pelican-plugins/tipue_search/tipue_search.py

# -*- coding: utf-8 -*-
"""
Tipue Search
============

A Pelican plugin to serialize generated HTML to JSON
that can be used by jQuery plugin - Tipue Search.

Copyright (c) Talha Mansoor
"""

from __future__ import unicode_literals

import os.path
import json
from bs4 import BeautifulSoup
from codecs import open

from pelican import signals


class Tipue_Search_JSON_Generator(object):

    def __init__(self, context, settings, path, theme, output_path, *null):

        self.output_path = output_path
        self.context = context
        self.siteurl = settings.get('SITEURL')
        self.json_nodes = []

    def create_json_node(self, page):

        if getattr(page, 'status', 'published') != 'published':
            return

        soup_title = BeautifulSoup(page.title.replace(' ', ' '))
        page_title = soup_title.get_text(' ', strip=True).replace('“', '"').replace('”', '"').replace('’', "'").replace('^', '^')

        soup_text = BeautifulSoup(page.content)
        page_text = soup_text.get_text(' ', strip=True).replace('“', '"').replace('”', '"').replace('’', "'").replace('¶', ' ').replace('^', '^')
        page_text = ' '.join(page_text.split())

        if getattr(page, 'category', 'None') == 'None':
            page_category = ''
        else:
            page_category = page.category.name

        page_url = self.siteurl + '/' + page.url

        node = {'title': page_title,
                'text': page_text,
                'tags': page_category,
                'loc': page_url}

        self.json_nodes.append(node)

    def generate_output(self, writer):
        path = os.path.join(self.output_path, 'tipuesearch_content.json')

        pages = self.context['pages'] + self.context['articles']

        for article in self.context['articles']:
            pages += article.translations

        for page in pages:
            self.create_json_node(page)
        root_node = {'pages': self.json_nodes}

        with open(path, 'w', encoding='utf-8') as fd:
            json.dump(root_node, fd, separators=(',', ':'))


def get_generators(generators):
    return Tipue_Search_JSON_Generator


def register():
    signals.get_generators.connect(get_generators)