Spaces:
Sleeping
Sleeping
Create app.py
Browse files
app.py
ADDED
|
@@ -0,0 +1,66 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import re
|
| 2 |
+
import PyPDF2
|
| 3 |
+
import gradio as gr
|
| 4 |
+
from transformers import pipeline
|
| 5 |
+
|
| 6 |
+
ner_pipeline = pipeline("ner", model="dslim/bert-base-NER", tokenizer="dslim/bert-base-NER", aggregation_strategy="simple")
|
| 7 |
+
|
| 8 |
+
def clean_resume_text(text):
|
| 9 |
+
text = re.sub(r'http\S+', ' ', text)
|
| 10 |
+
text = re.sub(r'#\S+', '', text)
|
| 11 |
+
text = re.sub(r'@\S+', ' ', text)
|
| 12 |
+
text = re.sub(r'[^\w\s]', ' ', text)
|
| 13 |
+
text = re.sub(r'[^\x00-\x7f]', ' ', text)
|
| 14 |
+
return re.sub(r'\s+', ' ', text).strip()
|
| 15 |
+
|
| 16 |
+
def extract_resume_text(file):
|
| 17 |
+
try:
|
| 18 |
+
reader = PyPDF2.PdfReader(file)
|
| 19 |
+
text = ""
|
| 20 |
+
for page in reader.pages:
|
| 21 |
+
extracted = page.extract_text()
|
| 22 |
+
if extracted:
|
| 23 |
+
text += extracted + " "
|
| 24 |
+
return text if text.strip() else "Error: No text extracted."
|
| 25 |
+
except Exception as e:
|
| 26 |
+
return f"Error reading PDF: {str(e)}"
|
| 27 |
+
|
| 28 |
+
def extract_entities_from_pdf(file):
|
| 29 |
+
resume_text = extract_resume_text(file)
|
| 30 |
+
if resume_text.startswith("Error"):
|
| 31 |
+
return resume_text
|
| 32 |
+
|
| 33 |
+
entities = ner_pipeline(resume_text)
|
| 34 |
+
|
| 35 |
+
result = {
|
| 36 |
+
"Persons": [],
|
| 37 |
+
"Organizations": [],
|
| 38 |
+
"Locations": [],
|
| 39 |
+
"Other": []
|
| 40 |
+
}
|
| 41 |
+
|
| 42 |
+
for entity in entities:
|
| 43 |
+
label = entity.get("entity_group")
|
| 44 |
+
word = entity.get("word")
|
| 45 |
+
if label == "PER":
|
| 46 |
+
result["Persons"].append(word)
|
| 47 |
+
elif label == "ORG":
|
| 48 |
+
result["Organizations"].append(word)
|
| 49 |
+
elif label == "LOC":
|
| 50 |
+
result["Locations"].append(word)
|
| 51 |
+
else:
|
| 52 |
+
result["Other"].append(word)
|
| 53 |
+
|
| 54 |
+
result["Cleaned_Text"] = clean_resume_text(resume_text)
|
| 55 |
+
return result
|
| 56 |
+
|
| 57 |
+
iface = gr.Interface(
|
| 58 |
+
fn=extract_entities_from_pdf,
|
| 59 |
+
inputs=gr.File(file_types=[".pdf"]),
|
| 60 |
+
outputs="json",
|
| 61 |
+
title="Resume Entity Extractor",
|
| 62 |
+
description="Upload a PDF resume. It will extract names, organizations, and locations using Hugging Face NER."
|
| 63 |
+
)
|
| 64 |
+
|
| 65 |
+
if __name__ == "__main__":
|
| 66 |
+
iface.launch()
|