#!/bin/python
"""
File:    make_atom.py
Version: Python 3.6.5
Author:  Nicholas Russo (http://njrusmc.net)
Purpose: CD process to build the atom.xml artifact which allows users to
         subscribe to the blog. The script finds all HTML blog files and
	 extracts the relevant information needed by Atom.
"""

import os
from datetime import datetime
from bs4 import BeautifulSoup
from jinja2 import Environment, FileSystemLoader

def get_blog_files(top):
    """
    Walk the filesystem for a given directory, grabbing only the HTML
    files. Return a list of strings in 'path/blog' format for future iteration,
    which will open up each file to perform parsing.
    """
    blog_files = []
    for path, dirs, files in os.walk(top=top):
        if not dirs:
            blog_files.extend([f'{path}/{blog}' for blog in files if 'html' in blog])
    return blog_files

def main():
    """
    feed_updated: str
    blogs:
      - title: str
        updated: str
        url: str
	summary: str
        categories:
          - str
          - str
    """

    # Create main data dictionary with an initial timestamp of a date
    # in the past, so any new updates will certainly overwrite it.
    # Instantiate an empty list of blogs (strings) for future extension.
    data = {
        'feed_updated': '2019-01-01',
        'blogs': []
    }

    # Iterate over the number of blogs found
    # FIX this needs to be more generic, not hard coded to 2019
    for blog_file in sorted(get_blog_files('blog/2019'), reverse=True):

        # Open the file, read the text, and parse HTML data using BS4.
        print(f'{blog_file} ... ', end='')
        with open(blog_file) as stream:
            html_text = stream.read()
        soup = BeautifulSoup(html_text, 'html.parser')

        # If the "active" tag is absent, the blog should not be added to the Atom feed.
        # Else the "active" tag is present, so keep processing.
        active_tag = soup.find('meta', attrs={'name': 'active'})
        if active_tag is None:
            print('inactive')
            continue

        # Split the date from the title, then convert the date to the proper Atom
        # format of YYYY-MM-DD from the blog format of DD Mmm YYYY.
        date_title_tag = soup.find('title')
        date_title_elems = date_title_tag.text.split(' - ')
        date_ymd = datetime.strptime(date_title_elems[0], '%d %b %Y').strftime('%Y-%m-%d')

        # If the current date is newer than the previous feed date, update
        # the overall feed date. This results in the newest blog governing
        # the feed update value.
        if date_ymd > data['feed_updated']:
            data['feed_updated'] = date_ymd

        # Find category string. If it exists, split on comma, otherwise
        # use an empty list to gracefully kill the jinja2 for loop
        category_tags = soup.find('meta', attrs={'name': 'category'})
        if category_tags:
            category_strs = category_tags.get('content').split(',')
        else:
            category_strs = []

        # The summary should be t
        body_elems = soup.find('body').findChildren(recursive=False)
        summary_str = body_elems[1]

        # Assemble temporary dict to add to overall blog list
        blog_dict = {
            'title': date_title_elems[1],
            'updated': date_ymd,
            'url': 'http://njrusmc.net/' + blog_file,
            'summary': summary_str,
            'categories': category_strs
        }
        data['blogs'].append(blog_dict)
        print('OK!')

    # Use jinja2 to build the proper Atom XML file for the blog
    j2_env = Environment(loader=FileSystemLoader('.'), trim_blocks=True, autoescape=True)
    template = j2_env.get_template('ci_scripts/atom_xml.j2')
    with open('blog/atom.xml', 'w') as atom_xml:
        atom_xml.write(template.render(data=data))

    print('XML data written to blog/atom.xml')

if __name__ == '__main__':
    main()
#
# EOF