tinami/pixa downloads now supported

This commit is contained in:
albert
2011-09-30 14:27:04 -04:00
parent 2287bc8d61
commit aa2b65c48b
13 changed files with 326 additions and 170 deletions

View File

@@ -1,115 +0,0 @@
class Download
class Error < Exception ; end
attr_accessor :source, :content_type, :file_path
def initialize(source, file_path)
@source = source
@file_path = file_path
end
def download!
http_get_streaming do |response|
self.content_type = response["Content-Type"]
File.open(file_path, "wb") do |out|
response.read_body(out)
end
end
after_download
end
def pixiv_rewrite(headers)
return unless source =~ /pixiv\.net/
headers["Referer"] = "http://www.pixiv.net"
# Don't download the small version
if source =~ %r!(/img/.+?/.+?)_m.+$!
match = $1
source.sub!(match + "_m", match)
end
# Download the big version if it exists
if source =~ %r!(\d+_p\d+)\.!
match = $1
repl = match.sub(/_p/, "_big_p")
big_source = source.sub(match, repl)
if pixiv_http_exists?(big_source)
self.source = big_source
end
end
end
def pixiv_http_exists?(url)
# example: http://img01.pixiv.net/img/as-special/15649262_big_p2.jpg
exists = false
uri = URI.parse(url)
Net::HTTP.start(uri.host, uri.port) do |http|
headers = {"Referer" => "http://www.pixiv.net", "User-Agent" => "#{Danbooru.config.app_name}/#{Danbooru.config.version}"}
http.request_head(uri.request_uri, headers) do |res|
if res.is_a?(Net::HTTPSuccess)
exists = true
end
end
end
exists
end
def before_download(headers)
pixiv_rewrite(headers)
end
def after_download
fix_image_board_sources
end
def http_get_streaming(options = {})
max_size = options[:max_size] || Danbooru.config.max_file_size
max_size = nil if max_size == 0 # unlimited
limit = 4
while true
url = URI.parse(source)
unless url.is_a?(URI::HTTP)
raise Error.new("URL must be HTTP")
end
Net::HTTP.start(url.host, url.port) do |http|
http.read_timeout = 10
headers = {
"User-Agent" => "#{Danbooru.config.safe_app_name}/#{Danbooru.config.version}"
}
before_download(headers)
url = URI.parse(source)
http.request_get(url.request_uri, headers) do |res|
case res
when Net::HTTPSuccess then
if max_size
len = res["Content-Length"]
raise Error.new("File is too large (#{len} bytes)") if len && len.to_i > max_size
end
yield(res)
return
when Net::HTTPRedirection then
if limit == 0 then
raise Error.new("Too many redirects")
end
source = res["location"]
limit -= 1
else
raise Error.new("HTTP error code: #{res.code} #{res.message}")
end
end # http.request_get
end # http.start
end # while
end # def
def fix_image_board_sources
if source =~ /\/src\/\d{12,}|urnc\.yi\.org|yui\.cynthia\.bne\.jp/
self.source = "Image board"
end
end
end

View File

@@ -0,0 +1,84 @@
module Downloads
class File
class Error < Exception ; end
attr_accessor :source, :content_type, :file_path
def initialize(source, file_path)
@source = source
@file_path = file_path
end
def download!
http_get_streaming do |response|
self.content_type = response["Content-Type"]
::File.open(file_path, "wb") do |out|
response.read_body(out)
end
end
after_download
end
def before_download(url, headers)
Strategies::Base.strategies.each do |strategy|
url, headers = strategy.new.rewrite(url, headers)
end
return [url, headers]
end
def after_download
fix_image_board_sources
end
def http_get_streaming(options = {})
max_size = options[:max_size] || Danbooru.config.max_file_size
max_size = nil if max_size == 0 # unlimited
limit = 4
while true
url = URI.parse(source)
unless url.is_a?(URI::HTTP)
raise Error.new("URL must be HTTP")
end
Net::HTTP.start(url.host, url.port) do |http|
http.read_timeout = 10
headers = {
"User-Agent" => "#{Danbooru.config.safe_app_name}/#{Danbooru.config.version}"
}
@source, headers = before_download(@source, headers)
url = URI.parse(source)
http.request_get(url.request_uri, headers) do |res|
case res
when Net::HTTPSuccess then
if max_size
len = res["Content-Length"]
raise Error.new("File is too large (#{len} bytes)") if len && len.to_i > max_size
end
yield(res)
return
when Net::HTTPRedirection then
if limit == 0 then
raise Error.new("Too many redirects")
end
source = res["location"]
limit -= 1
else
raise Error.new("HTTP error code: #{res.code} #{res.message}")
end
end # http.request_get
end # http.start
end # while
end # def
def fix_image_board_sources
if source =~ /\/src\/\d{12,}|urnc\.yi\.org|yui\.cynthia\.bne\.jp/
@source = "Image board"
end
end
end
end

View File

@@ -0,0 +1,13 @@
module Downloads
module Strategies
class Base
def self.strategies
[Pixiv, Tinami, Pixa]
end
def rewrite(url, headers)
return [url, headers]
end
end
end
end

View File

@@ -0,0 +1,19 @@
module Downloads
module Strategies
class Pixa < Base
def rewrite(url, headers)
if url =~ /https?:\/\/(?:\w+\.)?pixa\.cc/
url, headers = rewrite_headers(url, headers)
end
return [url, headers]
end
protected
def rewrite_headers(url, headers)
headers["Referer"] = "http://www.pixa.cc"
return [url, headers]
end
end
end
end

View File

@@ -0,0 +1,58 @@
module Downloads
module Strategies
class Pixiv < Base
def rewrite(url, headers)
if url =~ /https?:\/\/(?:\w+\.)?pixiv\.net/
url, headers = rewrite_headers(url, headers)
url, headers = rewrite_small_images(url, headers)
url, headers = rewrite_small_manga_pages(url, headers)
end
return [url, headers]
end
protected
def rewrite_headers(url, headers)
headers["Referer"] = "http://www.pixiv.net"
return [url, headers]
end
def rewrite_small_images(url, headers)
if url =~ %r!(/img/.+?/.+?)_m.+$!
match = $1
url.sub!(match + "_m", match)
end
return [url, headers]
end
def rewrite_small_manga_pages(url, headers)
if url =~ %r!(\d+_p\d+)\.!
match = $1
repl = match.sub(/_p/, "_big_p")
big_url = url.sub(match, repl)
if http_exists?(big_url)
url = big_url
end
end
return [url, headers]
end
def http_exists?(url)
# example: http://img01.pixiv.net/img/as-special/15649262_big_p2.jpg
exists = false
uri = URI.parse(url)
Net::HTTP.start(uri.host, uri.port) do |http|
headers = {"Referer" => "http://www.pixiv.net", "User-Agent" => "#{Danbooru.config.app_name}/#{Danbooru.config.version}"}
http.request_head(uri.request_uri, headers) do |res|
if res.is_a?(Net::HTTPSuccess)
exists = true
end
end
end
exists
end
end
end
end

View File

@@ -0,0 +1,19 @@
module Downloads
module Strategies
class Tinami < Base
def rewrite(url, headers)
if url =~ /https?:\/\/(?:\w+\.)?tinami\.com/
url, headers = rewrite_headers(url, headers)
end
return [url, headers]
end
protected
def rewrite_headers(url, headers)
headers["Referer"] = "http://www.tinami.com/view"
return [url, headers]
end
end
end
end

View File

@@ -238,7 +238,7 @@ class Upload < ActiveRecord::Base
# Downloads the file to destination_path
def download_from_source(destination_path)
download = Download.new(source, destination_path)
download = Downloads::File.new(source, destination_path)
download.download!
self.file_path = destination_path
self.content_type = download.content_type || file_ext_to_content_type(source)

View File

@@ -1,54 +0,0 @@
require 'test_helper'
class DownloadTest < ActiveSupport::TestCase
context "A post download" do
setup do
@source = "http://www.google.com/intl/en_ALL/images/logo.gif"
@tempfile = Tempfile.new("danbooru-test")
@download = Download.new(@source, @tempfile.path)
end
teardown do
@tempfile.close
end
should "stream a file from an HTTP source" do
@download.http_get_streaming do |resp|
assert_equal("200", resp.code)
assert(resp["Content-Length"].to_i > 0, "File should be larger than 0 bytes")
end
end
should "throw an exception when the file is larger than the maximum" do
assert_raise(Download::Error) do
@download.http_get_streaming(:max_size => 1) do |resp|
end
end
end
should "store the file in the tempfile path" do
@download.download!
assert_equal(@source, @download.source)
assert(File.exists?(@tempfile.path), "temp file should exist")
assert(File.size(@tempfile.path) > 0, "should have data")
end
should "initialize the content type" do
@download.download!
assert_match(/image\/gif/, @download.content_type)
end
end
context "a post download for a pixiv manga page" do
setup do
@source = "http://img65.pixiv.net/img/kiyoringo/21755794_p2.png"
@tempfile = Tempfile.new("danbooru-test")
@download = Download.new(@source, @tempfile.path)
end
should "download the big version" do
@download.download!
assert_equal("http://img65.pixiv.net/img/kiyoringo/21755794_big_p2.png", @download.source)
end
end
end

View File

@@ -0,0 +1,43 @@
require 'test_helper'
module Downloads
class FileTest < ActiveSupport::TestCase
context "A post download" do
setup do
@source = "http://www.google.com/intl/en_ALL/images/logo.gif"
@tempfile = Tempfile.new("danbooru-test")
@download = Downloads::File.new(@source, @tempfile.path)
end
teardown do
@tempfile.close
end
should "stream a file from an HTTP source" do
@download.http_get_streaming do |resp|
assert_equal("200", resp.code)
assert(resp["Content-Length"].to_i > 0, "File should be larger than 0 bytes")
end
end
should "throw an exception when the file is larger than the maximum" do
assert_raise(Downloads::File::Error) do
@download.http_get_streaming(:max_size => 1) do |resp|
end
end
end
should "store the file in the tempfile path" do
@download.download!
assert_equal(@source, @download.source)
assert(::File.exists?(@tempfile.path), "temp file should exist")
assert(::File.size(@tempfile.path) > 0, "should have data")
end
should "initialize the content type" do
@download.download!
assert_match(/image\/gif/, @download.content_type)
end
end
end
end

View File

@@ -0,0 +1,18 @@
require 'test_helper'
module Downloads
class NicoSeigaTest < ActiveSupport::TestCase
context "a download for a pixa image" do
setup do
@source = "http://img.tinami.com/illust2/img/330/4e85ecd880a8f.jpg"
@tempfile = Tempfile.new("danbooru-test")
@download = Downloads::File.new(@source, @tempfile.path)
end
should "work" do
@download.download!
assert_equal(201248, ::File.size(@tempfile.path))
end
end
end
end

View File

@@ -0,0 +1,18 @@
require 'test_helper'
module Downloads
class PixaTest < ActiveSupport::TestCase
context "a download for a pixa image" do
setup do
@source = "http://file0.pixa.cc/illustrations/6f/d6/3f/f9/51/61/29/72/23/ac/middle/sse.jpg?1317405928"
@tempfile = Tempfile.new("danbooru-test")
@download = Downloads::File.new(@source, @tempfile.path)
end
should "work" do
@download.download!
assert_equal(104627, ::File.size(@tempfile.path))
end
end
end
end

View File

@@ -0,0 +1,35 @@
require 'test_helper'
module Downloads
class PixivTest < ActiveSupport::TestCase
context "a download for a pixiv manga page" do
setup do
@source = "http://img65.pixiv.net/img/kiyoringo/21755794_p2.png"
@tempfile = Tempfile.new("danbooru-test")
@download = Downloads::File.new(@source, @tempfile.path)
@download.download!
end
should "instead download the big version" do
assert_equal("http://img65.pixiv.net/img/kiyoringo/21755794_big_p2.png", @download.source)
end
end
context "a download for a small image" do
setup do
@source = "http://img02.pixiv.net/img/wanwandoh/4348318_m.jpg"
@tempfile = Tempfile.new("danbooru-test")
@download = Downloads::File.new(@source, @tempfile.path)
@download.download!
end
should "instead download the original version" do
assert_equal("http://img02.pixiv.net/img/wanwandoh/4348318.jpg", @download.source)
end
should "work" do
assert_equal(185778, ::File.size(@tempfile.path))
end
end
end
end

View File

@@ -0,0 +1,18 @@
require 'test_helper'
module Downloads
class TinamiTest < ActiveSupport::TestCase
context "a download for a pixa image" do
setup do
@source = "http://img.tinami.com/illust2/img/330/4e85ecd880a8f.jpg"
@tempfile = Tempfile.new("danbooru-test")
@download = Downloads::File.new(@source, @tempfile.path)
end
should "work" do
@download.download!
assert_equal(201248, ::File.size(@tempfile.path))
end
end
end
end