Code

import openai
from scholarly import scholarly
from langchain.schema import HumanMessage
from langchain.chat_models import ChatOpenAI

#Function to search and retrieve papers from Google Scholar that match all keywords
def search_papers(keywords):
    search_query = scholarly.search_pubs(" ".join(keywords))
    papers = []
    for _ in range(100):  #Fetch up to 100 papers and then filter
        try:
            paper = next(search_query)
            # Check if all keywords are in the title or abstract
            title = paper['bib']['title'].lower()
            abstract = paper['bib'].get('abstract', '').lower()
            if all(keyword.lower() in title or keyword.lower() in abstract for keyword in keywords):
                papers.append(paper)
                if len(papers) >= 10:  #Limit to 10 papers
                    break
        except StopIteration:
            break
    return papers

#Function to summarize a list of papers using OpenAI's gpt-3.5-turbo model via LangChain
def summarize_papers(papers, temperature=0):
    summaries = []
    chat = ChatOpenAI(model_name="gpt-3.5-turbo", temperature=temperature)

    for paper in papers:
        title = paper['bib']['title']
        year = paper['bib'].get('pub_year', 'Unknown year')
        abstract = paper['bib'].get('abstract', 'No abstract available.')

        input_text = f"""Summarize the following research paper:

Title: {title}
Year: {year}

Abstract: {abstract}
"""
        prompt = HumanMessage(content=input_text)

        summary_response = chat([prompt])
        summary = summary_response.content.strip()

        summaries.append({
            'title': title,
            'year': year,
            'summary': summary
        })

    return summaries

#Prompt the user for keywords to search for papers
keywords_input = input("Enter the keywords to search for papers (separated by commas): ")
keywords = [keyword.strip() for keyword in keywords_input.split(",")]

#Search for papers using the provided keywords
papers = search_papers(keywords)

#Summarize the retrieved papers
summaries = summarize_papers(papers, temperature=0)  #Temperature set to 0 for deterministic results

#Print the summaries of the papers
for i, summary_info in enumerate(summaries, 1):
    print(f"Summary of paper {i}:\nTitle: {summary_info['title']}\nYear: {summary_info['year']}\nSummary: {summary_info['summary']}\n")