rss_reader.py

from argparse import ArgumentParser
from typing import Optional, Sequence
import requests
import xml.etree.ElementTree as ET
import json as json


class UnhandledException(Exception):
    pass


def rss_parser(
    xml: str,
    limit: int | None,
    json_mode: bool = False,
) -> list[str]:
    
    try:
        root = ET.fromstring(xml)
        channel = root.find("channel")
        
        if channel is None:
            raise ValueError("Invalid RSS feed format")

        channel_data = {
            "title":channel.findtext("title"),
            "link":channel.findtext("link"),
            "lastBuildDate":channel.findtext("lastBuildDate"),
            "pubDate":channel.findtext("pubDate"),
            "managinEditor":channel.findtext("managinEditor"),
            "description":channel.findtext("description"),
            }
        
        channel_data = {k: v for k, v in channel_data.items() if v}
            

        items_data: list[dict[str, str]] = []
        items = channel.findall("item")
        for item in items[:limit] if limit else items:
            item_data = {
                "title": item.findtext("title"),
                "author": item.findtext("author"),
                "pubDate": item.findtext("pubDate"),
                "link": item.findtext("link"),
                "category": item.findtext("category"),
                "description": item.findtext("description"),
            }

            item_data = {k: v for k, v in item_data.items() if v}
            
            items_data.append(item_data)

        channel_data |= {"items":items_data}

        if json_mode:
            return [json.dumps(channel_data, indent=4)]

        transformed_output = []

        if channel_data.get("title"):
            transformed_output.append(f"Feed: {channel_data['title']}")
        if channel_data.get("link"):
            transformed_output.append(f"Link: {channel_data['link']}")
        if channel_data.get("lastBuildDate"):
            transformed_output.append(f"Last Build Date: {channel_data['lastBuildDate']}")
        if channel_data.get("pubDate"):
            transformed_output.append(f"Publish Date: {channel_data['pubDate']}")
        if channel_data.get("language"):
            transformed_output.append(f"Language: {channel_data['language']}")
        if channel_data.get("managingEditor"):
            transformed_output.append(f"Editor: {channel_data['managingEditor']}")
        if channel_data.get("description"):
            transformed_output.append(f"Description: {channel_data['description']}")

        for item in channel_data["items"]:
            if item.get("title"):
                transformed_output.append(f"Title: {item['title']}")
            if item.get("author"):
                transformed_output.append(f"Author: {item['author']}")
            if item.get("pubDate"):
                transformed_output.append(f"Published: {item['pubDate']}")
            if item.get("link"):
                transformed_output.append(f"Link: {item['link']}")
            if item.get("category"):
                transformed_output.append(f"Categories: {item['category']}")
            if item.get("description"):
                transformed_output.append(item["description"])

        return transformed_output

    except ET.ParseError as e:
        raise UnhandledException(f"Failed to parse XML: {e}")


def main(argv: Optional[Sequence] = None):
    parser = ArgumentParser(
        prog="rss_reader",
        description="Pure Python command-line RSS reader.",
    )
    parser.add_argument("source", help="RSS URL", type=str, nargs="?")
    parser.add_argument(
        "--json", help="Print result as JSON in stdout", action="store_true"
    )
    parser.add_argument(
        "--limit", help="Limit news topics if this parameter provided", type=int
    )

    headers = {
        "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/111.0.0.0 Safari/537.36",
        }
    args = parser.parse_args(argv)
    xml = requests.get(args.source,headers=headers).text
    try:
        print("\n".join(rss_parser(xml, args.limit, args.json)))
        return 0
    except Exception as e:
        raise UnhandledException(e)


if __name__ == "__main__":
    main()