nushell/crates/nu-std/std/xml/mod.nu
Douglas 2a3805c164
Virtual std module subdirectories (#14040)
# Description

Uses "normal" module `std/<submodule>/mod.nu` instead of renaming the
files (as requested in #13842).

# User-Facing Changes

No user-facing changes other than in `view files` results. Imports
remain the same after this PR.

# Tests + Formatting

- 🟢 `toolkit fmt`
- 🟢 `toolkit clippy`
- 🟢 `toolkit test`
- 🟢 `toolkit test stdlib`

Also manually confirmed that it does not interfere with nupm, since we
did have a conflict at one point (and it's not possible to test here).

# Performance Tests

## Linux

### Nushell Startup - No config

```nu
bench --pretty -n 200  { <path_to>/nu -c "exit" }
```

| Release | Startup Time |
| --- | --- |
| 0.98.0 | 22ms 730µs 768ns +/- 1ms 515µs 942ns
| This commit | 9ms 312µs 68ns +/- 709µs 378ns
| Yesterday's nightly | 9ms 230µs 953ns +/- 9ms 67µs 689ns

### Nushell Startup - Load full standard library

Measures relative impact of a full `use std *`, which isn't recommended,
but worth tracking.

```nu
bench --pretty -n 200  { <path_to>/nu -c "use std *; exit" }
```

| Release | Startup Time |
| --- | --- |
| 0.98.0 | 23ms 10µs 636ns +/- 1ms 277µs 854ns
| This commit | 26ms 922µs 769ns +/- 562µs 538ns
| Yesterday's nightly | 28ms 133µs 95ns +/- 761µs 943ns
| `deprecated_dirs` removal PR * | 23ms 610µs 333ns +/- 369µs 436ns

\* Current increase is partially due to double-loading `dirs` with
removal warning in older version.

# After Submitting

Still TODO - Update standard library doc
2024-10-10 06:56:37 -05:00

219 lines
7.9 KiB
Plaintext

# Utility functions to read, change and create XML data in format supported
# by `to xml` and `from xml` commands
# Get all xml entries matching simple xpath-inspired query
export def xaccess [
path: list # List of steps. Each step can be a
# 1. String with tag name. Finds all children with specified name. Equivalent to `child::A` in xpath
# 2. `*` string. Get all children without any filter. Equivalent to `descendant` in xpath
# 3. Int. Select n-th among nodes selected by previous path. Equivalent to `(...)[1]` in xpath, but is indexed from 0.
# 4. Closure. Predicate accepting entry. Selects all entries among nodes selected by previous path for which predicate returns true.
] {
let input = $in
if ($path | is-empty) {
let path_span = (metadata $path).span
error make {
msg: 'Empty path provided'
label: {
text: 'Use a non-empty list of path steps'
span: $path_span
}
}
}
# In xpath first element in path is applied to root element
# this way it is possible to apply first step to root element
# of nu xml without unrolling one step of loop
mut values: any = ()
$values = {content: [ { content: $input } ] }
for $step in ($path) {
match ($step | describe) {
'string' => {
if $step == '*' {
$values = ($values.content | flatten)
} else {
$values = ($values.content | flatten | where tag == $step)
}
},
'int' => {
$values = [ ($values | get $step) ]
},
'closure' => {
$values = ($values | where {|x| do $step $x})
},
$type => {
let step_span = (metadata $step).span
error make {
msg: $'Incorrect path step type ($type)'
label: {
text: 'Use a string or int as a step'
span: $step_span
}
}
}
}
if ($values | is-empty) {
return []
}
}
$values
}
def xupdate-string-step [ step: string rest: list updater: closure ] {
let input = $in
# Get a list of elements to be updated and their indices
let to_update = ($input.content | enumerate | filter {|it|
let item = $it.item
$step == '*' or $item.tag == $step
})
if ($to_update | is-empty) {
return $input
}
let new_values = ($to_update.item | xupdate-internal $rest $updater)
mut reenumerated_new_values: any = ($to_update.index | zip $new_values | each {|x| {index: $x.0 item: $x.1}})
mut new_content = []
for it in ($input.content | enumerate) {
let item = $it.item
let idx = $it.index
let next = (if (not ($reenumerated_new_values | is-empty)) and $idx == $reenumerated_new_values.0.index {
let tmp = $reenumerated_new_values.0
$reenumerated_new_values = ($reenumerated_new_values | skip 1)
$tmp.item
} else {
$item
})
$new_content = ($new_content | append $next)
}
{tag: $input.tag attributes: $input.attributes content: $new_content}
}
def xupdate-int-step [ step: int rest: list updater: closure ] {
$in | enumerate | each {|it|
let item = $it.item
let idx = $it.index
if $idx == $step {
[ $item ] | xupdate-internal $rest $updater | get 0
} else {
$item
}
}
}
def xupdate-closure-step [ step: closure rest: list updater: closure ] {
$in | each {|it|
if (do $step $it) {
[ $it ] | xupdate-internal $rest $updater | get 0
} else {
$it
}
}
}
def xupdate-internal [ path: list updater: closure ] {
let input = $in
if ($path | is-empty) {
$input | each $updater
} else {
let step = $path.0
let rest = ($path | skip 1)
match ($step | describe) {
'string' => {
$input | each {|x| $x | xupdate-string-step $step $rest $updater}
},
'int' => {
$input | xupdate-int-step $step $rest $updater
},
'closure' => {
$input | xupdate-closure-step $step $rest $updater
},
$type => {
let step_span = (metadata $step).span
error make {
msg: $'Incorrect path step type ($type)'
label: {
text: 'Use a string or int as a step'
span: $step_span
}
}
}
}
}
}
# Update xml data entries matching simple xpath-inspired query
export def xupdate [
path: list # List of steps. Each step can be a
# 1. String with tag name. Finds all children with specified name. Equivalent to `child::A` in xpath
# 2. `*` string. Get all children without any filter. Equivalent to `descendant` in xpath
# 3. Int. Select n-th among nodes selected by previous path. Equivalent to `(...)[1]` in xpath, but is indexed from 0.
# 4. Closure. Predicate accepting entry. Selects all entries among nodes selected by previous path for which predicate returns true.
updater: closure # A closure used to transform entries matching path.
] {
{tag:? attributes:? content: [$in]} | xupdate-internal $path $updater | get content.0
}
# Get type of an xml entry
#
# Possible types are 'tag', 'text', 'pi' and 'comment'
export def xtype [] {
let input = $in
if (($input | describe) == 'string' or
($input.tag? == null and $input.attributes? == null and ($input.content? | describe) == 'string')) {
'text'
} else if $input.tag? == '!' {
'comment'
} else if $input.tag? != null and ($input.tag? | str starts-with '?') {
'pi'
} else if $input.tag? != null {
'tag'
} else {
error make {msg: 'Not an xml emtry. Check valid types of xml entries via `help to xml`'}
}
}
# Insert new entry to elements matching simple xpath-inspired query
export def xinsert [
path: list # List of steps. Each step can be a
# 1. String with tag name. Finds all children with specified name. Equivalent to `child::A` in xpath
# 2. `*` string. Get all children without any filter. Equivalent to `descendant` in xpath
# 3. Int. Select n-th among nodes selected by previous path. Equivalent to `(...)[1]` in xpath, but is indexed from 0.
# 4. Closure. Predicate accepting entry. Selects all entries among nodes selected by previous path for which predicate returns true.
new_entry: record # A new entry to insert into `content` field of record at specified position
position?: int # Position to insert `new_entry` into. If specified inserts entry at given position (or end if
# position is greater than number of elements) in content of all entries of input matched by
# path. If not specified inserts at the end.
] {
$in | xupdate $path {|entry|
match ($entry | xtype) {
'tag' => {
let new_content = if $position == null {
$entry.content | append $new_entry
} else {
let position = if $position > ($entry.content | length) {
$entry.content | length
} else {
$position
}
$entry.content | insert $position $new_entry
}
{tag: $entry.tag attributes: $entry.attributes content: $new_content}
},
_ => (error make {msg: 'Can insert entry only into content of a tag node'})
}
}
}