File size: 3,347 Bytes
1f0dd44 9b9dee3 4289365 9b9dee3 1f0dd44 9b9dee3 1f0dd44 9b9dee3 7edccf8 1f0dd44 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 |
import gradio as gr
from transformers import pipeline
import spacy
from docx import Document
import tempfile
import os
import torch
try:
nlp = spacy.load('en_core_web_sm')
except IOError:
import subprocess
subprocess.run(["python", "-m", "spacy", "download", "en_core_web_sm"])
nlp = spacy.load('en_core_web_sm')
#nlp = spacy.load('en_core_web_sm')
# Load models quietly
#nlp = spacy.load('en_core_web_sm')
device = 0 if torch.cuda.is_available() else -1
fill_mask = pipeline("fill-mask", model="bert-base-uncased", device=device) # Use GPU if available
# Cell 3: Define functions
def extract_text_from_docx(file_path):
try:
doc = Document(file_path)
return '\n'.join([para.text for para in doc.paragraphs])
except:
return ""
def save_updated_resume(text):
try:
temp_file = tempfile.NamedTemporaryFile(delete=False, suffix='.docx')
doc = Document()
for line in text.split('\n'):
doc.add_paragraph(line)
doc.save(temp_file.name)
return temp_file.name
except:
return None
def bert_suggest_replacements(text):
suggestions = []
improved_text = text
doc = nlp(text)
for token in doc:
if token.is_stop or not token.is_alpha or len(token.text) <= 3:
continue
masked_sentence = improved_text.replace(token.text, "[MASK]", 1)
try:
predictions = fill_mask(masked_sentence)
best_suggestion = predictions[0]['token_str']
if best_suggestion != token.text.lower():
improved_text = improved_text.replace(token.text, best_suggestion, 1)
suggestions.append(f"• Replace '{token.text}' with '{best_suggestion}'")
except:
continue
return improved_text, suggestions
def process_resume(file):
if file is None:
return "Please upload a resume file.", "No suggestions yet.", None
try:
resume_text = extract_text_from_docx(file.name)
if not resume_text:
return "Could not read the resume. Please ensure it's a valid .docx file.", "No suggestions available.", None
updated_text, suggestions = bert_suggest_replacements(resume_text)
output_file_path = save_updated_resume(updated_text)
suggestions_text = "Suggested Improvements:\n" + '\n'.join(suggestions) if suggestions else "No improvements suggested. Your resume looks good!"
return updated_text, suggestions_text, output_file_path
except:
return "An error occurred. Please try again with a different file.", "No suggestions available.", None
# Cell 4: Create and launch Gradio interface
iface = gr.Interface(
fn=process_resume,
inputs=gr.File(label="Upload Your Resume (.docx only)", file_types=[".docx"]),
outputs=[
gr.Textbox(label="Optimized Resume Text", lines=10),
gr.Textbox(label="ATS Improvement Suggestions", lines=10),
gr.File(label="Download Optimized Resume")
],
title="ATS Resume Optimizer",
description="Upload your .docx resume for ATS optimization. We'll suggest improvements and provide an updated version.",
theme="default",
css="""
.gradio-container {max-width: 800px; margin: auto;}
.output-markdown {white-space: pre-wrap;}
"""
)
iface.launch(share=True)
|