extern/nushell
1
0
Fork 1
mirror of https://github.com/nushell/nushell.git synced 2025-08-10 08:49:33 +02:00
Code Issues Packages Projects Releases Wiki Activity

This is part of on-going work with capabilities when working with

Browse Source 
tables and able to work with them for data processing & viewing
purposes. At the moment, certain ways to process said tables we
are able to view a histogram of a given column.

As usage matures, we may find certain core commands that could
be used ergonomically when working with tables on Nu.
This commit is contained in:
Andrés N. Robalino
2019-11-12 03:38:55 -05:00
parent 3163b0d362
commit 00b3c2036a
7 changed files with 147 additions and 109 deletions
Download Patch File Download Diff File Expand all files Collapse all files

97
src/commands/histogram.rs
View File

@ -1,10 +1,10 @@
use crate::commands::WholeStreamCommand;
use crate::commands::evaluate_by::evaluate;
use crate::commands::group_by::group;
use crate::commands::map_max_by::map_max;
use crate::commands::reduce_by::reduce;
use crate::commands::t_sort_by::columns_sorted;
use crate::commands::t_sort_by::t_sort;
use crate::commands::evaluate_by::evaluate;
use crate::commands::reduce_by::reduce;
use crate::commands::map_max_by::map_max;
use crate::commands::WholeStreamCommand;
use crate::data::TaggedDictBuilder;
use crate::errors::ShellError;
use crate::prelude::*;
@ -15,6 +15,7 @@ pub struct Histogram;
#[derive(Deserialize)]
pub struct HistogramArgs {
column_name: Tagged<String>,
rest: Vec<Tagged<String>>,
}
impl WholeStreamCommand for Histogram {
@ -23,11 +24,16 @@ impl WholeStreamCommand for Histogram {
}
fn signature(&self) -> Signature {
Signature::build("histogram").required(
"column_name",
SyntaxShape::String,
"the name of the column to graph by",
)
Signature::build("histogram")
.required(
"column_name",
SyntaxShape::String,
"the name of the column to graph by",
)
.rest(
SyntaxShape::Member,
"column name to give the histogram's frequency column",
)
}
fn usage(&self) -> &str {
@ -44,7 +50,7 @@ impl WholeStreamCommand for Histogram {
}
pub fn histogram(
HistogramArgs { column_name }: HistogramArgs,
HistogramArgs { column_name, rest }: HistogramArgs,
RunnableContext { input, name, .. }: RunnableContext,
) -> Result<OutputStream, ShellError> {
let stream = async_stream! {
@ -68,13 +74,24 @@ pub fn histogram(
let mut idx = 0;
let column_names_supplied: Vec<_> = rest.iter().map(|f| f.item.clone()).collect();
let frequency_column_name = if column_names_supplied.is_empty() {
"frecuency".to_string()
} else {
column_names_supplied[0].clone()
};
let column = (*column_name).clone();
if let Tagged { item: Value::Table(start), .. } = datasets.get(0).unwrap() {
for percentage in start.into_iter() {
let mut fact = TaggedDictBuilder::new(&name);
fact.insert_tagged("committer", group_labels.get(idx).unwrap().clone());
fact.insert_tagged(&column, group_labels.get(idx).unwrap().clone());
if let Tagged { item: Value::Primitive(Primitive::Int(ref num)), .. } = percentage.clone() {
fact.insert("activity", std::iter::repeat("*").take(num.to_i32().unwrap() as usize).collect::<String>());
fact.insert(&frequency_column_name, std::iter::repeat("*").take(num.to_i32().unwrap() as usize).collect::<String>());
}
idx = idx + 1;
@ -104,38 +121,36 @@ fn percentages(
} => {
let datasets: Vec<_> = datasets
.into_iter()
.map(|subsets| {
match subsets {
Tagged {
item: Value::Table(data),
..
} => {
let data = data
.into_iter()
.map(|d| match d {
Tagged {
item: Value::Primitive(Primitive::Int(n)),
..
} => {
let max = match max {
Tagged {
item: Value::Primitive(Primitive::Int(ref maxima)),
..
} => maxima.to_i32().unwrap(),
_ => 0,
};
.map(|subsets| match subsets {
Tagged {
item: Value::Table(data),
..
} => {
let data = data
.into_iter()
.map(|d| match d {
Tagged {
item: Value::Primitive(Primitive::Int(n)),
..
} => {
let max = match max {
Tagged {
item: Value::Primitive(Primitive::Int(ref maxima)),
..
} => maxima.to_i32().unwrap(),
_ => 0,
};
let n = { n.to_i32().unwrap() * 100 / max };
let n = { n.to_i32().unwrap() * 100 / max };
Value::number(n).tagged(&tag)
}
_ => Value::number(0).tagged(&tag),
})
.collect::<Vec<_>>();
Value::Table(data).tagged(&tag)
}
_ => Value::Table(vec![]).tagged(&tag),
Value::number(n).tagged(&tag)
}
_ => Value::number(0).tagged(&tag),
})
.collect::<Vec<_>>();
Value::Table(data).tagged(&tag)
}
_ => Value::Table(vec![]).tagged(&tag),
})
.collect();

40
src/commands/map_max_by.rs
View File

@ -81,29 +81,27 @@ pub fn map_max(
} => {
let datasets: Vec<_> = datasets
.into_iter()
.map(|subsets| {
match subsets {
Tagged {
item: Value::Table(data),
..
} => {
let data = data.into_iter().fold(0, |acc, value| match value {
Tagged {
item: Value::Primitive(Primitive::Int(n)),
..
} => {
if n.to_i32().unwrap() > acc {
n.to_i32().unwrap()
} else {
acc
}
.map(|subsets| match subsets {
Tagged {
item: Value::Table(data),
..
} => {
let data = data.into_iter().fold(0, |acc, value| match value {
Tagged {
item: Value::Primitive(Primitive::Int(n)),
..
} => {
if n.to_i32().unwrap() > acc {
n.to_i32().unwrap()
} else {
acc
}
_ => acc,
});
Value::number(data).tagged(&tag)
}
_ => Value::number(0).tagged(&tag),
}
_ => acc,
});
Value::number(data).tagged(&tag)
}
_ => Value::number(0).tagged(&tag),
})
.collect();

38
src/commands/t_sort_by.rs
View File

@ -57,25 +57,25 @@ fn t_sort_by(
RunnableContext { input, name, .. }: RunnableContext,
) -> Result<OutputStream, ShellError> {
Ok(OutputStream::new(async_stream! {
let values: Vec<Tagged<Value>> = input.values.collect().await;
let values: Vec<Tagged<Value>> = input.values.collect().await;
let column_grouped_by_name = if let Some(grouped_by) = group_by {
Some(grouped_by.item().clone())
} else {
None
};
let column_grouped_by_name = if let Some(grouped_by) = group_by {
Some(grouped_by.item().clone())
} else {
None
};
if show_columns {
for label in columns_sorted(column_grouped_by_name, &values[0], &name).iter() {
yield ReturnSuccess::value(label.clone());
}
} else {
match t_sort(column_grouped_by_name, None, &values[0], name) {
Ok(sorted) => yield ReturnSuccess::value(sorted),
Err(err) => yield Err(err)
}
if show_columns {
for label in columns_sorted(column_grouped_by_name, &values[0], &name).iter() {
yield ReturnSuccess::value(label.clone());
}
}))
} else {
match t_sort(column_grouped_by_name, None, &values[0], name) {
Ok(sorted) => yield ReturnSuccess::value(sorted),
Err(err) => yield Err(err)
}
}
}))
}
pub fn columns_sorted(
@ -125,7 +125,7 @@ pub fn columns_sorted(
keys.into_iter().map(|k| k.tagged(&origin_tag)).collect()
}
_ => vec![Value::string("default").tagged(&origin_tag)]
_ => vec![Value::string("default").tagged(&origin_tag)],
}
}
@ -238,7 +238,7 @@ mod tests {
use crate::data::meta::*;
use crate::Value;
use indexmap::IndexMap;
fn string(input: impl Into<String>) -> Tagged<Value> {
Value::string(input.into()).tagged_unknown()
}
@ -305,7 +305,7 @@ mod tests {
]
)
}
#[test]
fn sorts_the_tables() {
let group_by = String::from("date");