collection_hatena_entory_to_static_page.rb

# -*- coding: utf-8 -*-

require 'rubygems'
require 'oauth'
require 'rexml/document'
require "sqlite3"
require 'net/http'
require 'net/https'
require 'uri'
require 'time'

HATENA_ID = "nasust";
BLOG_ID = "showrtpath.hatenablog.com";
DB_PATH = "blog_entry.db";
API_KEY = "xxxxxxxxx";

PAGE_CAGEGORY_URL = {
	"ライフ" => "http://showrtpath.hatenablog.com/entry/life",
	"Showrtpath" => "http://showrtpath.hatenablog.com/entry/showrtpath",	
	"Apple/ガジェット" => "http://showrtpath.hatenablog.com/entry/apple_gadget",	
	"IOS開発" => "http://showrtpath.hatenablog.com/entry/developp_ios",	
	"WEB" => "http://showrtpath.hatenablog.com/entry/wev_dev",	
	"UI/UX" => "http://showrtpath.hatenablog.com/entry/ui_ux",	
	"ブックマーク" => "http://showrtpath.hatenablog.com/entry/bookmark",	
}

PAGE_CAGEGORY_ENTRY_ID = {
	"ライフ" => "12921228815716456765",
	"Showrtpath" => "12921228815716456900",
	"Apple/ガジェット" => "12921228815716456972",
	"IOS開発" => "12921228815716457009",
	"WEB" => "12921228815716466347",
	"UI/UX" => "12921228815716457088",
	"ブックマーク" => "12921228815716457122",
}

@page_category = {
	"ライフ" => ["ライフ","ガジェット"],
	"Showrtpath" => ["Showrtpathブラウザの紹介"],
	"Apple/ガジェット" => ["iphone","Mac "],
	"IOS開発" => ["iOS開発","iOS開発の記事のまとめ" , "ソースコード"],
	"WEB" => ["WEB"],
	"UI/UX" => ["UI/UX"],
	"ブックマーク" => ["nasustのbookmark"]
}

@page_category_data = {
	"ライフ" => [],
	"Showrtpath" => [],	
	"Apple/ガジェット" => [],	
	"IOS開発" => [],	
	"WEB" => [],	
	"UI/UX" => [],	
	"ブックマーク" => [],	
}


@db = SQLite3::Database.new( DB_PATH );
result= @db.execute "SELECT COUNT(*) FROM sqlite_master WHERE type='table' AND name='blog_entry';"
if result[0][0] == 0 ;
	sql = <<-SQL
create table blog_entry(
id INTEGER PRIMARY KEY,
title TEXT,
href TEXT UNIQUE,
image_url TEXT,
published TEXT,
tag TEXT
);
SQL
	@db.execute sql
end


def get_blog( next_page_url )

	http = Net::HTTP.new("blog.hatena.ne.jp",443);
	http.use_ssl = true;
	http.verify_mode = OpenSSL::SSL::VERIFY_NONE
	path =  "/#{HATENA_ID}/#{BLOG_ID}/atom/entry";
	if( next_page_url )
		path = path + "?" + next_page_url;
	end
	
	request = Net::HTTP::Get.new(path);
	request.basic_auth(HATENA_ID, API_KEY);
	response = http.request(request);
	
	if( response.code == "200" )
		#puts response.body;
		
		doc = REXML::Document.new(response.body);
		
		next_url = nil;
		doc.elements.each("*/link[@rel='next']"){ | element |
			next_url = element.attributes["href"];
		}
		
		doc.elements.each("*/entry"){ | element |
			blog_title = nil;
			blog_link = nil;
			published = nil;
			image_path = nil;
			category = [];
			
			#block link
			element.elements.each("link[@rel='alternate']"){ | element |
				blog_link = element.attributes["href"];
			}
			
			skip = false;
			PAGE_CAGEGORY_URL.each{ | key , url |
				if( url == blog_link )
					skip = true;
				end
			}
			if( skip )
				next;
			end
					
			element.elements.each("title"){ | element |
				blog_title = element.text;
			}
			element.elements.each("published"){ | element |
				published = element.text;
			}
			element.elements.each("category"){ | element |
				category << element.attributes["term"];
				#p element.attributes["term"];
			}

			element.elements.each("hatena:formatted-content"){ | element |
				begin				
					html = element.text;
					URI.extract( html , ["http","https"] ).each{ | url |
						if( url =~ /\.(jpg|gif|png)$/ )
							image_path = url;
							break;
						end
					}
					#p image_path;
				rescue
					p $!
				end
			}
			if( image_path == nil ||image_path.empty? )
				image_path = "http://cdn-ak.f.st-hatena.com/images/fotolife/n/nasust/20140104/20140104094040.jpg";
			end
			begin
				@db.execute "insert into blog_entry(title,href,image_url,published,tag) values ( ?, ? , ? , ? , ?)" , [ blog_title , blog_link , image_path , published , category.join(",") ]
			rescue
				next_url = nil;
				break;
			end
						
		}
		if(next_url)
			#puts next_url;
			uri = URI.parse(next_url);
			get_blog(uri.query);
		else
			puts "generate"
			generate_static_page();
		end
	end
end

def generate_static_page()
	@db.execute( "select * from blog_entry" ) do | row |
		title = row[1];
		href = row[2];
		image_url = row[3];
		published = row[4];
		tags = row[5];
		
		category_key = [];
		tags_split = tags.split(",");
		tags_split.each{ | tag |
			@page_category.each{ | key , match_tags |
				match_tags.each{ | match_tag |
					if( tag == match_tag )
						# puts key + ":" + title + " :" + tag + " : " + match_tag;
						category_key << key;
						break;
					end
				}
			}
		}
		category_key.uniq!;
		
		if( category_key.empty? )
			category_key << "ライフ"
		end
		
		category_key.each{ | key |
			@page_category_data[ key ] << {
				:title => title,
				:href => href,
				:image_url => image_url,
				:published => published,
				:tags => tags,
			};
		}
		
		#puts "blog_title: #{title} + blog_link : #{href} published:#{published} image_path:#{image_url} + category:#{ tags }\n\n"
		
	end
	
	@page_category_data.each{ | key , entrys |
		post_url = PAGE_CAGEGORY_URL[key];
		
		body = [];
		
		entrys = entrys.sort{| a , b |
			a_published = a[:published];
			b_published = b[:published];
			a_time = Time.parse(a_published);
			b_time = Time.parse(b_published);
			
			a_time <=> b_time;
		}
		
		entrys = entrys.reverse;
		
		entrys.each{ | entry |
			title = entry[:title];
			href = entry[:href];
			image_url = entry[:image_url];
			published = entry[:published];
			tags = entry[:tags];
			
			html=<<-HTML
<div class="entry_item">
	<div class="entry_image">
		<a href="#{href}"><img src="#{image_url}"></a>
	</div>
	<div class="title"><a href="#{href}">#{title}</a></div>
</div>
HTML
			body << html;
		}
		
		
		xml = <<-XML
<?xml version="1.0" encoding="utf-8"?>
<entry xmlns="http://www.w3.org/2005/Atom"
       xmlns:app="http://www.w3.org/2007/app">
  <title>#{key}</title>
  <author><name>#{HATENA_ID}</name></author>
  <updated>2013-07-14T00:00:00</updated>
  <content type="text/plain">
<![CDATA[
<br>
<div id="entry_container" class="category_columns js-masonry" data-masonry-options='{ "columnWidth": 175, "itemSelector": ".entry_item" }'>
#{body.join("\n")}
</div>
<script type="text/javascript" src="http://showrtpath.com/hatena/masonry.pkgd.min.js"></script>
]]>
  </content>
  <app:control>
    <app:draft>no</app:draft>
  </app:control>
</entry>
XML
		#puts xml;
		entry_id = PAGE_CAGEGORY_ENTRY_ID[key];
				
		http = Net::HTTP.new("blog.hatena.ne.jp",443);
		http.use_ssl = true;
		http.verify_mode = OpenSSL::SSL::VERIFY_NONE
	
		req = Net::HTTP::Put.new( "/#{HATENA_ID}/#{BLOG_ID}/atom/entry/#{entry_id}");
		req.basic_auth(HATENA_ID, API_KEY);
		req.body = xml;
		response = http.request(req);
		p response;
	}
	
end


get_blog(nil);