👻 Autumn's High Tech Study Log ✨ Learning Japanese at the speed of light with tech 🫨

Week 1 :partying_face:

Level 2 in Wanikani! I am going to take it a bit easy and not go overboard on lessons :slight_smile:

Jitai
I really LOVE Jitai on wanikani Jitai (字体): The font randomizer that fits I think it makes so much sense and it’s very fun.

I found this anki addon to do the same and some more https://ankiweb.net/shared/info/1476821676 , it works super well!! (fun fact: this works on all my cards. so now i have super weird non-japanese cards haha)
image
i could change this in the settings but maybe it’s good for me…

But hey! At least it’s fun fonts!

image

And now my cards have become a lil bit harder as I try to work out… What exactly is that supposed to be!

It is maybe working a bit too good.

Kanji God
My problem of “not knowing the kanji in my cards” might be solved by https://ankiweb.net/shared/info/1872210448

I installed it, but the cards it makes are so ugly (in the code). They encode it as base64 for some reason?

Physically they are pretty

I am thinking the card type just unmarshals this base64 and handles it, it makes it harder for me to edit it manually lol. not sure if i want that?

Essentially what I want to do:

  • make kanji anki cards from vocab cards i have mined (whether that’s look-a-head or just on the fly)
  • ability to add custom stories
  • know how to write it (important as i hand-write kanji)

This addon looks ahead to cards coming up (so if I see a card next week with a kanji, it will teach it to me now).

It also sees the kanji’s radicals & primitives and teaches me those first too, so I get like an order to my learning which is quite nice.

It also uses kanjii koohi stories which are always fun (some of their stories make me laugh out loud. being a fat rich man who greedily takes up all the space under the umbrella, but he’s not really fat. He’s just hiding a lot of gold under his clothes)

It does seem to work good. 辛 is a kanji I see frequently (personally) but it’s lvl44 in wanikani, it wouldn’t make much sense for me to wait 1+ years to see that before learning it. My goals really are to build up a baseline vocab / kanji (through wk etc) but also specifically the stuff I enjoy I need to learn :partying_face:

My only complaint is the format is a bit weird (it being base64), and that Migaku are not in my good books for creating software then abandoning it for their VC-funded SRS-suite which isn’t very good. I do not look forward to the day this addon breaks forever :frowning: (hopefully someone can & will maintain it though, as it is open source!)

Migaku

Migaku came out of the wood works with a cool update:

Previously my yomitan / asb / sharex setup was a bit troublesome. This makes it SO much easier!

The only bad thing is that because I am awkward and do not use the Migaku SRS (but rather send to Anki) I cannot easily keep track of what words I know…

I wrote a python script to use ankiconnect to go through my mature cards and print the words (sometimes sentences) I know. I think Migaku parses words… It then puts it into “known words” into anki (by hand) and I can highlight words I know. Anyway, here’s the script:

import requests
import json

url = "http://localhost:8765"

# JLAB
# gets cards
payload = json.dumps({
  "action": "findCards",
  "version": 6,
  "params": {
    "query": "deck:\"Default::Jlab's beginner course::Part 1: Listening comprehension\" prop:ivl>21"
  }
})
headers = {
  'Content-Type': 'application/json'
}

response = requests.request("POST", url, headers=headers, data=payload)
resp_txt = response.json()
# list of card IDs
card_ids = resp_txt["result"]

# get cards
payload = json.dumps({
  "action": "cardsInfo",
  "version": 6,
  "params": {
    "cards": card_ids
  }
})
headers = {
  'Content-Type': 'application/json'
}

response = requests.request("POST", url, headers=headers, data=payload)
resp_txt = response.json()

words = []

for i in resp_txt["result"]:
    words.append(i["fields"]["Jlab-Kanji"]['value'])


# Mining
# gets cards
payload = json.dumps({
  "action": "findCards",
  "version": 6,
  "params": {
    "query": "deck:\"Default::mining\" prop:ivl>21"
  }
})
headers = {
  'Content-Type': 'application/json'
}

response = requests.request("POST", url, headers=headers, data=payload)
resp_txt = response.json()
# list of card IDs
card_ids = resp_txt["result"]


payload = json.dumps({
  "action": "cardsInfo",
  "version": 6,
  "params": {
    "cards": card_ids
  }
})
headers = {
  'Content-Type': 'application/json'
}

response = requests.request("POST", url, headers=headers, data=payload)
resp_txt = response.json()


for i in resp_txt["result"]:
    try:
        words.append(i["fields"]["Key"]["value"])
    except error as e:
        print("failed to work on")
        print(i)

text = "\n".join(map(str, words))
print(text)

And for all cards (so I don’t add the same card twice. The plan is to use the above script to find i+1 content, and the below script to let any tools I am using know this word is in my queue and I do not need to re-add it to my queue :stuck_out_tongue: )

import requests
import json
import re

url = "http://localhost:8765"

# JLAB
# gets cards
payload = json.dumps({
  "action": "findCards",
  "version": 6,
  "params": {
    "query": "deck:\"Default::Jlab's beginner course::Part 1: Listening comprehension\""
  }
})
headers = {
  'Content-Type': 'application/json'
}

response = requests.request("POST", url, headers=headers, data=payload)
resp_txt = response.json()
# list of card IDs
card_ids = resp_txt["result"]

# get cards
payload = json.dumps({
  "action": "cardsInfo",
  "version": 6,
  "params": {
    "cards": card_ids
  }
})
headers = {
  'Content-Type': 'application/json'
}

response = requests.request("POST", url, headers=headers, data=payload)
resp_txt = response.json()

words = []

for i in resp_txt["result"]:
    try:
        words.append(i["fields"]["Jlab-Kanji"]['value']).strip()
    except:
        continue


# Mining
# gets cards
payload = json.dumps({
  "action": "findCards",
  "version": 6,
  "params": {
    "query": "deck:\"Default::mining\""
  }
})
headers = {
  'Content-Type': 'application/json'
}

response = requests.request("POST", url, headers=headers, data=payload)
resp_txt = response.json()
# list of card IDs
card_ids = resp_txt["result"]


payload = json.dumps({
  "action": "cardsInfo",
  "version": 6,
  "params": {
    "cards": card_ids
  }
})
headers = {
  'Content-Type': 'application/json'
}

response = requests.request("POST", url, headers=headers, data=payload)
resp_txt = response.json()


for i in resp_txt["result"]:
    # try excepts handle all my note types
    try:
        w = (i["fields"]["Key"]["value"])
        words.append(w.strip())
    except:
        try:
            w = (i['fields']['Word']['value'])
            words.append(w.strip())
        except:
            word = i['fields']['Target Word']['value']
            cleaned_string = re.sub(r'\[.*?\]|\s', '', word)
            words.append(cleaned_string)

text = "\n".join(map(str, words))

with open('output.txt', 'w', encoding='utf-8') as f:
    f.write(text)

print("wrote to file")

It’s ugly, but it works! I have 282 words which are mature! That’s so many for me haha.

I have around 2000 words total, about 160 i have mined and the rest are from prebuilt decks.

Thoughts on Migaku so far

I think it’s probably the best immersion tool for people who don’t like to get down & dirty with their tools. For me personally it does not have features I like (Jitai!, Kanji God, fsrs etc) and being “locked in” into an ecosystem isn’t the best.

Even in this forum post I took advantage of how open Anki is to get a list of all words I know, where “know” is what I define rather than someone else.

However, I LOVE how easy it is to use and how seamless it is compared to the other options out there. I am using it, but sending data to Anki instead :slight_smile:

immersion this week

  • suzume (eng subs)
  • demon slayer (eng subs. Yes I finished the entire swordsmith arc in 1 night haha)
  • Old Enough 1 episode (eng + jpn subs, sentence mining. the kids subs are built into the video so its a bit annoying, the narrator has his own proper soft-subs though!)
  • My neighbor totoro (jpn + eng subs, sentence mined) but only like 20 mins
  • shirokuma cafe episode 2 (jpn only, sentence mined)
  • The way of the house husband episodes 1, 2 & 3 (jpn, sentence mined)

The way of the house husband is so very funny. It’s about an ex-Yakuza boss who becomes a househusband and takes it very seriously.

  1. You learn a lot of words / vocab on everyday life as he’s a househusband. So he goes shopping, uses points cards, buys flowers, cleans etc.
  2. It’s very funny because he’s an ex-Yakuza boss and sometimes takes it a bit seriously, or people fight him becuase of his background and he’s like “No! My green onions!”

What didn’t go so well
No pimsleur or CKA training this week :frowning:

ToDo list for next week:

  • Don’t study sentence mine cards, instead do 10 kanji / radicals a day for cards in my sentence mining. I think because it includes radicals I am a bit “behind” on my sentence mined cards, so to get up to speed I should just do kanji next week.
  • This includes Wanikani too :slight_smile:
  • I have been doing a Katakana deck on Anki for a few weeks just to memorise them (as I found them a bit harder than Hiragana). I finish that next week!
  • Sentence mine more, it’s so much fun!
  • I should focus more on my job tbh. so I just think focus on anki + WK, and sentence mine if i have spare time!
  • I use 2 card types in anki, can I reduce this to 1? Does it matter?

to explore

  • My only problem is migaku not knowing what words i have already sentence mined… I can fix this by using their SRS, but I don’t wannnnaaaaa
  • they have this weird thing where u can track words. so learning words (only through their srs) = yellow, green = known (through srs or through importing known words), and tracked = red (only used in the UI for now). I am thinking each type i make a flashcard i track it, and when i look at it and go “i know what that is” i move it to known… lmaooo
  • Right now not much of a problem because of how few words I know. Will only be a problem in the future, and I am hoping some techniques will exist by then to handle this!

To review next week:

  • Is kanji god actually working? There’s a lot of kanji and when I search them I can’t entirely find what cards they’re from. Might be because they’re radicals / used in other kanji? Need to keep an eye on this :slight_smile:

Oh, and here’s some stats.


6 Likes