app.py

import streamlit as st
import pandas as pd
from transformers import pipeline
from stqdm import stqdm
from simplet5 import SimpleT5
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM


@st.cache
def load_t5():
    model = AutoModelForSeq2SeqLM.from_pretrained("t5-base")

    tokenizer = AutoTokenizer.from_pretrained("t5-base")
    return model, tokenizer


@st.cache(allow_output_mutation=False)
def custom_model():
    return pipeline("summarization", model="my_awesome_sum/")


@st.cache
def convert_df(df):
    # IMPORTANT: Cache the conversion to prevent computation on every rerun
    return df.to_csv().encode("utf-8")


@st.cache
def load_one_line_summarizer(model):
    return model.load_model("t5", "snrspeaks/t5-one-line-summary")


st.set_page_config(layout="wide", page_title="Amazon Review Summarizer")
st.title("Amazon Review Summarizer")

uploaded_file = st.file_uploader("Choose a file", type=["xlsx", "xls", "csv"])
summarizer_option = st.selectbox(
    "Select Summarizer",
    ("Custom trained on the dataset", "t5-base", "t5-one-line-summary"),
)
ps = st.empty()
if st.button("Process"):
    if uploaded_file is not None:

        df = pd.read_excel(uploaded_file)
        columns = df.columns.values.tolist()
        columns = [x.lower() for x in columns]
        df.columns = columns
        print(summarizer_option)
        if summarizer_option == "Custom trained on the dataset":
            model = custom_model()
            print(summarizer_option)
            text = df["text"].values.tolist()
            progress_text = "Summarization in progress. Please wait."
            summary = []

            for x in stqdm(range(len(text))):
                try:
                    summary.append(
                        model(
                            f"summarize: {text[x]}", max_length=50, early_stopping=True
                        )[0]["summary_text"]
                    )
                except:
                    pass
            output = pd.DataFrame(
                {"text": df["text"].values.tolist(), "summary": summary}
            )
            csv = convert_df(output)
            st.download_button(
                label="Download data as CSV",
                data=csv,
                file_name=f"{summarizer_option}_df.csv",
                mime="text/csv",
            )
        if summarizer_option == "t5-base":
            model, tokenizer = load_t5()
            text = df["text"].values.tolist()
            summary = []
            for x in stqdm(range(10)):

                tokens_input = tokenizer.encode(
                    "summarize: " + text[x],
                    return_tensors="pt",
                    max_length=tokenizer.model_max_length,
                    truncation=True,
                )
                summary_ids = model.generate(
                    tokens_input,
                    min_length=80,
                    max_length=150,
                    length_penalty=20,
                    num_beams=2,
                )
                summary_gen = tokenizer.decode(summary_ids[0], skip_special_tokens=True)
                summary.append(summary_gen)

            output = pd.DataFrame(
                {"text": df["text"].values.tolist()[0:10], "summary": summary}
            )
            csv = convert_df(output)
            st.download_button(
                label="Download data as CSV",
                data=csv,
                file_name=f"{summarizer_option}_df.csv",
                mime="text/csv",
            )

        if summarizer_option == "t5-one-line-summary":
            model = SimpleT5()
            text = df["text"].values.tolist()

            load_one_line_summarizer(model=model)

            summary = []
            for x in stqdm(range(10)):
                try:
                    summary.append(model.predict(text[x])[0])
                except:
                    pass
            output = pd.DataFrame(
                {"text": df["text"].values.tolist()[0:10], "summary": summary}
            )
            csv = convert_df(output)
            st.download_button(
                label="Download data as CSV",
                data=csv,
                file_name=f"{summarizer_option}_df.csv",
                mime="text/csv",
            )