app.py

from gliner import GLiNER
import re
import fitz 
import gradio as gr

model = GLiNER.from_pretrained("gliner-community/gliner_large-v2.5", load_tokenizer=True)




def clean_text(text):
    # Remove all escape characters
    cleaned_text = re.sub(r'[\n\r\t\f\v]', ' ', text)

    # Remove any other non-printable characters
    cleaned_text = re.sub(r'[^\x20-\x7E]', '', cleaned_text)

    # Replace multiple spaces with a single space
    cleaned_text = re.sub(r'\s+', ' ', cleaned_text)

    # Strip leading and trailing whitespace
    cleaned_text = cleaned_text.strip()

    return cleaned_text


def pdf2text(file_path):

  with fitz.open(file_path) as doc:
      text = ""
      for page in doc:
          text += page.get_text()

  return clean_text(text)


def ner(text, labels, threshold) :
    labels = labels.split(",")
    labels = [label.strip() for label in labels]
    print(labels)


    return {
        "text": text,
        "entities": [
            {
                "entity": entity["label"],
                "word": entity["text"],
                "start": entity["start"],
                "end": entity["end"],
                "score": 0,
            }
            for entity in model.predict_entities(
                text, labels, flat_ner=True, threshold=threshold
            )
        ],
    }

def parser(file_path, labels, threshold):
  text = pdf2text(file_path)
  return ner(text, labels, threshold)


# Define a custom CSS style
custom_css = """
body {
    background-color: #f0f8ff;
    font-family: 'Arial', sans-serif;
}
.container {
    margin: auto;
    padding: 20px;
    border-radius: 10px;
    box-shadow: 0 4px 6px rgba(0, 0, 0, 0.1);
}
h1 {
    color: #3d1ad9;
    text-align: center;
}
#file_upload {
    display: flex;
    justify-content: center;
    margin-bottom: 20px;
}
"""


with gr.Blocks(css=custom_css) as demo:
    gr.HTML("<h1>AI-Powered Resume Parser</h1>")
    gr.HTML("<p style='text-align: center;'>This application extracts important data from your resume using innovative NLP methods. This tool's key advantage is that, in contrast to conventional resume parsers, it is generalized(Thanks to GLiNER team), meaning it functions in accordance with your needs. Simply enter the labels (NER) that you wish to extract, then adjust the threshold and submit the resume. Magic will happen in a few seconds.</p>")

    
    with gr.Row() as row:
        labels = gr.Textbox(
            label="Labels",
            placeholder="Enter your labels here (comma separated)",
            scale=2,
        )
        threshold = gr.Slider(
            0,
            1,
            value=0.3,
            step=0.01,
            label="Threshold",
            info="Lower the threshold to increase how many entities get predicted.",
            scale=0,
        )

    with gr.Row():
        file_input = gr.File(label="Upload Resume",
                file_types=['.pdf'],
                 elem_id="file_upload"
                )
        
    with gr.Row():
        parse_button = gr.Button("Parse Resume")

    with gr.Row():
        output = gr.HighlightedText(label="Parsed Resume",
                           combine_adjacent=True
                           )

    parse_button.click(fn=parser, inputs=[file_input,labels, threshold], outputs=output)

    gr.HTML("<p style='text-align: center;'>Our resume parser can identify and extract important details such as personal information, education, work experience, skills, and more. Simply upload your resume and let our AI do the work!</p>")
    
# Launch the interface
demo.queue()
demo.launch(share=True, debug=True)