Function chunkText

chunkText(text: string, max_length: number, separators?: string[]): string[]
Splits text into trimmed, non-empty chunks, preferring separators over forced splits.
Parameters
- text: string
  The text to split.
- max_length: number
  Maximum chunk length in UTF-16 code units (positive safe integer).
- separators: string[] = ...
  Literal split points in priority order (default: ['\n', ' ', '.']).
Returns string[]
The trimmed, non-empty chunks.
Throws
If max_length is not a positive safe integer.
Remarks
- For text exceeding the limit, uses the rightmost eligible occurrence of the first matching separator: its start must be at least Math.floor(max_length / 2) and its end within the limit. Separators stay in the preceding chunk unless trimmed.
- Trims every chunk and skips empty chunks.
- Never splits surrogate pairs; a non-BMP code point with max_length = 1 is the sole length-limit exception (2 units). Unpaired surrogates are preserved.
- Code-point safety is not grapheme safety: combining marks and ZWJ sequences may split.
Example
```
chunkText('hello world', 5);     // ['hello', 'world']
chunkText('hello world', 7);     // ['hello', 'world']
chunkText('abcdefghij', 5);      // ['abcde', 'fghij']
chunkText('a b c d e', 4);       // ['a b', 'c d', 'e']
```
- Defined in string/chunk.ts:27