nushell/crates/nu-command/src/strings/split/chars.rs
Stefan Holderbach 2aeb77bd3e
Fix output signature of split chars/words (#9739)
# Description
Those two commands did *not* vectorize over the input in the pure sense
as they performed a flat map. Now they return a list for each string
that gets split by them.

```
["foo" "bar"] | split chars
```

## Before 

```
╭───┬───╮
│ 0 │ f │
│ 1 │ o │
│ 2 │ o │
│ 3 │ b │
│ 4 │ a │
│ 5 │ r │
╰───┴───╯
```

## After
```
╭───┬───────────╮
│ 0 │ ╭───┬───╮ │
│   │ │ 0 │ f │ │
│   │ │ 1 │ o │ │
│   │ │ 2 │ o │ │
│   │ ╰───┴───╯ │
│ 1 │ ╭───┬───╮ │
│   │ │ 0 │ b │ │
│   │ │ 1 │ a │ │
│   │ │ 2 │ r │ │
│   │ ╰───┴───╯ │
╰───┴───────────╯
```
2023-07-23 17:06:41 -05:00

167 lines
5.1 KiB
Rust

use crate::grapheme_flags;
use nu_protocol::{
ast::Call,
engine::{Command, EngineState, Stack},
Category, Example, PipelineData, ShellError, Signature, Span, Type, Value,
};
use unicode_segmentation::UnicodeSegmentation;
#[derive(Clone)]
pub struct SubCommand;
impl Command for SubCommand {
fn name(&self) -> &str {
"split chars"
}
fn signature(&self) -> Signature {
Signature::build("split chars")
.input_output_types(vec![
(Type::String, Type::List(Box::new(Type::String))),
(
Type::List(Box::new(Type::String)),
Type::List(Box::new(Type::List(Box::new(Type::String)))),
),
])
.allow_variants_without_examples(true)
.switch("grapheme-clusters", "split on grapheme clusters", Some('g'))
.switch(
"code-points",
"split on code points (default; splits combined characters)",
Some('c'),
)
.category(Category::Strings)
}
fn usage(&self) -> &str {
"Split a string into a list of characters."
}
fn search_terms(&self) -> Vec<&str> {
vec!["character", "separate", "divide"]
}
fn examples(&self) -> Vec<Example> {
vec![
Example {
description: "Split the string into a list of characters",
example: "'hello' | split chars",
result: Some(Value::List {
vals: vec![
Value::test_string("h"),
Value::test_string("e"),
Value::test_string("l"),
Value::test_string("l"),
Value::test_string("o"),
],
span: Span::test_data(),
}),
},
Example {
description: "Split on grapheme clusters",
example: "'🇯🇵ほげ' | split chars -g",
result: Some(Value::List {
vals: vec![
Value::test_string("🇯🇵"),
Value::test_string(""),
Value::test_string(""),
],
span: Span::test_data(),
}),
},
Example {
description: "Split multiple strings into lists of characters",
example: "['hello', 'world'] | split chars",
result: Some(Value::test_list(vec![
Value::test_list(vec![
Value::test_string("h"),
Value::test_string("e"),
Value::test_string("l"),
Value::test_string("l"),
Value::test_string("o"),
]),
Value::test_list(vec![
Value::test_string("w"),
Value::test_string("o"),
Value::test_string("r"),
Value::test_string("l"),
Value::test_string("d"),
]),
])),
},
]
}
fn run(
&self,
engine_state: &EngineState,
_stack: &mut Stack,
call: &Call,
input: PipelineData,
) -> Result<PipelineData, ShellError> {
split_chars(engine_state, call, input)
}
}
fn split_chars(
engine_state: &EngineState,
call: &Call,
input: PipelineData,
) -> Result<PipelineData, ShellError> {
let span = call.head;
let graphemes = grapheme_flags(call)?;
input.map(
move |x| split_chars_helper(&x, span, graphemes),
engine_state.ctrlc.clone(),
)
}
fn split_chars_helper(v: &Value, name: Span, graphemes: bool) -> Value {
match v.span() {
Ok(v_span) => {
if let Ok(s) = v.as_string() {
Value::List {
vals: if graphemes {
s.graphemes(true)
.collect::<Vec<_>>()
.into_iter()
.map(move |x| Value::string(x, v_span))
.collect()
} else {
s.chars()
.collect::<Vec<_>>()
.into_iter()
.map(move |x| Value::string(x, v_span))
.collect()
},
span: v_span,
}
} else {
Value::Error {
error: Box::new(ShellError::PipelineMismatch {
exp_input_type: "string".into(),
dst_span: name,
src_span: v_span,
}),
}
}
}
Err(error) => Value::Error {
error: Box::new(error),
},
}
}
#[cfg(test)]
mod test {
use super::*;
#[test]
fn test_examples() {
use crate::test_examples;
test_examples(SubCommand {})
}
}