app.py

import gradio as gr
from sentence_transformers import SentenceTransformer
import pandas as pd
import pickle

def make_clickable_both(val): 
    name, url = val.split('#')
    print(name+"\n")
    print(url+"\n")
    return f'<a href="{url}">{name}</a>'

def find(query):
    def get_detailed_instruct(task_description: str, query: str) -> str:
        return f'Instruct: {task_description}\nQuery: {query}'
    
    # Each query must come with a one-sentence instruction that describes the task
    task = 'Given a web search query, retrieve relevant passages that answer the query'
    queries = [
        get_detailed_instruct(task, query)
    ]
    print("cekpoin0\n")
    
    quran = pd.read_csv('quran-simple-clean.txt', delimiter="|")
    
    file = open('quran-splitted.sav','rb')
    quran_splitted = pickle.load(file)
    
    model = SentenceTransformer('intfloat/multilingual-e5-large-instruct')
    
    documents = quran_splitted['text'].tolist()
    # document_embeddings = model.encode(documents, convert_to_tensor=True, normalize_embeddings=True)
    # filename = 'encoded_quran_text_split_multilingual-e5-large-instruct.sav'
    # pickle.dump(embeddings, open(filename, 'wb'))
    file = open('encoded_quran_text_split_multilingual-e5-large-instructs.sav','rb')
    document_embeddings = pickle.load(file)
    print("cekpoin1\n")
    
    query_embeddings = model.encode(queries, convert_to_tensor=True, normalize_embeddings=True)
    scores = (query_embeddings @ document_embeddings.T) * 100
    print("cekpoin2\n")

    # insert the similarity value to dataframe & sort it
    file = open('quran-splitted.sav','rb')
    quran_splitted = pickle.load(file)
    quran_splitted['similarity'] = scores.tolist()[0]
    sorted_quran = quran_splitted.sort_values(by='similarity', ascending=False)
    print("cekpoin3\n")
    
    #results = ""
    results = pd.DataFrame()
    i = 0
    while i<20:
        result = sorted_quran.iloc[i]
        result_quran = quran.loc[(quran['sura']==result['sura']) & (quran['aya']==result['aya'])]
        results = pd.concat([results, result_quran])
        #results = results + result_quran['text'].item()+" (Q.S "+str(result['sura']).rstrip('.0')+":"+str(result['aya']).rstrip('.0')+")\n"
        i=i+1

    results['text'] = results['text'] + '#' + 'https://quran.com/'+results['sura'].astype(str)+':'+results['aya'].astype(str)+'/tafsirs/en-tafisr-ibn-kathir'
    results = results.style.format({'text': make_clickable_both})
    
    #return sorted_quran
    return results
    
demo = gr.Interface(fn=find, inputs="textbox", outputs=gr.Dataframe(headers=['sura', 'aya', 'text'],wrap=True,datatype=["str", "str", "html"]))
#demo = gr.Interface(fn=find, inputs="textbox", outputs="textbox")
    
if __name__ == "__main__":
    demo.launch(share=True)