Generate a random sentence of length N from a given text. Logic:

1. Randomly select a word
2. The next word will be the subsequent contagious word from the previously selected word. If the next word occurs multiple times in the sentence, select randomly.
3. Continue until length fulfill required length

Special case: If the chosen word has no next word (last word in the sentence, use first word as the "next word" - assume circular)

Example
text = "hello this is a flexport interview and this is a hello and that was cool"
n = 4

• say first word randomly selected is "flexport"
• next word after "flexport" in the text is "interview". "interview" only occurs once.
• next word after "interview" in the text is "and". "and" occurs twice at index 6 & 11. Randomly select one of the indices. Say we select the one at index 6
• next word after "and" at index 6 is "this". "this" occurs at twice at index 1 & 7. Randomly selected at 7.

Output: "flexport interview and this"

Solution proposed

1. create hash of words and indexAt occurances

{
	"flexport": [4],
	"and": [6,11],
	"this": [1,7],
	...
}

2. use a counter and iteratively add new words into the output

// pick number x where 0 <= x < words.length
n = rand(words.length)

while count < length {
	ans += words[n]
	// compute next n
	indexAts = occurences[(n+1)%words.length]
	n = rand(indexAts.length)
	count++
}

Space complexity: O(n)
Time complexity: O(n)

Any thoughts of a better approach?