1882-Add min, max in addition to average for acting lists (#1969)

* Converting average.rs to math.rs

* Making some progress towards math

Examples and unit tests failing, also think I found a bug with passing in strings

* Fix typos

* Found issue with negative numbers

* Add some comments

* Split commands like in split and str_ but do not register?

* register commands in cli

* Address clippy warnings

* Fix bad examples

* Make the example failure message more helpful

* Replace unwraps

* Use compare_values to avoid coercion issues

* Remove unneeded code
This commit is contained in:
Arash Outadi 2020-06-13 14:49:57 -07:00 committed by GitHub
parent 40673e4599
commit a042f407c1
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
12 changed files with 440 additions and 79 deletions

View File

@ -345,7 +345,10 @@ pub fn create_default_context(
whole_stream_command(Headers),
// Data processing
whole_stream_command(Histogram),
whole_stream_command(Math),
whole_stream_command(Average),
whole_stream_command(Minimum),
whole_stream_command(Maximum),
whole_stream_command(Sum),
// File format output
whole_stream_command(To),

View File

@ -8,7 +8,6 @@ pub(crate) mod alias;
pub(crate) mod append;
pub(crate) mod args;
pub(crate) mod autoview;
pub(crate) mod average;
pub(crate) mod build_string;
pub(crate) mod cal;
pub(crate) mod calc;
@ -68,6 +67,7 @@ pub(crate) mod lines;
pub(crate) mod ls;
#[allow(unused)]
pub(crate) mod map_max_by;
pub(crate) mod math;
pub(crate) mod merge;
pub(crate) mod mkdir;
pub(crate) mod mv;
@ -135,7 +135,6 @@ pub(crate) use command::{
pub(crate) use alias::Alias;
pub(crate) use append::Append;
pub(crate) use average::Average;
pub(crate) use build_string::BuildString;
pub(crate) use cal::Cal;
pub(crate) use calc::Calc;
@ -151,6 +150,7 @@ pub(crate) use du::Du;
pub(crate) use each::Each;
pub(crate) use echo::Echo;
pub(crate) use is_empty::IsEmpty;
pub(crate) use math::Math;
pub(crate) use update::Update;
pub(crate) mod kill;
pub(crate) use kill::Kill;
@ -198,6 +198,7 @@ pub(crate) use lines::Lines;
pub(crate) use ls::Ls;
#[allow(unused_imports)]
pub(crate) use map_max_by::MapMaxBy;
pub(crate) use math::{Average, Maximum, Minimum};
pub(crate) use merge::Merge;
pub(crate) use mkdir::Mkdir;
pub(crate) use mv::Move;

View File

@ -1,28 +1,28 @@
use crate::commands::math::utils::calculate;
use crate::commands::WholeStreamCommand;
use crate::prelude::*;
use crate::utils::data_processing::{reducer_for, Reduce};
use bigdecimal::FromPrimitive;
use bigdecimal::{FromPrimitive, Zero};
use nu_errors::ShellError;
use nu_protocol::hir::{convert_number_to_u64, Number, Operator};
use nu_protocol::{Dictionary, Primitive, ReturnSuccess, Signature, UntaggedValue, Value};
use num_traits::identities::Zero;
use nu_protocol::{
hir::{convert_number_to_u64, Number, Operator},
Primitive, Signature, UntaggedValue, Value,
};
use indexmap::map::IndexMap;
pub struct Average;
pub struct SubCommand;
#[async_trait]
impl WholeStreamCommand for Average {
impl WholeStreamCommand for SubCommand {
fn name(&self) -> &str {
"average"
"math average"
}
fn signature(&self) -> Signature {
Signature::build("average")
Signature::build("math average")
}
fn usage(&self) -> &str {
"Average the values."
"Gets the average of a list of numbers"
}
async fn run(
@ -30,75 +30,32 @@ impl WholeStreamCommand for Average {
args: CommandArgs,
registry: &CommandRegistry,
) -> Result<OutputStream, ShellError> {
average(RunnableContext {
input: args.input,
registry: registry.clone(),
shell_manager: args.shell_manager,
host: args.host,
ctrl_c: args.ctrl_c,
current_errors: args.current_errors,
name: args.call_info.name_tag,
raw_input: args.raw_input,
})
calculate(
RunnableContext {
input: args.input,
registry: registry.clone(),
shell_manager: args.shell_manager,
host: args.host,
ctrl_c: args.ctrl_c,
current_errors: args.current_errors,
name: args.call_info.name_tag,
raw_input: args.raw_input,
},
average,
)
.await
}
fn examples(&self) -> Vec<Example> {
vec![Example {
description: "Average a list of numbers",
example: "echo [100 0 100 0] | average",
result: Some(vec![UntaggedValue::decimal(50).into()]),
description: "Get the average of a list of numbers",
example: "echo [-50 100.0 25] | math average",
result: Some(vec![UntaggedValue::decimal(25).into()]),
}]
}
}
async fn average(
RunnableContext {
mut input, name, ..
}: RunnableContext,
) -> Result<OutputStream, ShellError> {
let values: Vec<Value> = input.drain_vec().await;
if values.iter().all(|v| v.is_primitive()) {
match avg(&values, name) {
Ok(result) => Ok(OutputStream::one(ReturnSuccess::value(result))),
Err(err) => Err(err),
}
} else {
let mut column_values = IndexMap::new();
for value in values {
if let UntaggedValue::Row(row_dict) = value.value {
for (key, value) in row_dict.entries.iter() {
column_values
.entry(key.clone())
.and_modify(|v: &mut Vec<Value>| v.push(value.clone()))
.or_insert(vec![value.clone()]);
}
}
}
let mut column_totals = IndexMap::new();
for (col_name, col_vals) in column_values {
match avg(&col_vals, &name) {
Ok(result) => {
column_totals.insert(col_name, result);
}
Err(err) => return Err(err),
}
}
Ok(OutputStream::one(ReturnSuccess::value(
UntaggedValue::Row(Dictionary {
entries: column_totals,
})
.into_untagged_value(),
)))
}
}
fn avg(values: &[Value], name: impl Into<Tag>) -> Result<Value, ShellError> {
let name = name.into();
pub fn average(values: &[Value], name: &Tag) -> Result<Value, ShellError> {
let sum = reducer_for(Reduce::Sum);
let number = BigDecimal::from_usize(values.len()).expect("expected a usize-sized bigdecimal");
@ -156,12 +113,12 @@ fn avg(values: &[Value], name: impl Into<Tag>) -> Result<Value, ShellError> {
#[cfg(test)]
mod tests {
use super::Average;
use super::SubCommand;
#[test]
fn examples_work_as_expected() {
use crate::examples::test as test_examples;
test_examples(Average {})
test_examples(SubCommand {})
}
}

View File

@ -0,0 +1,135 @@
use crate::commands::WholeStreamCommand;
use crate::prelude::*;
use nu_errors::ShellError;
use nu_protocol::{ReturnSuccess, Signature, UntaggedValue};
pub struct Command;
#[async_trait]
impl WholeStreamCommand for Command {
fn name(&self) -> &str {
"math"
}
fn signature(&self) -> Signature {
Signature::build("math")
}
fn usage(&self) -> &str {
r#"Use mathematical functions (average, min, max) to aggregate list of numbers or tables
math average
math min
math max
"#
}
async fn run(
&self,
_args: CommandArgs,
registry: &CommandRegistry,
) -> Result<OutputStream, ShellError> {
Ok(OutputStream::one(Ok(ReturnSuccess::Value(
UntaggedValue::string(crate::commands::help::get_help(&Command, &registry.clone()))
.into_value(Tag::unknown()),
))))
}
}
#[cfg(test)]
mod tests {
use super::*;
use crate::commands::math::{
average::average, max::maximum, min::minimum, utils::MathFunction,
};
use nu_plugin::test_helpers::value::{decimal, int};
use nu_protocol::Value;
#[test]
fn examples_work_as_expected() {
use crate::examples::test as test_examples;
test_examples(Command {})
}
#[test]
fn test_math_functions() {
struct TestCase {
description: &'static str,
values: Vec<Value>,
expected_err: Option<ShellError>,
// Order is: avg, min, max
expected_res: Vec<Result<Value, ShellError>>,
}
let tt: Vec<TestCase> = vec![
TestCase {
description: "Empty data should throw an error",
values: Vec::new(),
expected_err: Some(ShellError::unexpected("Expected data")),
expected_res: Vec::new(),
},
TestCase {
description: "Single value",
values: vec![int(10)],
expected_err: None,
expected_res: vec![Ok(decimal(10)), Ok(int(10)), Ok(int(10))],
},
TestCase {
description: "Multiple Values",
values: vec![int(10), int(30), int(20)],
expected_err: None,
expected_res: vec![Ok(decimal(20)), Ok(int(10)), Ok(int(30))],
},
TestCase {
description: "Mixed Values",
values: vec![int(10), decimal(26.5), decimal(26.5)],
expected_err: None,
expected_res: vec![Ok(decimal(21)), Ok(int(10)), Ok(decimal(26.5))],
},
TestCase {
description: "Negative Values",
values: vec![int(10), int(-11), int(-14)],
expected_err: None,
expected_res: vec![Ok(decimal(-5)), Ok(int(-14)), Ok(int(10))],
},
TestCase {
description: "Mixed Negative Values",
values: vec![int(10), decimal(-11.5), decimal(-13.5)],
expected_err: None,
expected_res: vec![Ok(decimal(-5)), Ok(decimal(-13.5)), Ok(int(10))],
},
// TODO-Uncomment once Issue: https://github.com/nushell/nushell/issues/1883 is resolved
// TestCase {
// description: "Invalid Mixed Values",
// values: vec![int(10), decimal(26.5), decimal(26.5), string("math")],
// expected_err: Some(ShellError::unimplemented("something")),
// expected_res: vec![],
// },
];
let test_tag = Tag::unknown();
for tc in tt.iter() {
let tc: &TestCase = tc; // Just for type annotations
let math_functions: Vec<MathFunction> = vec![average, minimum, maximum];
let results = math_functions
.iter()
.map(|mf| mf(&tc.values, &test_tag))
.collect_vec();
if tc.expected_err.is_some() {
assert!(
results.iter().all(|r| r.is_err()),
"Expected all functions to error for test-case: {}",
tc.description,
);
} else {
for (i, res) in results.into_iter().enumerate() {
assert_eq!(
res, tc.expected_res[i],
"math function {} failed on test-case {}",
i, tc.description
);
}
}
}
}
}

View File

@ -0,0 +1,69 @@
use crate::commands::math::utils::calculate;
use crate::commands::WholeStreamCommand;
use crate::prelude::*;
use crate::utils::data_processing::{reducer_for, Reduce};
use nu_errors::ShellError;
use nu_protocol::{Signature, UntaggedValue, Value};
pub struct SubCommand;
#[async_trait]
impl WholeStreamCommand for SubCommand {
fn name(&self) -> &str {
"math max"
}
fn signature(&self) -> Signature {
Signature::build("math max")
}
fn usage(&self) -> &str {
"Get the maximum of a list of numbers or tables"
}
async fn run(
&self,
args: CommandArgs,
registry: &CommandRegistry,
) -> Result<OutputStream, ShellError> {
calculate(
RunnableContext {
input: args.input,
registry: registry.clone(),
shell_manager: args.shell_manager,
host: args.host,
ctrl_c: args.ctrl_c,
current_errors: args.current_errors,
name: args.call_info.name_tag,
raw_input: args.raw_input,
},
maximum,
)
.await
}
fn examples(&self) -> Vec<Example> {
vec![Example {
description: "Find the maximum of list of numbers",
example: "echo [-50 100 25] | math max",
result: Some(vec![UntaggedValue::int(100).into()]),
}]
}
}
pub fn maximum(values: &[Value], _name: &Tag) -> Result<Value, ShellError> {
let max_func = reducer_for(Reduce::Maximum);
max_func(Value::nothing(), values.to_vec())
}
#[cfg(test)]
mod tests {
use super::SubCommand;
#[test]
fn examples_work_as_expected() {
use crate::examples::test as test_examples;
test_examples(SubCommand {})
}
}

View File

@ -0,0 +1,69 @@
use crate::commands::math::utils::calculate;
use crate::commands::WholeStreamCommand;
use crate::prelude::*;
use crate::utils::data_processing::{reducer_for, Reduce};
use nu_errors::ShellError;
use nu_protocol::{Signature, UntaggedValue, Value};
pub struct SubCommand;
#[async_trait]
impl WholeStreamCommand for SubCommand {
fn name(&self) -> &str {
"math min"
}
fn signature(&self) -> Signature {
Signature::build("math min")
}
fn usage(&self) -> &str {
"Finds the minimum within a list of numbers or tables"
}
async fn run(
&self,
args: CommandArgs,
registry: &CommandRegistry,
) -> Result<OutputStream, ShellError> {
calculate(
RunnableContext {
input: args.input,
registry: registry.clone(),
shell_manager: args.shell_manager,
host: args.host,
ctrl_c: args.ctrl_c,
current_errors: args.current_errors,
name: args.call_info.name_tag,
raw_input: args.raw_input,
},
minimum,
)
.await
}
fn examples(&self) -> Vec<Example> {
vec![Example {
description: "Get the minimum of a list of numbers",
example: "echo [-50 100 25] | math min",
result: Some(vec![UntaggedValue::int(-50).into()]),
}]
}
}
pub fn minimum(values: &[Value], _name: &Tag) -> Result<Value, ShellError> {
let min_func = reducer_for(Reduce::Minimum);
min_func(Value::nothing(), values.to_vec())
}
#[cfg(test)]
mod tests {
use super::SubCommand;
#[test]
fn examples_work_as_expected() {
use crate::examples::test as test_examples;
test_examples(SubCommand {})
}
}

View File

@ -0,0 +1,10 @@
pub mod average;
pub mod command;
pub mod max;
pub mod min;
pub mod utils;
pub use average::SubCommand as Average;
pub use command::Command as Math;
pub use max::SubCommand as Maximum;
pub use min::SubCommand as Minimum;

View File

@ -0,0 +1,52 @@
use crate::prelude::*;
use nu_errors::ShellError;
use nu_protocol::{Dictionary, ReturnSuccess, UntaggedValue, Value};
use indexmap::map::IndexMap;
pub type MathFunction = fn(values: &[Value], tag: &Tag) -> Result<Value, ShellError>;
pub async fn calculate(
RunnableContext {
mut input, name, ..
}: RunnableContext,
mf: MathFunction,
) -> Result<OutputStream, ShellError> {
let values: Vec<Value> = input.drain_vec().await;
if values.iter().all(|v| v.is_primitive()) {
match mf(&values, &name) {
Ok(result) => Ok(OutputStream::one(ReturnSuccess::value(result))),
Err(err) => Err(err),
}
} else {
let mut column_values = IndexMap::new();
for value in values {
if let UntaggedValue::Row(row_dict) = value.value {
for (key, value) in row_dict.entries.iter() {
column_values
.entry(key.clone())
.and_modify(|v: &mut Vec<Value>| v.push(value.clone()))
.or_insert(vec![value.clone()]);
}
}
}
let mut column_totals = IndexMap::new();
for (col_name, col_vals) in column_values {
match mf(&col_vals, &name) {
Ok(result) => {
column_totals.insert(col_name, result);
}
Err(err) => return Err(err),
}
}
Ok(OutputStream::one(ReturnSuccess::value(
UntaggedValue::Row(Dictionary {
entries: column_totals,
})
.into_untagged_value(),
)))
}
}

View File

@ -126,6 +126,7 @@ pub fn compute_values(
}
}
/// If left is {{ Operator }} right
pub fn compare_values(
operator: Operator,
left: &UntaggedValue,

View File

@ -29,8 +29,10 @@ pub fn test(cmd: impl WholeStreamCommand + 'static) {
.iter()
.zip(result.iter())
.all(|(e, a)| values_equal(e, a)),
"example produced unexpected result: {}",
"example command produced unexpected result.\ncommand: {}\nexpected: {:?}\nactual:{:?}",
example.example,
expected,
result,
);
}
}

View File

@ -8,6 +8,9 @@ use nu_source::{SpannedItem, Tag, Tagged, TaggedItem};
use nu_value_ext::{get_data_by_key, ValueExt};
use num_traits::Zero;
// Re-usable error messages
const ERR_EMPTY_DATA: &str = "Cannot perform aggregate math operation on empty data";
pub fn columns_sorted(
_group_by_name: Option<String>,
value: &Value,
@ -198,6 +201,9 @@ pub fn evaluate(
}
pub fn sum(data: Vec<Value>) -> Result<Value, ShellError> {
if data.is_empty() {
return Err(ShellError::unexpected(ERR_EMPTY_DATA));
}
let mut acc = Value::zero();
for value in data {
match value.value {
@ -214,6 +220,56 @@ pub fn sum(data: Vec<Value>) -> Result<Value, ShellError> {
Ok(acc)
}
pub fn max(data: Vec<Value>) -> Result<Value, ShellError> {
let mut biggest = data
.first()
.ok_or_else(|| ShellError::unexpected(ERR_EMPTY_DATA))?
.value
.clone();
for value in data.iter() {
if let Ok(greater_than) = compare_values(Operator::GreaterThan, &value.value, &biggest) {
if greater_than {
biggest = value.value.clone();
}
} else {
return Err(ShellError::unexpected(format!(
"Could not compare\nleft: {:?}\nright: {:?}",
biggest, value.value
)));
}
}
Ok(Value {
value: biggest,
tag: Tag::unknown(),
})
}
pub fn min(data: Vec<Value>) -> Result<Value, ShellError> {
let mut smallest = data
.first()
.ok_or_else(|| ShellError::unexpected(ERR_EMPTY_DATA))?
.value
.clone();
for value in data.iter() {
if let Ok(greater_than) = compare_values(Operator::LessThan, &value.value, &smallest) {
if greater_than {
smallest = value.value.clone();
}
} else {
return Err(ShellError::unexpected(format!(
"Could not compare\nleft: {:?}\nright: {:?}",
smallest, value.value
)));
}
}
Ok(Value {
value: smallest,
tag: Tag::unknown(),
})
}
fn formula(
acc_begin: Value,
calculator: Box<dyn Fn(Vec<Value>) -> Result<Value, ShellError> + Send + Sync + 'static>,
@ -233,11 +289,15 @@ pub fn reducer_for(
) -> Box<dyn Fn(Value, Vec<Value>) -> Result<Value, ShellError> + Send + Sync + 'static> {
match command {
Reduce::Sum | Reduce::Default => Box::new(formula(Value::zero(), Box::new(sum))),
Reduce::Minimum => Box::new(|_, values| min(values)),
Reduce::Maximum => Box::new(|_, values| max(values)),
}
}
pub enum Reduce {
Sum,
Minimum,
Maximum,
Default,
}
@ -250,6 +310,8 @@ pub fn reduce(
let reduce_with = match reducer {
Some(cmd) if cmd == "sum" => reducer_for(Reduce::Sum),
Some(cmd) if cmd == "min" => reducer_for(Reduce::Minimum),
Some(cmd) if cmd == "max" => reducer_for(Reduce::Maximum),
Some(_) | None => reducer_for(Reduce::Default),
};

View File

@ -7,11 +7,11 @@ fn can_average_numbers() {
r#"
open sgml_description.json
| get glossary.GlossDiv.GlossList.GlossEntry.Sections
| average
| math average
| echo $it
"#
));
println!("{:?}", actual.err);
assert_eq!(actual.out, "101.5")
}
@ -19,7 +19,7 @@ fn can_average_numbers() {
fn can_average_bytes() {
let actual = nu!(
cwd: "tests/fixtures/formats",
"ls | sort-by name | skip 1 | first 2 | get size | average | format \"{$it}\" | echo $it"
"ls | sort-by name | skip 1 | first 2 | get size | math average | format \"{$it}\" | echo $it"
);
assert_eq!(actual.out, "1.6 KB");