import streamlit as st
import sparknlp
import os
import pandas as pd
from sparknlp.base import *
from sparknlp.annotator import *
from pyspark.ml import Pipeline
from sparknlp.pretrained import PretrainedPipeline
# Page configuration
st.set_page_config(
layout="wide",
page_title="Spark NLP Demos App",
initial_sidebar_state="auto"
)
# CSS for styling
st.markdown("""
""", unsafe_allow_html=True)
@st.cache_resource
def init_spark():
return sparknlp.start()
@st.cache_resource
def create_pipeline(model):
documentAssembler = DocumentAssembler()\
.setInputCol("text")\
.setOutputCol("document")
use = UniversalSentenceEncoder.pretrained("tfhub_use", "en")\
.setInputCols(["document"])\
.setOutputCol("sentence_embeddings")
sentimentdl = ClassifierDLModel.pretrained(model)\
.setInputCols(["sentence_embeddings"])\
.setOutputCol("sentiment")
nlpPipeline = Pipeline(stages = [documentAssembler, use, sentimentdl])
return nlpPipeline
def fit_data(pipeline, data):
empty_df = spark.createDataFrame([['']]).toDF('text')
pipeline_model = pipeline.fit(empty_df)
model = LightPipeline(pipeline_model)
results = model.fullAnnotate(data)[0]
return results['sentiment'][0].result
# Set up the page layout
st.markdown('
State-of-the-Art Emotion Detecter in Tweets with Spark NLP
', unsafe_allow_html=True)
# Sidebar content
model = st.sidebar.selectbox(
"Choose the pretrained model",
["classifierdl_use_emotion"],
help="For more info about the models visit: https://sparknlp.org/models"
)
# Reference notebook link in sidebar
link = """
"""
st.sidebar.markdown('Reference notebook:')
st.sidebar.markdown(link, unsafe_allow_html=True)
# Load examples
examples = [
"I am SO happy the news came out in time for my birthday this weekend! My inner 7-year-old cannot WAIT!",
"That moment when you see your friend in a commercial. Hahahaha!",
"My soul has just been pierced by the most evil look from @rickosborneorg. A mini panic attack & chill in bones followed soon after.",
"For some reason I woke up thinkin it was Friday then I got to school and realized its really Monday -_-",
"I'd probably explode into a jillion pieces from the inablility to contain all of my if I had a Whataburger patty melt right now. #drool",
"These are not emotions. They are simply irrational thoughts feeding off of an emotion",
"Found out im gonna be with sarah bo barah in ny for one day!!! Eggcitement :)",
"That awkward moment when you find a perfume box full of sensors!",
"Just home from group celebration - dinner at Trattoria Gianni, then Hershey Felder's performance - AMAZING!!",
"Nooooo! My dad turned off the internet so I can't listen to band music!"
]
st.subheader("Automatically identify Joy, Surprise, Fear, Sadness in Tweets using out pretrained Spark NLP DL classifier.")
selected_text = st.selectbox("Select a sample", examples)
custom_input = st.text_input("Try it for yourself!")
if custom_input:
selected_text = custom_input
elif selected_text:
selected_text = selected_text
st.subheader('Selected Text')
st.write(selected_text)
# Initialize Spark and create pipeline
spark = init_spark()
pipeline = create_pipeline(model)
output = fit_data(pipeline, selected_text)
# Display output sentence
if output == 'joy':
st.markdown("""This seems like a {} tweet. 😂
""".format('joyous'), unsafe_allow_html=True)
elif output == 'surprise':
st.markdown("""This seems like a {} tweet. 😊
""".format('surprised'), unsafe_allow_html=True)
elif output == 'sadness':
st.markdown("""This seems like a {} tweet. 😟
""".format('sad'), unsafe_allow_html=True)
elif output == 'fear':
st.markdown("""This seems like a {} tweet. 😱
""".format('fearful'), unsafe_allow_html=True)