nushell/crates/nu-command/src/formats/from/ssv.rs
Christian Friedow 66ad83c15c
from ssv --aligned-columns should separate lines by character index instead of byte index (#8558)
# Description

## Symptom
Lines which are input into `from ssv --aligned-columns` are split
incorrectly of they contain utf-8 characters which have the length of
multiple bytes. Notice how the values of the `Bars` column bleeds into
the `Security` column in the following output (the big grey areas are
censored data ;) ):

![before-patch](https://user-images.githubusercontent.com/17351844/226757737-be7ca493-5c64-4a91-9153-984df515bb8c.png)

## Problem
The function behind `from ssv --aligned-columns` splits lines into
fields by byte index (which is default behavior of str.get(...) in Rust)
instead of character index. If the header row has a different length in
bytes than the remaining table rows, the split is executed incorrectly.

## Solution
The function behind `from ssv --aligned-columns1 now separates lines by
character index instead of byte index. This productes the following
(correct) output (the big grey areas are censored data ;) ):

![after-patch](https://user-images.githubusercontent.com/17351844/226757850-7acaebf3-2d40-4f85-b76e-64e465254bda.png)
2023-03-22 17:54:18 -05:00

509 lines
16 KiB
Rust

use indexmap::map::IndexMap;
use nu_engine::CallExt;
use nu_protocol::ast::Call;
use nu_protocol::engine::{Command, EngineState, Stack};
use nu_protocol::{
Category, Example, IntoPipelineData, PipelineData, ShellError, Signature, Span, Spanned,
SyntaxShape, Type, Value,
};
#[derive(Clone)]
pub struct FromSsv;
const DEFAULT_MINIMUM_SPACES: usize = 2;
impl Command for FromSsv {
fn name(&self) -> &str {
"from ssv"
}
fn signature(&self) -> Signature {
Signature::build("from ssv")
.input_output_types(vec![(Type::String, Type::Table(vec![]))])
.switch(
"noheaders",
"don't treat the first row as column names",
Some('n'),
)
.switch("aligned-columns", "assume columns are aligned", Some('a'))
.named(
"minimum-spaces",
SyntaxShape::Int,
"the minimum spaces to separate columns",
Some('m'),
)
.category(Category::Formats)
}
fn usage(&self) -> &str {
"Parse text as space-separated values and create a table. The default minimum number of spaces counted as a separator is 2."
}
fn examples(&self) -> Vec<Example> {
vec![Example {
example: r#"'FOO BAR
1 2' | from ssv"#,
description: "Converts ssv formatted string to table",
result: Some(Value::List { vals: vec![Value::Record { cols: vec!["FOO".to_string(), "BAR".to_string()], vals: vec![Value::test_string("1"), Value::test_string("2")], span: Span::test_data() }], span: Span::test_data() }),
}, Example {
example: r#"'FOO BAR
1 2' | from ssv -n"#,
description: "Converts ssv formatted string to table but not treating the first row as column names",
result: Some(
Value::List { vals: vec![Value::Record { cols: vec!["column1".to_string(), "column2".to_string()], vals: vec![Value::test_string("FOO"), Value::test_string("BAR")], span: Span::test_data() }, Value::Record { cols: vec!["column1".to_string(), "column2".to_string()], vals: vec![Value::test_string("1"), Value::test_string("2")], span: Span::test_data() }], span: Span::test_data() }),
}]
}
fn run(
&self,
engine_state: &EngineState,
stack: &mut Stack,
call: &Call,
input: PipelineData,
) -> Result<PipelineData, ShellError> {
from_ssv(engine_state, stack, call, input)
}
}
enum HeaderOptions<'a> {
WithHeaders(&'a str),
WithoutHeaders,
}
fn parse_aligned_columns<'a>(
lines: impl Iterator<Item = &'a str>,
headers: HeaderOptions,
separator: &str,
) -> Vec<Vec<(String, String)>> {
fn construct<'a>(
lines: impl Iterator<Item = &'a str>,
headers: Vec<(String, usize)>,
) -> Vec<Vec<(String, String)>> {
lines
.map(|l| {
headers
.iter()
.enumerate()
.map(|(i, (header_name, start_position))| {
let char_index_start = match l.char_indices().nth(*start_position) {
Some(idx) => idx.0,
None => *start_position,
};
let val = match headers.get(i + 1) {
Some((_, end)) => {
if *end < l.len() {
let char_index_end = match l.char_indices().nth(*end) {
Some(idx) => idx.0,
None => *end,
};
l.get(char_index_start..char_index_end)
} else {
l.get(char_index_start..)
}
}
None => l.get(char_index_start..),
}
.unwrap_or("")
.trim()
.into();
(header_name.clone(), val)
})
.collect()
})
.collect()
}
let find_indices = |line: &str| {
let values = line
.split(&separator)
.map(str::trim)
.filter(|s| !s.is_empty());
values
.fold(
(0, vec![]),
|(current_pos, mut indices), value| match line[current_pos..].find(value) {
None => (current_pos, indices),
Some(index) => {
let absolute_index = current_pos + index;
indices.push(absolute_index);
(absolute_index + value.len(), indices)
}
},
)
.1
};
let parse_with_headers = |lines, headers_raw: &str| {
let indices = find_indices(headers_raw);
let headers = headers_raw
.split(&separator)
.map(str::trim)
.filter(|s| !s.is_empty())
.map(String::from)
.zip(indices);
let columns = headers.collect::<Vec<(String, usize)>>();
construct(lines, columns)
};
let parse_without_headers = |ls: Vec<&str>| {
let mut indices = ls
.iter()
.flat_map(|s| find_indices(s))
.collect::<Vec<usize>>();
indices.sort_unstable();
indices.dedup();
let headers: Vec<(String, usize)> = indices
.iter()
.enumerate()
.map(|(i, position)| (format!("column{}", i + 1), *position))
.collect();
construct(ls.iter().map(|s| s.to_owned()), headers)
};
match headers {
HeaderOptions::WithHeaders(headers_raw) => parse_with_headers(lines, headers_raw),
HeaderOptions::WithoutHeaders => parse_without_headers(lines.collect()),
}
}
fn parse_separated_columns<'a>(
lines: impl Iterator<Item = &'a str>,
headers: HeaderOptions,
separator: &str,
) -> Vec<Vec<(String, String)>> {
fn collect<'a>(
headers: Vec<String>,
rows: impl Iterator<Item = &'a str>,
separator: &str,
) -> Vec<Vec<(String, String)>> {
rows.map(|r| {
headers
.iter()
.zip(r.split(separator).map(str::trim).filter(|s| !s.is_empty()))
.map(|(a, b)| (a.to_owned(), b.to_owned()))
.collect()
})
.collect()
}
let parse_with_headers = |lines, headers_raw: &str| {
let headers = headers_raw
.split(&separator)
.map(str::trim)
.map(str::to_owned)
.filter(|s| !s.is_empty())
.collect();
collect(headers, lines, separator)
};
let parse_without_headers = |ls: Vec<&str>| {
let num_columns = ls.iter().map(|r| r.len()).max().unwrap_or(0);
let headers = (1..=num_columns)
.map(|i| format!("column{i}"))
.collect::<Vec<String>>();
collect(headers, ls.into_iter(), separator)
};
match headers {
HeaderOptions::WithHeaders(headers_raw) => parse_with_headers(lines, headers_raw),
HeaderOptions::WithoutHeaders => parse_without_headers(lines.collect()),
}
}
fn string_to_table(
s: &str,
noheaders: bool,
aligned_columns: bool,
split_at: usize,
) -> Vec<Vec<(String, String)>> {
let mut lines = s.lines().filter(|l| !l.trim().is_empty());
let separator = " ".repeat(std::cmp::max(split_at, 1));
let (ls, header_options) = if noheaders {
(lines, HeaderOptions::WithoutHeaders)
} else {
match lines.next() {
Some(header) => (lines, HeaderOptions::WithHeaders(header)),
None => return vec![],
}
};
let f = if aligned_columns {
parse_aligned_columns
} else {
parse_separated_columns
};
f(ls, header_options, &separator)
}
fn from_ssv_string_to_value(
s: &str,
noheaders: bool,
aligned_columns: bool,
split_at: usize,
span: Span,
) -> Value {
let rows = string_to_table(s, noheaders, aligned_columns, split_at)
.iter()
.map(|row| {
let mut dict = IndexMap::new();
for (col, entry) in row {
dict.insert(
col.to_string(),
Value::String {
val: entry.to_string(),
span,
},
);
}
Value::from(Spanned { item: dict, span })
})
.collect();
Value::List { vals: rows, span }
}
fn from_ssv(
engine_state: &EngineState,
stack: &mut Stack,
call: &Call,
input: PipelineData,
) -> Result<PipelineData, ShellError> {
let name = call.head;
let noheaders = call.has_flag("noheaders");
let aligned_columns = call.has_flag("aligned-columns");
let minimum_spaces: Option<Spanned<usize>> =
call.get_flag(engine_state, stack, "minimum-spaces")?;
let (concat_string, _span, metadata) = input.collect_string_strict(name)?;
let split_at = match minimum_spaces {
Some(number) => number.item,
None => DEFAULT_MINIMUM_SPACES,
};
Ok(
from_ssv_string_to_value(&concat_string, noheaders, aligned_columns, split_at, name)
.into_pipeline_data_with_metadata(metadata),
)
}
#[cfg(test)]
mod tests {
use super::*;
fn owned(x: &str, y: &str) -> (String, String) {
(String::from(x), String::from(y))
}
#[test]
fn it_trims_empty_and_whitespace_only_lines() {
let input = r#"
a b
1 2
3 4
"#;
let result = string_to_table(input, false, true, 1);
assert_eq!(
result,
vec![
vec![owned("a", "1"), owned("b", "2")],
vec![owned("a", "3"), owned("b", "4")]
]
);
}
#[test]
fn it_deals_with_single_column_input() {
let input = r#"
a
1
2
"#;
let result = string_to_table(input, false, true, 1);
assert_eq!(result, vec![vec![owned("a", "1")], vec![owned("a", "2")]]);
}
#[test]
fn it_uses_first_row_as_data_when_noheaders() {
let input = r#"
a b
1 2
3 4
"#;
let result = string_to_table(input, true, true, 1);
assert_eq!(
result,
vec![
vec![owned("column1", "a"), owned("column2", "b")],
vec![owned("column1", "1"), owned("column2", "2")],
vec![owned("column1", "3"), owned("column2", "4")]
]
);
}
#[test]
fn it_allows_a_predefined_number_of_spaces() {
let input = r#"
column a column b
entry 1 entry number 2
3 four
"#;
let result = string_to_table(input, false, true, 3);
assert_eq!(
result,
vec![
vec![
owned("column a", "entry 1"),
owned("column b", "entry number 2")
],
vec![owned("column a", "3"), owned("column b", "four")]
]
);
}
#[test]
fn it_trims_remaining_separator_space() {
let input = r#"
colA colB colC
val1 val2 val3
"#;
let trimmed = |s: &str| s.trim() == s;
let result = string_to_table(input, false, true, 2);
assert!(result
.iter()
.all(|row| row.iter().all(|(a, b)| trimmed(a) && trimmed(b))));
}
#[test]
fn it_keeps_empty_columns() {
let input = r#"
colA col B col C
val2 val3
val4 val 5 val 6
val7 val8
"#;
let result = string_to_table(input, false, true, 2);
assert_eq!(
result,
vec![
vec![
owned("colA", ""),
owned("col B", "val2"),
owned("col C", "val3")
],
vec![
owned("colA", "val4"),
owned("col B", "val 5"),
owned("col C", "val 6")
],
vec![
owned("colA", "val7"),
owned("col B", ""),
owned("col C", "val8")
],
]
);
}
#[test]
fn it_can_produce_an_empty_stream_for_header_only_input() {
let input = "colA col B";
let result = string_to_table(input, false, true, 2);
let expected: Vec<Vec<(String, String)>> = vec![];
assert_eq!(expected, result);
}
#[test]
fn it_uses_the_full_final_column() {
let input = r#"
colA col B
val1 val2 trailing value that should be included
"#;
let result = string_to_table(input, false, true, 2);
assert_eq!(
result,
vec![vec![
owned("colA", "val1"),
owned("col B", "val2 trailing value that should be included"),
]]
);
}
#[test]
fn it_handles_empty_values_when_noheaders_and_aligned_columns() {
let input = r#"
a multi-word value b d
1 3-3 4
last
"#;
let result = string_to_table(input, true, true, 2);
assert_eq!(
result,
vec![
vec![
owned("column1", "a multi-word value"),
owned("column2", "b"),
owned("column3", ""),
owned("column4", "d"),
owned("column5", "")
],
vec![
owned("column1", "1"),
owned("column2", ""),
owned("column3", "3-3"),
owned("column4", "4"),
owned("column5", "")
],
vec![
owned("column1", ""),
owned("column2", ""),
owned("column3", ""),
owned("column4", ""),
owned("column5", "last")
],
]
);
}
#[test]
fn input_is_parsed_correctly_if_either_option_works() {
let input = r#"
docker-registry docker-registry=default docker-registry=default 172.30.78.158 5000/TCP
kubernetes component=apiserver,provider=kubernetes <none> 172.30.0.2 443/TCP
kubernetes-ro component=apiserver,provider=kubernetes <none> 172.30.0.1 80/TCP
"#;
let aligned_columns_noheaders = string_to_table(input, true, true, 2);
let separator_noheaders = string_to_table(input, true, false, 2);
let aligned_columns_with_headers = string_to_table(input, false, true, 2);
let separator_with_headers = string_to_table(input, false, false, 2);
assert_eq!(aligned_columns_noheaders, separator_noheaders);
assert_eq!(aligned_columns_with_headers, separator_with_headers);
}
#[test]
fn test_examples() {
use crate::test_examples;
test_examples(FromSsv {})
}
}