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: str) -> str:
    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 Exception:
            fixed_text += word + " "
    return fixed_text.encode('utf-8').decode()

def paper_data(paper_data_json: str, wait_time: int = 5) -> str:
    result_data = {"status": "success", "data": {}}
    papers_by_category = json.loads(paper_data_json)
    for category, papers in papers_by_category.items():
        print(f"Processing category: {category}")
        result_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) if title else ""
            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, mindmap = None, 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:
                fixed_title = html.escape(fix_text(title).strip())
                fixed_citation = html.escape(fix_text(citation).strip())
                fixed_summary = html.escape(str(summary).strip())
                fixed_mindmap = html.escape(str(mindmap).strip())
                post_status = post_blog(doi, fixed_title, category, fixed_summary, fixed_mindmap, fixed_citation, ACCESS_KEY, wait_time)
            except Exception as e:
                print(f"Error posting blog '{title}': {e}")
                continue
            result_data["data"][category][paper_id] = {
                "id": paper_id,
                "doi": doi,
                "title": fixed_title,
                "category": category,
                "posted": post_status,
                "citation": fixed_citation,
                "summary": fixed_summary,
                "mindmap": fixed_mindmap,
            }
    return json.dumps(result_data, indent=4, ensure_ascii=False)

def post_blogpost(uaccess_key: str, wait_time: int = 5) -> str:
    if uaccess_key != ACCESS_KEY:
        return False
    data = fetch_paper_data_with_category(uaccess_key)
    processed_data = paper_data(data, wait_time)
    try:
        send_email(processed_data)
        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 processed_data

def test(uaccess_key: str) -> str:
    if uaccess_key != ACCESS_KEY:
        return False
    test_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",
            }
        }
    }
    return paper_data(json.dumps(test_data, ensure_ascii=False, indent=4))

if __name__ == '__main__':
    result = test(ACCESS_KEY)
    print(result)