Spaces:
Runtime error
Runtime error
Commit
·
8fd005b
1
Parent(s):
2ba0142
Create app.py
Browse files
app.py
ADDED
|
@@ -0,0 +1,215 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import subprocess
|
| 2 |
+
import re
|
| 3 |
+
import pandas as pd
|
| 4 |
+
import plotly.express as px
|
| 5 |
+
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
|
| 6 |
+
from prettytable import PrettyTable
|
| 7 |
+
import streamlit as st
|
| 8 |
+
|
| 9 |
+
#st.title('Code Generation on the CoNaLa Dataset')
|
| 10 |
+
|
| 11 |
+
import subprocess
|
| 12 |
+
import re
|
| 13 |
+
import pandas as pd
|
| 14 |
+
import plotly.express as px
|
| 15 |
+
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
|
| 16 |
+
from prettytable import PrettyTable
|
| 17 |
+
|
| 18 |
+
#browser.gatherUsageStats=False
|
| 19 |
+
|
| 20 |
+
class CodeGenerator:
|
| 21 |
+
def __init__(self):
|
| 22 |
+
self.tokenizer = AutoTokenizer.from_pretrained("AhmedSSoliman/MarianCG-CoNaLa-Large")
|
| 23 |
+
self.model = AutoModelForSeq2SeqLM.from_pretrained("AhmedSSoliman/MarianCG-CoNaLa-Large")
|
| 24 |
+
|
| 25 |
+
def generate_code(self, nl_input):
|
| 26 |
+
input_ids = self.tokenizer.encode(nl_input, return_tensors="pt")
|
| 27 |
+
output_ids = self.model.generate(input_ids)
|
| 28 |
+
output_code = self.tokenizer.decode(output_ids[0], skip_special_tokens=True)
|
| 29 |
+
return output_code
|
| 30 |
+
|
| 31 |
+
|
| 32 |
+
|
| 33 |
+
|
| 34 |
+
|
| 35 |
+
def check_code(self, code):
|
| 36 |
+
with open("temp.py", "w") as f:
|
| 37 |
+
f.write(code)
|
| 38 |
+
result = subprocess.run(["flake8", "--count", "temp.py"], stdout=subprocess.PIPE, stderr=subprocess.PIPE)
|
| 39 |
+
output = result.stdout.decode()
|
| 40 |
+
error = result.stderr.decode()
|
| 41 |
+
|
| 42 |
+
|
| 43 |
+
return output
|
| 44 |
+
#return self._process_output(output, error)
|
| 45 |
+
|
| 46 |
+
def check_code_list(self, code_list):
|
| 47 |
+
output = ""
|
| 48 |
+
error = ""
|
| 49 |
+
for code in code_list:
|
| 50 |
+
with open("temp.py", "w") as f:
|
| 51 |
+
f.write(code)
|
| 52 |
+
result = subprocess.run(["flake8", "--count", "temp.py"], stdout=subprocess.PIPE, stderr=subprocess.PIPE)
|
| 53 |
+
output += result.stdout.decode()
|
| 54 |
+
error += result.stderr.decode()
|
| 55 |
+
|
| 56 |
+
return self._process_output(output, error)
|
| 57 |
+
|
| 58 |
+
def _process_output(self, output, error):
|
| 59 |
+
if output:
|
| 60 |
+
output_counts = self._get_error_counts(output)
|
| 61 |
+
self.show_variables_in_table(output_counts, output)
|
| 62 |
+
self.visualize_all_errors(output_counts)
|
| 63 |
+
self.visualize_error_types(output_counts)
|
| 64 |
+
|
| 65 |
+
return self._format_error_counts(output_counts)
|
| 66 |
+
else:
|
| 67 |
+
error_counts = self._get_error_counts(error)
|
| 68 |
+
self.show_variables_in_table(output_counts, output)
|
| 69 |
+
self.visualize_all_errors(error_counts)
|
| 70 |
+
self.visualize_error_types(error_counts)
|
| 71 |
+
|
| 72 |
+
return self._format_error_counts(error_counts)
|
| 73 |
+
|
| 74 |
+
def _get_error_counts(self, output):
|
| 75 |
+
error_counts = {}
|
| 76 |
+
error_messages = re.findall(r"temp.py:(\d+):\d+: (\w\d+)", output)
|
| 77 |
+
for message in error_messages:
|
| 78 |
+
error_type = message[1]
|
| 79 |
+
if error_type in error_counts:
|
| 80 |
+
error_counts[error_type] += 1
|
| 81 |
+
else:
|
| 82 |
+
error_counts[error_type] = 1
|
| 83 |
+
return error_counts
|
| 84 |
+
|
| 85 |
+
def _format_error_counts(self, error_counts):
|
| 86 |
+
error_message = "\n".join([f"{error_type}: {count}" for error_type, count in error_counts.items()])
|
| 87 |
+
return error_message
|
| 88 |
+
|
| 89 |
+
def visualize_all_errors(self, error_counts):
|
| 90 |
+
for error_type, count in error_counts.items():
|
| 91 |
+
print(f"{error_type}: {count}\n")
|
| 92 |
+
|
| 93 |
+
|
| 94 |
+
def visualize_error_types(self, error_counts):
|
| 95 |
+
df = pd.DataFrame({'Error Type': list(error_counts.keys()), 'Count': list(error_counts.values())})
|
| 96 |
+
fig = px.bar(df, x='Count', y='Error Type', title='Error Occurrences in The Generated Code')
|
| 97 |
+
fig.update_layout(
|
| 98 |
+
title={
|
| 99 |
+
'text': "Error Occurrences in The Generated Code",
|
| 100 |
+
'x': 0.5,
|
| 101 |
+
'y': 0.96,
|
| 102 |
+
'xanchor': 'center',
|
| 103 |
+
'yanchor': 'top'
|
| 104 |
+
},
|
| 105 |
+
xaxis_title="Error Counts",
|
| 106 |
+
yaxis_title="Error Codes"
|
| 107 |
+
)
|
| 108 |
+
fig.show()
|
| 109 |
+
|
| 110 |
+
def show_variables_in_table(self, output_counts, output):
|
| 111 |
+
table = PrettyTable()
|
| 112 |
+
table.field_names = ["Error Code", "Message"]
|
| 113 |
+
table.add_row([output_counts, output])
|
| 114 |
+
#table.add_row(["Error", error])
|
| 115 |
+
print(table)
|
| 116 |
+
|
| 117 |
+
def display_variables(self, output, error):
|
| 118 |
+
output_df = pd.DataFrame({"Output": [output]})
|
| 119 |
+
error_df = pd.DataFrame({"Error": [error]})
|
| 120 |
+
display(pd.concat([output_df, error_df], axis=1))
|
| 121 |
+
|
| 122 |
+
|
| 123 |
+
|
| 124 |
+
|
| 125 |
+
|
| 126 |
+
|
| 127 |
+
import autopep8
|
| 128 |
+
import black
|
| 129 |
+
import isort
|
| 130 |
+
import pylint.lint
|
| 131 |
+
import autoimport
|
| 132 |
+
from yapf.yapflib.yapf_api import FormatCode # reformat a string of code
|
| 133 |
+
|
| 134 |
+
class PythonCodeFormatter:
|
| 135 |
+
def __init__(self, code):
|
| 136 |
+
self.code = code.replace('▁', ' ').strip()
|
| 137 |
+
|
| 138 |
+
|
| 139 |
+
def load_code_from_file(self, filename):
|
| 140 |
+
# Load the code to be fixed
|
| 141 |
+
with open(filename, 'r') as f:
|
| 142 |
+
self.code = f.read()
|
| 143 |
+
|
| 144 |
+
def format(self):
|
| 145 |
+
try:
|
| 146 |
+
# Use isort to sort and organize the imports
|
| 147 |
+
formatted_code = isort.code(self.code)
|
| 148 |
+
|
| 149 |
+
# Use black to format the code
|
| 150 |
+
formatted_code = black.format_str(formatted_code, mode=black.Mode())
|
| 151 |
+
|
| 152 |
+
# Use autoimport to add a missing import statement
|
| 153 |
+
formatted_code = autoimport.fix_code(formatted_code)
|
| 154 |
+
|
| 155 |
+
# Use autopep8 to fix any remaining issues
|
| 156 |
+
formatted_code = autopep8.fix_code(formatted_code)
|
| 157 |
+
|
| 158 |
+
formatted_code, changed = FormatCode(formatted_code)
|
| 159 |
+
|
| 160 |
+
return formatted_code
|
| 161 |
+
|
| 162 |
+
except RuntimeError as error:
|
| 163 |
+
if str(error) == 'Project root not found.':
|
| 164 |
+
return formatted_code
|
| 165 |
+
else:
|
| 166 |
+
raise # re-raise the error if it's not the one we're looking for
|
| 167 |
+
|
| 168 |
+
except ValueError as error:
|
| 169 |
+
return formatted_code
|
| 170 |
+
|
| 171 |
+
return formatted_code
|
| 172 |
+
|
| 173 |
+
|
| 174 |
+
def save(self, filename):
|
| 175 |
+
# Save the fixed code to a file
|
| 176 |
+
with open(filename, 'w') as f:
|
| 177 |
+
f.write(self.code)
|
| 178 |
+
|
| 179 |
+
|
| 180 |
+
|
| 181 |
+
|
| 182 |
+
|
| 183 |
+
code_generator = CodeGenerator()
|
| 184 |
+
|
| 185 |
+
|
| 186 |
+
# Streamlit app
|
| 187 |
+
def main():
|
| 188 |
+
st.title('Code Generator and Error Checker')
|
| 189 |
+
nl_input = st.text_area('Enter natural language input for code generation')
|
| 190 |
+
if st.button('Generate Code'):
|
| 191 |
+
# Generate code
|
| 192 |
+
output_code = code_generator.generate_code(nl_input)
|
| 193 |
+
st.subheader('Generated Code')
|
| 194 |
+
st.code(output_code, language='python')
|
| 195 |
+
|
| 196 |
+
# Check code for errors
|
| 197 |
+
st.subheader('Error Check')
|
| 198 |
+
error_message = code_generator.check_code(output_code)
|
| 199 |
+
st.write('Error Counts:')
|
| 200 |
+
st.write(error_message)
|
| 201 |
+
|
| 202 |
+
|
| 203 |
+
st.subheader('Error Correction')
|
| 204 |
+
formatter = PythonCodeFormatter(output_code)
|
| 205 |
+
formatted_code = formatter.format()
|
| 206 |
+
st.write('Code after correction:')
|
| 207 |
+
st.write(formatted_code)
|
| 208 |
+
#st.subheader('Code after correction:')
|
| 209 |
+
#st.code(formatted_code, language='python')
|
| 210 |
+
|
| 211 |
+
|
| 212 |
+
|
| 213 |
+
if __name__ == '__main__':
|
| 214 |
+
|
| 215 |
+
main()
|