Your challenge is to turn a Japanese word and a dictionary pitch accent number into a new string where the rises and falls in pitch are marked: e.g. (2, ウシロ) →
To help you out with this, I'll explain a little about Japanese phonology.
Background: on moras
For the purpose of this challenge, we will write Japanese words in katakana, the simplest Japanese syllabary. A word consists of one or more moras, each of which consists of
optionally followed by one of:
For example, シャッキン consists of 4 moras: シャ, ッ, キ, ン.
※ The three optional characters are small versions of katakana
ヤユヨ (ya yu yo).
For example, キャ is ki + small ya, pronounced kya (1 mora), whereas キヤ is kiya (2 moras).
Background: pitch accents
Japanese words are described by a certain contour of high and low pitches. For example, ハナス (hanasu, speak) is pronounced
ナ / \ ハ ス–···
meaning the pitch goes up after ハ ha, then falls after ナ na, and stays low after ス su.
(That is to say: Unaccented grammar particles after su will be low again).
We might describe this contour as “low-high-low-(low…)”, or
LHL(L) for short.
In a dictionary, the pitch of this word would be marked as 2, because the pitch falls after the second mora. There are a few possible pitch patterns that occur in Tōkyō dialect Japanese, and they are all given a number:
0, which represents
LHHH…HH(H). This is heibangata, monotone form.
イ–バ–ン–··· ··· / or / ヘ ナ
1, which represents
HLLL…LL(H). This is atamadakagata, “head-high” form.
イ キ \ or \ ノ–チ–··· ···
n ≥ 2, which represents
LHHH…(length n) followed by a drop to
For example, センチメエトル has pitch accent number 4:
ン–チ–メ / \ セ エ–ト–ル–···
Such a word must have ≥ n moras, of course.
※ Note the difference between ハシ  and ハシ :
シ–··· シ / vs / \ ハ ハ ···
You can hear the different pitch accent numbers demonstrated here ♪.
Given an integer n ≥ 0 (a pitch accent number as above) and a valid word of at least n moras, insert
\ into it at points where those symbols would occur in the diagrams above; i.e. at points in the word where the pitch rises or falls.
The output for (2, ハナス) would be:
And the output for (0, ヘイバン) would be:
Remember: you must correctly handle sequences like キョ or チャ as a single mora.
The output for (1, キョウ) is
\, you may pick any other pair of distinct, non-katakana Unicode symbols.
This is code-golf: the objective is to write the shortest solution, measured in bytes.
One per line, in the format
n word → expected_output:
0 ナ → ナ/ 0 コドモ → コ/ドモ 0 ワタシ → ワ/タシ 0 ガッコウ → ガ/ッコウ 1 キ → キ\ 1 キヤ → キ\ヤ 1 キャ → キャ\ 1 ジショ → ジ\ショ 1 チュウゴク → チュ\ウゴク 1 ナニ → ナ\ニ 1 シュイ → シュ\イ 1 キョウ → キョ\ウ 1 キャンバス → キャ\ンバス 2 キヨカ → キ/ヨ\カ 2 キョカ → キョ/カ\ 2 ココロ → コ/コ\ロ 2 ジテンシャ → ジ/テ\ンシャ 3 センセイ → セ/ンセ\イ 3 タクサン → タ/クサ\ン 4 アタラシイ → ア/タラシ\イ 4 オトウト → オ/トウト\ 6 ジュウイチガツ → ジュ/ウイチガツ\
Generated by this reference solution.