|
|
import gradio as gr |
|
|
from keybert import KeyBERT |
|
|
|
|
|
|
|
|
kw_model = KeyBERT(model="all-MiniLM-L6-v2") |
|
|
|
|
|
|
|
|
EXAMPLES = { |
|
|
"Scientific Abstract": """ |
|
|
Compatibility of systems of linear constraints over the set of natural numbers. |
|
|
Criteria of compatibility of a system of linear Diophantine equations, strict inequations, |
|
|
and nonstrict inequations are considered. Upper bounds for components of a minimal set of solutions |
|
|
and algorithms of construction of minimal generating sets of solutions for all types of systems are given. |
|
|
""", |
|
|
"News Article": """ |
|
|
Machine learning is revolutionizing the way we interact with technology. |
|
|
Artificial intelligence systems are becoming more sophisticated, enabling automated decision making |
|
|
and pattern recognition at unprecedented scales. Deep learning algorithms continue to improve, |
|
|
making breakthroughs in natural language processing and computer vision. |
|
|
""", |
|
|
"Technical Documentation": """ |
|
|
The user interface provides intuitive navigation through contextual menus and adaptive layouts. |
|
|
System responses are optimized for performance while maintaining high reliability standards. |
|
|
Database connections are pooled to minimize resource overhead and maximize throughput. |
|
|
""" |
|
|
} |
|
|
|
|
|
def extract_keywords(text, num_words, ngram_range, diversity, use_mmr): |
|
|
|
|
|
min_n, max_n = map(int, ngram_range.split('-')) |
|
|
|
|
|
|
|
|
keywords = kw_model.extract_keywords( |
|
|
text, |
|
|
keyphrase_ngram_range=(min_n, max_n), |
|
|
stop_words='english', |
|
|
top_n=num_words, |
|
|
use_mmr=use_mmr, |
|
|
diversity=diversity if use_mmr else None |
|
|
) |
|
|
|
|
|
|
|
|
result = [] |
|
|
for keyword, score in keywords: |
|
|
result.append(f"• {keyword:<30} (score: {score:.4f})") |
|
|
|
|
|
return "\n".join(result) if result else "No keywords found." |
|
|
|
|
|
def load_example(example_name): |
|
|
return EXAMPLES.get(example_name, "") |
|
|
|
|
|
|
|
|
with gr.Blocks(title="KeyBERT Keyword Extraction") as demo: |
|
|
gr.Markdown("# 🔑 Keyword extraction using KeyBERT") |
|
|
gr.Markdown("**Developed by : Venugopal Adep**") |
|
|
gr.Markdown("Extract keywords using BERT embeddings and semantic similarity") |
|
|
|
|
|
with gr.Row(): |
|
|
with gr.Column(scale=2): |
|
|
input_text = gr.Textbox( |
|
|
label="Input Text", |
|
|
placeholder="Enter your text here...", |
|
|
lines=8 |
|
|
) |
|
|
example_dropdown = gr.Dropdown( |
|
|
choices=list(EXAMPLES.keys()), |
|
|
label="Load Example Text" |
|
|
) |
|
|
|
|
|
with gr.Column(scale=1): |
|
|
ngram_range = gr.Dropdown( |
|
|
choices=["1-1", "1-2", "1-3", "2-2", "2-3", "3-3"], |
|
|
value="1-2", |
|
|
label="Keyword Length (N-gram Range)" |
|
|
) |
|
|
|
|
|
num_words = gr.Slider( |
|
|
minimum=1, |
|
|
maximum=20, |
|
|
value=10, |
|
|
step=1, |
|
|
label="Number of Keywords" |
|
|
) |
|
|
|
|
|
use_mmr = gr.Checkbox( |
|
|
label="Use Maximal Marginal Relevance", |
|
|
value=True |
|
|
) |
|
|
|
|
|
diversity = gr.Slider( |
|
|
minimum=0.1, |
|
|
maximum=1.0, |
|
|
value=0.5, |
|
|
step=0.1, |
|
|
label="Diversity (MMR)", |
|
|
interactive=True |
|
|
) |
|
|
|
|
|
extract_btn = gr.Button("Extract Keywords", variant="primary") |
|
|
|
|
|
output_text = gr.Textbox( |
|
|
label="Extracted Keywords", |
|
|
lines=12, |
|
|
interactive=False |
|
|
) |
|
|
|
|
|
|
|
|
example_dropdown.change( |
|
|
load_example, |
|
|
inputs=[example_dropdown], |
|
|
outputs=[input_text] |
|
|
) |
|
|
|
|
|
extract_btn.click( |
|
|
extract_keywords, |
|
|
inputs=[ |
|
|
input_text, |
|
|
num_words, |
|
|
ngram_range, |
|
|
diversity, |
|
|
use_mmr |
|
|
], |
|
|
outputs=[output_text] |
|
|
) |
|
|
|
|
|
demo.launch() |
