comiccrawler 2017.8.13

An image crawler with extendible modules and gui.

Comic Crawler 是用來扒圖的一支 Python Script。擁有簡易的下載管理員、圖書館功能、 與方便的擴充能力。

2017.3.25 更新

• 此版本用 node_vm2 取代 pyExecJs

  • execjs 可以用來執行 JavaScript，但是沒有任何的安全機制。若是從網站下載的 JavaScript 包含惡意程式（如︰刪光你的資料、破壞作業系統、病毒……等），以 execjs 執行是完全無法防止的。

  • node_vm2 用 vm2 執行 JavaScript，比 execjs 多了一層沙箱防護。

  • vm2 需要 Node.js >= 6。

2016.12.20 更新

• 此版本修改了檔案的命名規則

  • 原先檔名若包含不合法的字元 /\?|<>:"*，會被替換成底線 _

  • 改版後則會替換為對應的全型字元 ／＼？｜＜＞：＂＊

  • 若是有包含這類字元的存檔，會因為替換規則不同而讀不到舊版底線的檔案

  • 請在更新後執行 comiccrawler migrate 指令，會自動對舊版的存檔重命名

• 此版本修改了設定檔的格式

  • 項目名稱區分大小寫

  • 若是要求填入 cookie 資訊，會以 cookie_ 為前綴

2016.6.4 更新

• 此版本修改了存檔的運作方式，建議在更新前先將存檔備份

• 改版後，所有「未使用中」的任務資料會存到 ~/comiccrawler/pool/ 資料夾

• ~/comiccrawler/pool.json 不再儲存 episode 相關資訊

• 任務下載時，會再從 pool 資料夾中讀出 episode 相關資訊

• 目的為減少不必要的記憶體使用量

下載和安裝（Windows）

Comic Crawler is on PyPI. 安裝完 python 後，可以直接用 pip 指令自動安裝。

Install Python

你需要 Python 3.4 以上。安裝檔可以從它的 官方網站 下載。

安裝時記得要選「Add python.exe to path」，才能使用 pip 指令。

Install Node.js

Comic Crawler 使用 Node.js 來分析需要執行 JavaScript 的網站。

至少需要 6.0 以上的版本︰ https://nodejs.org/

Install Comic Crawler

在 cmd 底下輸入以下指令︰

pip install comiccrawler

更新時︰

pip install --upgrade comiccrawler

最後在 cmd 底下輸入以下指令執行 Comic Crawler︰

comiccrawler gui

Supported domains

99.hhxxee.com chan.sankakucomplex.com comic.acgn.cc comic.ck101.com comic.sfacg.com danbooru.donmai.us deviantart.com e-hentai.org exhentai.org ikanman.com imgbox.com konachan.com m.dmzj.com manhua.dmzj.com nijie.info raw.senmanga.com seiga.nicovideo.jp smp.yoedge.com tel.dm5.com tsundora.com tuchong.com tumblr.com tw.seemh.com tw.weibo.com wix.com www.8comic.com www.99comic.com www.aacomic.com www.buka.cn www.cartoonmad.com www.chuixue.com www.comicbus.com www.comicvip.com www.dm5.com www.dmzj.com www.facebook.com www.flickr.com www.hhcomic.cc www.hhssee.com www.hhxiee.com www.iibq.com www.pixiv.net www.seemh.com yande.re

使用說明

As a CLI tool:

Usage:
  comiccrawler [--profile=<profile>] (
    domains |
    download <url> [--dest=<save_path>] |
    gui |
    migrate
  )
  comiccrawler (--help | --version)

Commands:
  domains    列出支援的網址
  download   下載指定的 url
  gui        啟動主視窗
  migrate    將舊存檔更名為新存檔

Options:
  --profile  指定設定檔存放的資料夾（預設為 "~/comiccrawler"）
  --dest     設定下載目錄（預設為 "."）
  --help     顯示幫助訊息
  --version  顯示版本

or you can use it in your python script:

from comiccrawler.core import Mission, analyze, download

# create a mission
m = Mission(url="http://example.com")
analyze(m)

# select the episodes you want
for ep in m.episodes:
    if ep.title != "chapter 123":
        ep.skip = True

# download to savepath
download(m, "path/to/save")

圖形介面

• 在文字欄貼上網址後點「加入連結」或是按 Enter

• 若是剪貼簿裡有支援的網址，且文字欄同時是空的，程式會自動貼上

• 對著任務右鍵，可以選擇把任務加入圖書館。圖書館內的任務，在每次程式啟動時，都會檢查是否有更新。

設定檔

[DEFAULT]
; 設定下載完成後要執行的程式，會傳入下載資料夾的位置
runafterdownload =

; 啟動時自動檢查圖書館更新
libraryautocheck = true

; 下載目的資料夾。相對路徑會根據設定檔資料夾的位置。
savepath = download

; 開啟 grabber 偵錯
errorlog = false

; 每隔 5 分鐘自動存檔
autosave = 5

; 存檔時使用下載時的原始檔名而不用頁碼
originalfilename = false

; 自動轉換集數名稱中數字的格式，可以用於補0
; 例︰第1集 -> 第001集
; 詳細的格式指定方式請參考 https://docs.python.org/3/library/string.html#format-specification-mini-language
titlenumberformat = {:03d}

• 設定檔位於 ~\comiccrawler\setting.ini。可以在執行時指定 --profile 選項以變更預設的位置。（在 Windows 中 ~ 會被展開為 %HOME% 或 %USERPROFILE%）

• 執行一次 comiccrawler gui 後關閉，設定檔會自動產生

• 各別的網站會有自己的設定，通常是要填入一些登入相關資訊

• 設定檔會在重新啟動後生效。若 ComicCrawler 正在執行中，可以點「重載設定檔」來載入新設定

• 各別網站的設定不會互相影響。假如在 [DEFAULT] 設 savepath = a；在 [Pixiv] 設 savepath = b，那麼從 pixiv 下載的都會存到 b 資料夾，其它的就用預設值，存到 a 資料夾。

runafterdownload

• Pixiv Ugoku to MP4

Module example

Starting from version 2016.4.21, you can add your own module to ~/comiccrawler/mods/module_name.py.

#! python3
"""
This is an example to show how to write a comiccrawler module.

"""

import re
from urllib.parse import urljoin
from comiccrawler.core import Episode
from configparser import ConfigParser

# The header used in grabber method
header = {}

# The cookies
cookie = {}

# Match domain. Support sub-domain, which means "example.com" will match
# "*.example.com"
domain = ["www.example.com", "comic.example.com"]

# Module name
name = "Example"

# With noepfolder = True, Comic Crawler won't generate subfolder for each
# episode.
noepfolder = False

# Wait 5 seconds between each download.
rest = 5

# Specific user settings. The key is case-sensitive.
config = {
    # The config value can only be str
    "use_largest_image": "true",

    # These special config starting with `cookie__` will be automatically
    # used when grabbing html or image.
    "cookie_user": "user-default-value",
    "cookie_hash": "hash-default-value"
}

USE_LARGEST_IMAGE = True

def load_config():
    """This function will be called each time the config reloaded. Optional
    """
    global USE_LARGE_IMAGE
    USE_LARGE_IMAGE = ConfigParser.BOOLEAN_STATES.get(config["use_largest_image"].lower())

def get_title(html, url):
    """Return mission title.

    Title will be used in saving filepath, so be sure to avoid duplicate title.
    """
    return re.search("<h1 id='title'>(.+?)</h1>", html).group(1)

def get_episodes(html, url):
    """Return episode list.

    The episode list should be sorted by date, oldest first.
    If the episode list is multi-pages, specify the url of next page in
    get_next_page.
    """
    match_list = re.findall("<a href='(.+?)'>(.+?)</a>", html)
    return [Episode(title, urljoin(url, ep_url))
            for ep_url, title in match_list]

def get_images(html, url):
    """Get the URL of all images.

    The return value could be:

    -  A list of image.
    -  A generator yielding image.
    -  An image, when there is only one image in current page.

    Comic Crawler treats following types as an image:

    -  str - the url of the image
    -  callable - return an url when called
    -  comiccrawler.core.Image - use it to provide customized filename.

    While receiving the value, it is converted to a Image instance. See ``comiccrawler.core.Image.create()``.

    If the episode has multi-pages, specify the url of next page in
    get_next_page.

    Use generator in caution! If your generator raised any error between
    two images, next call to the generator will always result in
    StopIteration, which means that Comic Crawler will think it had crawled
    all images and navigate to next page. If you need to use grabhtml()
    between each pages (i.e. may raise HTTPError), you should return a list
    of callback!
    """
    return re.findall("<img src='(.+?)'>", html)

def get_next_page(html, url):
    """Return the url of the next page."""
    match = re.search("<a id='nextpage' href='(.+?)'>next</a>", html)
    if match:
        return match.group(1)

def errorhandler(error, crawler):
    """Downloader will call errorhandler if there is an error happened when
    downloading image. Normally you can just ignore this function.
    """
    pass

def imagehandler(ext, b):
    """If this function exist, Comic Crawler will call it before saving
    image to disk, letting the module be able to edit the image.

    @ext  A str of image extension. Including "." (e.g. ".jpg")
    @b    The bytes object of the image.

    It should return a (result_ext, result_b) tuple.
    """
    return (ext, b)

Todos

• Make grabber be able to return verbose info?

• Need a better error log system.

• Support pool in Sankaku.

• Add module.get_episode_id to make the module decide how to compare episodes.

• Use HEAD to grab final URL before requesting the image?

Changelog

• 2017.8.13

  • Fix: sankaku login issue. #66

• 2017.6.14

  • Fix: comicbus analzye issue.

• 2017.5.29

  • Fix: 99 module. #63

• 2017.5.26

  • Fix: ikanman analyze issue.

• 2017.5.22

  • Fix: comicbus analyze issue. #62

• 2017.5.19

  • Add nijie module. #58

  • Add core.clean_tags.

  • Fix: check update button doesn’t work after update checking failed. #59

  • Fix: analyzation failed in comicbus. #61

• 2017.5.5

  • Fix: use raw <title> as title in search result (pixiv).

  • Add .wmv, .mov, and .psd into valid file extensions.

• 2017.4.26

  • Change: use table view in dm5. #54

  • Fix: runafterdownload is parsed incorrectly on windows.

• 2017.4.24

  • Fix: starred expression inside list.

• 2017.4.23

  • Fix: compat with python 3.4, starred expression can only occur inside function call.

  • Update node_vm2 to 0.3.0.

• 2017.4.22

  • Add .bmp to valid file extensions.

  • Fix: unable to check update for multi-page sites.

• 2017.4.18

  • Add senmanga. #49

  • Add yoedge. #47

  • Fix: header parser issue. See https://www.ptt.cc/bbs/Python/M.1492438624.A.BBC.html

  • Fix: escape trailing dots in file path. #46

  • Add: double-click to launch explorer.

  • Add: batch analyze panel. #45

• 2017.4.6

  • Fix: run after download doesn’t work properly if path contains spaces.

  • Fix: VMError with ugoku in pixiv.

  • Fix: automatic update check doesn’t record update time when failing.

• 2017.4.3

  • Fix: analyze error in dA.

  • Fix: subdomain changed in exh.

  • Fix: vm error in hh.

  • Add .url utils, .core.CycleList, .error.HTTPError.

  • Add aacomic.

  • Update pyxcute to 0.4.1.

• 2017.3.26

  • Fix: cleanup the old files.

  • Update pythreadworker to 0.8.0.

• 2017.3.25

  • Switch to node_vm2, drop pyexecjs.

  • Add login check in exh.

  • Switch to pylint, drop pyflakes.

  • Drop module manhuadao.

  • Update pyxcute.

  • Refactor.

• 2017.3.9

  • Add –profile option. #36

• 2017.3.6

  • Update seemh. #35

  • Escape title in pixiv.

  • Strip non-printable characters in safefilepath.

• 2017.2.5

  • Add www.dmzj.com module. #33

  • Fix: Sometime the title doesn’t include chapter number in buka. #33

• 2017.1.10

  • Add: nowebp option in ikanman. #31

  • Add weibo module.

  • Add tuchong module.

  • Fix: update table safe_tk error.

  • Change: existence check will only check original filename when originalfilename option is true.

• 2017.1.6

  • Add: Table class in gui.

  • Add: titlenumberformat option in setting.ini. #30 by @kuanyui.

  • Change: use Table to display domain list.

• 2017.1.3.1

  • Fix: schema error (konachan).

  • Fix: original filename should be extracted from final url instead of request url.

  • Add: now the module can specify image filename with comiccrawler.core.Image.

• 2017.1.3

  • Fix: original option doesn’t work (exh).

• 2016.12.20

  • Change how config works. This will affect the sites requiring cookie information.

  • Comic Crawler can save cookie back to config now!

  • Change how safefilepath works. Use escape table.

  • Make io.move support folders.

  • Add io.exists.

  • Add migrate command.

  • Add originalfilename option.

• 2016.12.6

  • Fix: imghdr can’t reconize .webp in Python 3.4.

• 2016.12.1

  • Fix: analyze error in wix.

  • Fix: mimetypes.guess_extension is not reliable with application/octet-stream

  • Add .webp to valid file type.

• 2016.11.27

  • Fix hhxiee module. Use new domain www.hhssee.com.

• 2016.11.25

  • Support cartoonmad.

• 2016.11.2

  • Fix: scaling issue on Windows XP.

  • Fix: login-check in deviantart.

  • Use desktop3 to open folder. #16

  • Fix: GUI crahsed if scaling < 1.

• 2016.10.8

  • Fix: math.inf is only available in python 3.5.

• 2016.10.4

  • Fix: can not download video in flickr.

  • Fix: use cookie in grabimg.

• 2016.9.30

  • Add params option to grabber.

  • Add flickr module.

• 2016.9.27

  • Fix: image pattern in buka.

  • Fix: add hhcomic domain.

• 2016.9.11

  • Fix: failed to read file encoded with utf-8-sig.

  • Fix: ignore empty posts in tumblr.

• 2016.8.24.1

  • Use better method to find next page in tumblr.

  • Fix unicode referer bug in grabber.

  • Update match pattern to avoid redirect in tumblr. See https://github.com/kennethreitz/requests/issues/3078.

  • Fix get_title error in tumblr that the title might be empty.

• 2016.8.24

  • Fix 429 error still raised by analyze_info.

  • Fix next page pattern in tumblr.

• 2016.8.22

  • Support hhxiee.

  • Fix get_episodes error in ck101.

  • Suppress 429 error when analyzing.

  • Change title format in yendere. Support pools.

• 2016.8.19

  • Fix title not found error in dm5.

• 2016.8.8

  • Use a safer method in write_file.

  • Add mission_lock for thread safe.

  • Use str as runafterdownload.

  • Use float as autosave.

  • Add debug log.

  • Rewrite analyzer. Episodes shouldn’t have same title.

• 2016.7.2

  • Fix context menu popup bug on linux.

  • Fix update checking stops after finished mission.

• 2016.7.1

  • Use cross-platform startfile (incomplete).

  • Use clam theme for GUI under linux.

  • Fix the error message of update checking failure.

  • Update checking won’t block GUI thread anymore.

  • Update pythreadworker to 0.6.

  • Fix import syntax in gui.get_scale.

• 2016.6.30

  • Support high dpi displays.

  • Don’t show error in library thread. Only warn the user when update checking fails.

• 2016.6.25

  • API changed. Now the errorhandler will recieve (error, crawler) instead of (error, episode).

  • Add errorhandler in seemh. It will try to use different host if downloading failed.

  • Drop mission to the bottom when update checking failed. Update checking process will stop if it had retried 10 times.

• 2016.6.14.1

  • Pass pyflakes and fix a bunch of typo.

• 2016.6.14

  • Fix: always re-init in crawlpage loop!

• 2016.6.12

  • Use GBK instead of GB2312 in grabber.

  • Add the ability to get title from non-user page in nico.

  • Fix: unable to add mission in chuixue.

  • Fix: unable to download image in nico.

  • Fix: episode is lost after changing the name of the mission.

  • Fix: unable to recheck update after login error.

• 2016.6.10

  • Change how to handle HTTP 429 error. Let the mission drop.

  • Add login check in sankaku.

  • Support .jpe(.jpg), .webm file types.

• 2016.6.4

  • Change how saved data works. Comic Crawler will write inactive mission data into ~/comiccrawler/pool/ folder to save the memory.

  • Fix regex in dA.

  • Fix sankaku’s hang. Do not suppress 429 error in grabber.

• 2016.6.3

  • Minor change to save/load file function to avoid unnecessary copy.

  • Comic Crawler will now execute runafterdownload command both from the default section and the module section.

• 2016.5.30

  • Add module.imagehandler, which can edit the image file before saving to disk.

  • Write frame info into ugoku zip in pixiv.

• 2016.5.28

  • Change how config work. Now you can specify different setting in each sections. (e.g. use different savepath with different module)

  • Save frame info about ugoku in pixiv.

  • Drop config.update in module.load_config.

  • Try to support additional info in get_images.

• 2016.5.24

  • Support buka.

• 2016.5.20

  • Find server by executing js in seemh.

• 2016.5.15

  • Fix dependency scheme.

• 2016.5.2

  • Use Conten-Type header to guess file extension.

  • Fix a bug that the thread is not removed when recived DOWNLOAD_INVALID.

  • Pause download when meeting 509 error in exh.

  • Add .mp4 to valid file types.

• 2016.5.1.1

  • Fix a bug that Comic Crawler doesn’t retry when the first connection failed.

  • Add Episode.image, so the module can supply image list during constructing Episode.

• 2016.5.1

  • Support wix.com.

• 2016.4.27

  • Domain changed in seemh.

• 2016.4.26.1

  • Fix charset encoding bug.

• 2016.4.26

  • Fix config bug with upper-case key.

  • Check urls of old episodes to avoid unnecessary analyzing.

  • Add option to get original image in exh. It will cost 5x of viewing limit.

• 2016.4.22.3

  • Fix retry-after hanged bug.

  • Fix cnfig override bug. Use ComicCrawler section to replace DEFAULT section.

  • Support account login in sankaku.

  • Support HTTP error log before raising.

  • Show next page url while analyzing.

• 2016.4.22.2

  • Move to pythreadworker 0.5.0

• 2016.4.22.1

  • Support loading module in python3.4.

• 2016.4.22

  • Fix setup.py. Use find_packages.

• 2016.4.21

  • Big rewrite.

  • Move to requests.

  • Move to pythreadworker 0.4.0.

  • Add the ability to load module from ~/comiccrawler/mods

  • Drop migrate command.

• 2016.4.20

  • Update install_requires.

• 2016.4.13

  • Fix facebook bug.

  • Move to doit.

• 2016.4.8

  • Fix get_next_page error.

  • Fix key error in CLI.

• 2016.4.4

  • Use new API!

  • Analyzer will check the last episode to decide whether to analyze all pages.

  • Support multiple images in one page.

  • Change how getimgurl and getimgurls work.

• 2016.4.2

  • Add tumblr module.

  • Enhance: support sub-domain in mods.get_module.

• 2016.3.27

  • Fix: handle deleted post (konachan).

  • Fix: enhance dialog. try to fix #8.

• 2016.2.29

  • Fix: use latest comicview.js (8comic).

• 2016.2.27

  • Fix: lastcheckupdate doesn’t work.

  • Add: comicbus domain (8comic).

• 2016.2.15.1

  • Fix: can not add mission.

• 2016.2.15

  • Add lastcheckupdate setting. Now the library will only automatically check updates once a day.

  • Refactor. Use MissionProxy, Mission doesn’t inherit UserWorker anymore.

• 2016.1.26

  • Change: checking updates won’t affect mission which is downloading.

  • Fix: page won’t skip if the savepath contains “~”.

  • Add: a new url pattern in facebook.

• 2016.1.17

  • Fix: an url matching issue in Facebook.

  • Enhance: downloader will loop through other episodes rather than stop current mission on crawlpage error.

• 2016.1.15

  • Fix: ComicCrawler doesn’t save session during downloading.

• 2016.1.13

  • Handle HTTPError 429.

• 2016.1.12

  • Add facebook module.

  • Add circular option in module. Which should be set to True if downloader doesn’t know which is the last page of the album. (e.g. Facebook)

• 2016.1.3

  • Fix downloading failed in seemh.

• 2015.12.9

  • Fix build-time dependencies.

• 2015.11.8

  • Fix next page issue in danbooru.

• 2015.10.25

  • Support nico seiga.

  • Try to fix MemoryError when writing files.

• 2015.10.9

  • Fix unicode range error in gui. See http://is.gd/F6JfjD

• 2015.10.8

  • Fix an error that unable to skip episode in pixiv module.

• 2015.10.7

  • Fix errors that unable to create folder if title contains “{}” characters.

• 2015.10.6

  • Support search page in pixiv module.

• 2015.9.29

  • Support http://www.chuixue.com.

• 2015.8.7

  • Fixed sfacg bug.

• 2015.7.31

  • Fixed: libraryautocheck option does not work.

• 2015.7.23

  • Add module dmzj_m. Some expunged manga may be accessed from mobile page. http://manhua.dmzj.com/name => http://m.dmzj.com/info/name.html

• 2015.7.22

  • Fix bug in module eight.

• 2015.7.17

  • Fix episode selecting bug.

• 2015.7.16

  • Added:

    • Cleanup unused missions after session loads.

    • Handle ajax episode list in seemh.

    • Show an error if no update to download when clicking “download updates”.

    • Show an error if failing to load session.

  • Changed:

    • Always use “UPDATE” state if the mission is not complete after re-analyzing.

    • Create backup if failing to load session instead of moving them to “invalid-save” folder.

    • Check edit flag in MissionManager.save().

  • Fixed:

    • Can not download “updated” mission.

    • Update checking will stop on error.

    • Sankaku module is still using old method to create Episode.

• 2015.7.15

  • Add module seemh.

• 2015.7.14

  • Refactor: pull out download_manager, mission_manager.

  • Enhance content_write: use os.replace.

  • Fix mission_manager save loop interval.

• 2015.7.7

  • Fix danbooru bug.

  • Fix dmzj bug.

• 2015.7.6

  • Fix getepisodes regex in exh.

• 2015.7.5

  • Add error handler to dm5.

  • Add error handler to acgn.

• 2015.7.4

  • Support imgbox.

• 2015.6.22

  • Support tsundora.

• 2015.6.18

  • Fix url quoting issue.

• 2015.6.14

  • Enhance safeprint. Use echo command.

  • Enhance content_write. Add append=False option.

  • Enhance Crawler. Cache imgurl.

  • Enhance grabber. Add cookie=None option. Change errorlog behavior.

  • Fix grabber unicode encoding issue.

  • Some module update.

• 2015.6.13

  • Fix clean_finished

  • Fix console_download

  • Enhance get_by_state

Author

• eight [e-mail]