This uses multiple methods to save your personal data from a MAL (MyAnimeList) account, focused on [anime|manga]lists/episode history and forum posts your account has made.
I wanted to use the API whenever possible here, but the information returned by the API is so scarce, or endpoints don't really exist at all, so you can't really get a lot of info out of it. As far as I could figure out, it doesn't have a history endpoint, or any way to retrieve how many times you've rewatched a show, so this uses:
malexport update lists- Theload.jsonendpoint (unauthenticated) to backup myanime/mangalist (by most recently updated, as that's useful in many contexts)malexport update api-lists- The MAL API endpoint (authenticated) to backup myanime/mangalist. This includes a lot of metadata for each entry, and also works for private lists (assuming you go through the OAuth flow with the private account logged in your browser)- Selenium (so requires your MAL Username/Password; stored locally) to:
malexport update history- Individually grab episode/chapter history data (i.e., this). Note: the datetimes on these depend on what timezone you have set in your MAL settingsmalexport update export- Download the MAL export (the giant XML files), since those have rewatch information, and better datesmalexport update messages- Downloads/Updates your received and sent messages (DMs)
malexport update forum- Uses the MAL API (docs) to grab forum postsmalexport update friends- Uses Jikan to update your friends
The defaults here are far more on the safe side when scraping. If data fails to download you may have been flagged as a bot and may need to open MAL in your browser to solve a captcha.
For most people, this'll take a few hours to populate the initial cache, and then, and then a few minutes every few days (would recommend doing it at least once every 3 weeks, since it uses recent history to update history entries) to update it.
Requires python3.8+
To install with pip, run:
pip install malexport
For your MyAnimeList API Info, you can use 'other' as the 'App Type', 'hobbyist' as 'Purpose of Use', and http://localhost as the redirect URI. This only requires a Client ID, not both a Client ID and a Secret
Since this uses selenium, that requires a chromedriver binary somewhere on your system. That's typically available in package repositories, else see here. If this isn't able to find the file, set the MALEXPORT_CHROMEDRIVER_LOCATION environment variable, like: MALEXPORT_CHROMEDRIVER_LOCATION=C:\\Downloads\\chromedriver.exe malexport ...
I left some shell functions I commonly use to query my data in malexport.sh, to use that set the MAL_USERNAME variable to your account name, and then source malexport.sh in your shell startup. Should work on both bash/zsh
malexport update all can be run to run all the updaters or malexport update [forum|history|lists|export|friends|messages] can be run to update one of them. Each of those require you to pass a -u malUsername. This stores everything (except for the MAL API Client ID) on an account-by-account basis, so its possible to backup multiple accounts
If you want to hide the chromedriver, you can run this like MALEXPORT_CHROMEDRIVER_HIDDEN=1 malexport update ...
For the update lists command, this uses the unauthenticated load.json endpoint, which is what is used on modern lists as MAL. Therefore, its contents might be slightly different depending on your settings. To get the most info out of it, I'd recommend going to your list preferences and enabling all of the columns so that metadata is returned. Also, this assumes the European date format for lists.
Credentials are asked for the first time they're needed, and then stored in ~/.config/malexport (overwrite with MALEXPORT_CFG). Data by default is stored in ~/.local/share/malexport (overwrite with MALEXPORT_DIR). Lots of other things here are configurable with environment variables:
malexport/common.py:18:REQUEST_WAIT_TIME: int = int(os.environ.get("MALEXPORT_REQUEST_WAIT_TIME", 10))
malexport/exporter/messages.py:27:TILL_SAME_LIMIT = int(os.environ.get("MALEXPORT_THREAD_LIMIT", 10))
malexport/exporter/driver.py:26:HIDDEN_CHROMEDRIVER = bool(int(os.environ.get("MALEXPORT_CHROMEDRIVER_HIDDEN", 0)))
malexport/exporter/driver.py:27:CHROME_LOCATION: Optional[str] = os.environ.get("MALEXPORT_CHROMEDRIVER_LOCATION")
malexport/exporter/driver.py:30:TEMP_DOWNLOAD_BASE = os.environ.get("MALEXPORT_TEMPDIR", tempfile.gettempdir())
malexport/exporter/history.py:47:TILL_SAME_LIMIT = int(os.environ.get("MALEXPORT_EPISODE_LIMIT", 5))
malexport/exporter/export_downloader.py:21:TRY_EXPORT_TIMES = int(os.environ.get("MALEXPORT_EXPORT_TRIES", 3))
malexport/exporter/mal_session.py:31:MALEXPORT_REDIRECT_URI = os.environ.get("MALEXPORT_REDIRECT_URI", "http://localhost")
malexport/paths.py:24:mal_id_cache_dir = os.environ.get("MAL_ID_CACHE_DIR", os.path.join(cache_dir, "mal-id-cache"))
malexport/paths.py:29: os.environ.get("MALEXPORT_ZIP_BACKUPS", os.path.join(local_directory, "malexport_zips"))
malexport/parse/common.py:30:CUTOFF_DATE = int(os.environ.get("MALEXPORT_CUTOFF_DATE", date.today().year + 5))
To show debug logs set export MALEXPORT_LOGS=10 (uses logging levels).
If you use 2FA you can set the MALEXPORT_2FA variable, like MALEXPORT_2FA=1 malexport update ... when running this, that adds a prompt to wait for you to login before continuing
I generally don't interface with the CLI interface here and instead use the my.mal.export in HPI. That handles configuring accounts/locating my data on disk
The parse subcommand includes corresponding commands which take the saved data, clean it up a bit into easier to manipulate representations. Those each have python functions which can be imported from malexport.parse, or called from the CLI to produce JSON.
The most useful is probably combine, which combines the xml, api-lists, history and lists data.
Otherwise, this acts on the data files (Reminder that data by default is stored in ~/.local/share/malexport):
$ malexport parse xml ./animelist.xml | jq '.entries[106]'
{
"anime_id": 31646,
"title": "3-gatsu no Lion",
"media_type": "TV",
"episodes": 22,
"my_id": 0,
"watched_episodes": 22,
"start_date": "2020-07-01",
"finish_date": "2020-08-09",
"rated": null,
"score": 9,
"storage": null,
"storage_value": 0,
"status": "Completed",
"comments": "",
"times_watched": 0,
"rewatch_value": null,
"priority": "LOW",
"tags": "",
"rewatching": false,
"rewatching_ep": 0,
"discuss": true,
"sns": "default",
"update_on_import": false
}parse list converts some of the status int enumerations (status/airing status) into the corresponding string values, and parses date strings like '04-09-20' to '09-04-2020':
malexport parse list ./animelist.json | jq '.[0]':
{
"status": "On Hold",
"score": 6,
"tags": "Slice of Life",
"rewatching": false,
"watched_episodes": 8,
"title": "Shiroi Suna no Aquatope",
"episodes": 24,
"airing_status": "Currently Airing",
"id": 46093,
"studios": [
{
"id": 132,
"name": "P.A. Works"
}
],
"licensors": [],
"season": {
"year": 2021,
"season": "Summer"
},
"has_episode_video": true,
"has_promotion_video": true,
"has_video": true,
"video_url": "/anime/46093/Shiroi_Suna_no_Aquatope/video",
"url": "/anime/46093/Shiroi_Suna_no_Aquatope",
"image_path": "https://cdn.myanimelist.net/r/96x136/images/anime/1932/114952.jpg?s=12d30d08dd16eb006e02f73d9dc14a8f",
"is_added_to_list": false,
"media_type": "TV",
"rating": "PG-13",
"start_date": "2021-07-10",
"finish_date": null,
"air_start_date": "2021-07-09",
"air_end_date": null,
"days": 53,
"storage": "",
"priority": "Low"
}If you want exact dates, I'd recommend using the xml export, as there's some estimation that has to done for the list export since the dates aren't absolute (e.g. 04-09-20 could be 2020 or 1920)
malexport parse forum -u malUsername extracts posts made by your user to JSON
malexport parse history -u malUsername | jq '.[] | select(.title == "Akira")'
{
"mal_id": 47,
"list_type": "anime",
"title": "Akira",
"entries": [
{
"at": "2016-02-02 21:47:00",
"number": 1
}
]
}'number' in this case refers to the chapter or episode number
As some random examples, using this from the python, or the CLI:
Which season do I have the most completed from?
>>> Counter([a.season for a in malexport.parse.parse_list("animelist.json", malexport.parse.ListType.ANIME) if a.score is not None and a.status == "Completed" if a.season is not None]).most_common(1)
[(Season(year=2016, season='Spring'), 73)]Or, you can use jq to mangle it into whatever you want. Heres a mess of pipes to create a graph of your Completed ratings, using termgraph:
$ malexport parse list ./animelist.json | jq '.[] | select(.status == "Completed") | .score' | grep -vx 0 | sort | uniq -c | awk '{ print $2 " " $1}' | termgraph | sort -n
1 : ▇▇▇▇▇▇▇▇▇ 158.00
2 : ▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 652.00
3 : ▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 847.00
4 : ▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 791.00
5 : ▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 562.00
6 : ▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 384.00
7 : ▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 263.00
8 : ▇▇▇▇▇▇ 103.00
9 : ▇▇ 47.00
10: ▏ 5.00
This includes a command to recover deleted MAL entries (in other words, a MAL moderator completely deleted the entry from the site, which automatically removes it from your list) which you previously had on your list, by recovering deleted items from zipfiles.
This requires hpi to automatically unzip files, install with python3 -m pip install malexport[recover] or directly with pip install hpi
malexport recover-deleted backup, saves to ~/.local/share/malexport_zips (can overwrite default location with MALEXPORT_ZIP_BACKUPS)
To figure out which entries are deleted, this uses mal-id-cache. To update the local cache of IDs, run:
$ malexport recover-deleted approved-update
Updated mal-id-cache to commit 9c0cbdeac567671c0970c79ee99531edc2d89b0b
$ malexport recover-deleted approved-ids-stats
Approved Anime: 23930
Approved Manga: 62212Then, you can run malexport recover-deleted recover to find deleted entries:
malexport recover-deleted recover
For example, mine look like:
python3 -m malexport recover-deleted recover -F \
| jq 'values[].anime | .[].XMLData | "\(.anime_id) \(.title) \(.score)/10 \(.start_date)"' -r6852 Ahiru no Otegara 3/10 2016-09-08
42142 X-Men Openings 4/10 2020-06-04
37879 Benghuai 3: Reburn 4/10 2018-09-14
38411 Eiga Daisuki Pompo-san 2/10 2018-09-27
29293 Isu 6/10 2016-09-10
38426 Koutetsujou no Kabaneri: Ran - Hajimaru Michiato 4/10 2019-02-13
10584 Mononoke Dance 4/10 2017-01-15
29949 Nami 1/10 2016-08-12
13675 Taisei Kensetsu: Bosporus Kaikyou Tunnel 4/10 2016-09-02
38065 Taisei Kensetsu: Singapore 4/10 2018-07-31
21441 Taisei Kensetsu: Sri Lanka Kousokudouro 3/10 2016-10-24
25883 Taisei Kensetsu: Vietnam Noi Bai Kuukou 4/10 2016-09-02
33234 Kaibutsu-kun: Suna Majin wo Yattsukero no Maki / Kaibutsu-kun to Haniwa Kaishin no Maki 4/10 2018-09-27
30245 Kamaishi no "Kiseki": Inochi wo Mamoru Tokubetsu Jugyou 2/10 2016-12-21
23399 Minami no Shima no Dera-chan 3/10 2016-04-26
None of those IDs exist anymore on MAL, so these backups are the only way to get metadata or my history/data for them
I backup my list once every 3 days, and have a corresponding bleanser (backup cleanser) file to remove redundant backups (ones that don't introduce 'new deleted' entries).
The manual-history command lets you locally save episode/chapter (history) data for anime or manga.
Since its not trivial to mark a single episode as watched on MAL; you have to:
- go to the anime page
- update the entry to the episode you watched
- then, reset the episode count to completed
- then, go to advanced edit, click the 'History' button and remove the latest entry created by you resetting the episode count
Sometimes I just rewatch particular episodes and not an entire show, this lets me do that and save that to a file.
That requires autotui and pyfzf-iter, which can be installed with pip install malexport[manual]
There is a parse command to parse the manual history file, and its combined into the history data when using combine