- Migrate all nodes to the V3 node API; old workflows migrate automatically - Add help pages for every node, opened from a ? icon in the title bar - Minor fixes and tooltip, placeholder and default improvements - Bump version to 3.0.0 Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
43 lines
1.3 KiB
Markdown
43 lines
1.3 KiB
Markdown
# ✂️ String Text Splitter
|
|
|
|
Splits text on a delimiter and/or on line breaks, and gives you both the first
|
|
piece and everything after it.
|
|
|
|
## Inputs
|
|
|
|
- **input_string** — The text to split.
|
|
- **delimiter** — The character to split on. Leave empty to split only on line
|
|
breaks (requires the toggle below).
|
|
- **split_at_linebreaks** — Also treat line breaks as split points. Works with
|
|
or without a delimiter.
|
|
- **target_indices** — Comma-separated 0-based positions to pull out, e.g.
|
|
`0,2`. Leave empty to skip this output.
|
|
|
|
## Outputs
|
|
|
|
- **first_chunk** — Everything before the first split point.
|
|
- **remainder** — Everything after it, still unsplit.
|
|
- **chunk_list** — Every piece, as a list. Empty pieces are dropped, so
|
|
repeated delimiters do not create blanks.
|
|
- **targeted_chunks** — Just the pieces named in `target_indices`, one per
|
|
line.
|
|
|
|
## How it works
|
|
|
|
With no delimiter and line-break splitting off, nothing is split: the text
|
|
comes out of `first_chunk` unchanged. Out-of-range target indices are ignored,
|
|
and a malformed `target_indices` prints a console warning rather than failing
|
|
the run.
|
|
|
|
## Examples
|
|
|
|
```
|
|
input_string: part1|part2|part3
|
|
delimiter: |
|
|
|
|
first_chunk: part1
|
|
remainder: part2|part3
|
|
chunk_list: [part1, part2, part3]
|
|
target_indices 0,2 -> targeted_chunks: "part1\npart3"
|
|
```
|