.9994508859743607,
9: 0.9997507867114407,
10: 0.9999112969650892,
11: 0.9999788802297832,
0: 0.9999831041838266,
12: 0.9999873281378701,
22: 0.9999957760459568,
14: 1.0000000000000002}
for key in tok_map.keys():
if rand_no < tok_map[key]:
num_sub_tokens_label = key
break
else:
num_sub_tokens_label = ny
tokenizer = AutoTokenizer.from_pretrained("microsoft/graphcodebert-base")
model = AutoModelForMaskedLM.from_pretrained("microsoft/graphcodebert-base")
model.load_state_dict(torch.load('model_26_2'))
model.eval()
X_init = X
X_init = X_init.replace("[MASK]", " [MASK] ")
X_init = X_init.replace("[MASK]", " ".join([tokenizer.mask_token] * num_sub_tokens_label))
tokens = tokenizer.encode_plus(X_init, add_special_tokens=False,return_tensors='pt')
input_id_chunki = tokens['input_ids'][0].split(510)
input_id_chunks = []
mask_chunks = []
mask_chunki = tokens['attention_mask'][0].split(510)
for tensor in input_id_chunki:
input_id_chunks.append(tensor)
for tensor in mask_chunki:
mask_chunks.append(tensor)
xi = torch.full((1,), fill_value=101)
yi = torch.full((1,), fill_value=1)
zi = torch.full((1,), fill_value=102)
for r in range(len(input_id_chunks)):
input_id_chunks[r] = torch.cat([xi, input_id_chunks[r]],dim = -1)
input_id_chunks[r] = torch.cat([input_id_chunks[r],zi],dim=-1)
mask_chunks[r] = torch.cat([yi, mask_chunks[r]],dim=-1)
mask_chunks[r] = torch.cat([mask_chunks[r],yi],dim=-1)
di = torch.full((1,), fill_value=0)
for i in range(len(input_id_chunks)):
pad_len = 512 - input_id_chunks[i].shape[0]
if pad_len > 0:
for p in range(pad_len):
input_id_chunks[i] = torch.cat([input_id_chunks[i],di],dim=-1)
mask_chunks[i] = torch.cat([mask_chunks[i],di],dim=-1)
vb = torch.ones_like(input_id_chunks[0])
fg = torch.zeros_like(input_id_chunks[0])
maski = []
for l in range(len(input_id_chunks)):
masked_pos = []
for i in range(len(input_id_chunks[l])):
if input_id_chunks[l][i] == tokenizer.mask_token_id: #103
if i != 0 and input_id_chunks[l][i-1] == tokenizer.mask_token_id:
continue
masked_pos.append(i)
maski.append(masked_pos)
input_ids = torch.stack(input_id_chunks)
att_mask = torch.stack(mask_chunks)
outputs = model(input_ids, attention_mask = att_mask)
last_hidden_state = outputs[0].squeeze()
l_o_l_sa = []
sum_state = []
for t in range(num_sub_tokens_label):
c = []
l_o_l_sa.append(c)
if len(maski) == 1:
masked_pos = maski[0]
for k in masked_pos:
for t in range(num_sub_tokens_label):
l_o_l_sa[t].append(last_hidden_state[k+t])
else:
for p in range(len(maski)):
masked_pos = maski[p]
for k in masked_pos:
for t in range(num_sub_tokens_label):
if (k+t) >= len(last_hidden_state[p]):
l_o_l_sa[t].append(last_hidden_state[p+1][k+t-len(last_hidden_state[p])])
continue
l_o_l_sa[t].append(last_hidden_state[p][k+t])
for t in range(num_sub_tokens_label):
sum_state.append(l_o_l_sa[t][0])
for i in range(len(l_o_l_sa[0])):
if i == 0:
continue
for t in range(num_sub_tokens_label):
sum_state[t] = sum_state[t] + l_o_l_sa[t][i]
yip = len(l_o_l_sa[0])
# qw = []
er = ""
val = 0.0
for t in range(num_sub_tokens_label):
sum_state[t] /= yip
idx = torch.topk(sum_state[t], k=5, dim=0)[1]
probs = F.softmax(sum_state[t], dim=0)
wor = [tokenizer.decode(i.item()).strip() for i in idx]
cnt = 0
for kl in wor:
if all(char.isalpha() for char in kl):
# qw.append(kl.lower())
er+=kl
break
cnt+=1
val = val - torch.log(probs[idx[cnt]])
val = val/num_sub_tokens_label
vali = round(val.item(), 2)
# print(er)
# astr = ""
# for j in range(len(qw)):
# mock = ""
# mock+= qw[j]
# if (j+2) < len(qw) and ((mock+qw[j+1]+qw[j+2]) in eng_dict):
# mock +=qw[j+1]
# mock +=qw[j+2]
# j = j+2
# elif (j+1) < len(qw) and ((mock+qw[j+1]) in eng_dict):
# mock +=qw[j+1]
# j = j+1
# if len(astr) == 0:
# astr+=mock
# else:
# astr+=mock.capitalize()
er = er+" (with PLL value of: "+str(vali)+")"
return er, vali
def meet(X, ni):
if len(ni) == 0:
ni = 0
ni = int(ni)
if ni == 0:
print_str,vali = greet(X,ni)
elif ni == -1:
tot_pll = 100.00
print_str = ""
fin_out = "The highest confidence prediction is: "
add_out = ""
for r in range(6):
er, pll = greet(X, 6-r)
print_str+= er
print_str+='\n'
if (pll - tot_pll) > 0.1 and tot_pll < 1:
break
elif pll >= tot_pll:
continue
else:
add_out = er
tot_pll = pll
print_str= print_str+fin_out+add_out
else:
print_str,vali = greet(X,ni)
return print_str
title = "Rename a variable in a Java class"
description = """This model is a fine-tuned GraphCodeBERT model fine-tuned to output higher-quality variable names for Java classes. Long classes are handled by the
model. Replace any variable name with a "[MASK]" to get an identifier renaming.
In the input box for the number of tokens, specify a number from 1 to 6 indicating the number of tokens in the variable name. Feel free to test multiple values. Use 0 to get a randomly sampled number. Use -1 to get the best recommendation, although this will be slower
"""
ex = [["""import java.io.*;
public class x {
public static void main(String[] args) {
String f = "file.txt";
BufferedReader [MASK] = null;
String l;
try {
[MASK] = new BufferedReader(new FileReader(f));
while ((l = [MASK].readLine()) != null) {
System.out.println(l);
}
} catch (IOException e) {
e.printStackTrace();
} finally {
try {
if ([MASK] != null) [MASK].close();
} catch (IOException ex) {
ex.printStackTrace();
}
}
}
}""", -1], ["""import java.net.*;
import java.io.*;
public class s {
public static void main(String[] args) throws IOException {
ServerSocket [MASK] = new ServerSocket(8000);
try {
Socket s = [MASK].accept();
PrintWriter pw = new PrintWriter(s.getOutputStream(), true);
BufferedReader br = new BufferedReader(new InputStreamReader(s.getInputStream()));
String i;
while ((i = br.readLine()) != null) {
pw.println(i);
}
} finally {
if ([MASK] != null) [MASK].close();
}
}
}""", -1], ["""import java.io.*;
import java.util.*;
public class y {
public static void main(String[] args) {
String [MASK] = "data.csv";
String l = "";
String cvsSplitBy = ",";
try (BufferedReader br = new BufferedReader(new FileReader([MASK]))) {
while ((l = br.readLine()) != null) {
String[] z = l.split(cvsSplitBy);
System.out.println("Values [field-1= " + z[0] + " , field-2=" + z[1] + "]");
}
} catch (IOException e) {
e.printStackTrace();
}
}
}""", -1]]
# We instantiate the Textbox class
textbox = gr.Textbox(label="Type Java code snippet:", placeholder="replace variable with [MASK]", lines=10)
textbox1 = gr.Textbox(label="Number of tokens in name:", placeholder="0 for randomly sampled number of tokens and -1 for automatic number of token selection",lines=1)
gr.Interface(title = title, description = description, examples = ex, fn=meet, inputs=[
textbox, textbox1
], outputs="text").launch()
# In[ ]: