main.py

import json
import os
import time
import dotenv
import html
from summarize_paper import summarize_paper
from fetch_data import fetch_paper_data_with_category
from post_blog import post_blog
from send_mail import send_email

dotenv.load_dotenv()
access_key = os.getenv("ACCESS_KEY")

def fix_text(text):
    text = html.escape(text.encode('utf-8').decode('utf-8').replace("â¦", "..., "))
    fixed_text = ""
    for word in text.split():
        try:
            fixed_text += word.encode('latin1').decode('utf-8')+" "
        except:
            fixed_text += word+" "
    return fixed_text.encode('utf-8').decode()

def paper_data(paper_data, wait_time=5):
    data = {"status": "success"}
    data['data'] = {}
    paper_data = json.loads(paper_data)
    for category, papers in paper_data.items():
            print(f"Processing category: {category}")
            data['data'][category] = {}
            for paper_id, details in papers.items():
                doi = details.get("doi")
                pdf_url = details.get("pdf_url")
                title = details.get("title")
                title = html.escape(title)
                citation = details.get("citation")
                if not all([paper_id, doi, pdf_url, title, citation]):
                    print(f"Skipping paper with ID: {paper_id} (missing details)")
                    continue
                summary = None
                mindmap = None
                max_retries = 3
                retry_count = 0
                while (not summary or not mindmap) and retry_count < max_retries:
                    try:
                        summary, mindmap = summarize_paper(pdf_url, paper_id, access_key)
                        if summary and mindmap:
                            break
                    except Exception as e:
                        print(f"Error summarizing paper {paper_id}: {e}")
                    retry_count += 1
                    if retry_count < max_retries:
                        print(f"Retrying paper {paper_id} in 3 minutes")
                        time.sleep(3*60)
                if not summary or not mindmap:
                    print(f"Failed to summarize paper {paper_id} after {max_retries} attempts")
                    continue
                try:
                    title = fix_text(title)
                    citation = fix_text(citation)
                    title = html.escape(str(title).strip())
                    citation = html.escape(str(citation).strip())
                    summary = html.escape(str(summary).strip())
                    mindmap = html.escape(str(mindmap).strip())
                    status = post_blog(doi, title, category, summary, mindmap, citation, access_key, wait_time)
                except Exception as e:
                    print(f"Error posting blog '{title}': {e}")
                    continue
                data['data'][category][paper_id] = {
                    "id": paper_id,
                    "doi": doi,
                    "title": title,
                    "category": category,
                    "posted": status,
                    "citation": citation,
                    "summary": summary,
                    "mindmap": mindmap,
                }
    data = json.dumps(data, indent=4, ensure_ascii=False)
    return data

def post_blogpost(uaccess_key, wait_time=5):
    if uaccess_key != access_key:
        return False
    data = fetch_paper_data_with_category(uaccess_key)
    pdata = paper_data(data, wait_time)
    try:
        send_email(pdata)
        print("\n-------------------------------------------------------\nMail Sent\n-------------------------------------------------------\n")
    except Exception as e:
        print(f"\n-------------------------------------------------------\nError sending mail: {e}\n-------------------------------------------------------\n")
    finally:
        print("\n-------------------------------------------------------\nProcess Completed\n-------------------------------------------------------\n")
    return pdata

def test(uaccess_key):
    if uaccess_key != access_key:
        return False
    data = {
        "Economics": {
            "2501.00578":{
                "paper_id":"2501.00578",
                "doi":"https://doi.org/10.1002/alz.14328",
                "title":"Bound-State Beta Decay of $\\mathbf{\\mathrm{^{205}{Tl}^{81+}}}$ Ions and the LOREX Project",
                "category":"Economics",
                "pdf_url":"https://arxiv.org/pdf/2501.00578",
                "citation":"Miller, A. D. (2025). The limits of tolerance (Version 1). arXiv. https://doi.org/10.48550/ARXIV.2501.00578",
            },
        },
    }
    status = paper_data(json.dumps(data, ensure_ascii=False, indent=4))
    return status
                
if __name__ == '__main__':
    data = test(access_key)
    print(data)