From e25a795cf62aa3159d94269de65ba23499a42029 Mon Sep 17 00:00:00 2001
From: Horasal <1991933+horasal@users.noreply.github.com>
Date: Fri, 25 Aug 2023 09:21:17 +0900
Subject: [PATCH] Add encoding auto-detection for `decode` (#10030)
# Description
Allow `decode` command to guess the encoding of input if no encoding
name is given.
# User-Facing Changes
* `decode` now has an optional parameter instead of required one. User
can just run `decode` to let the command automatically detect encoding
and convert it to utf-8.
* Based on the detect result, user may have to give a encoding name
or get informed that the input is not supported by `decode`
# Tests + Formatting
* `cargo fmt --all -- --check` : OK
* `cargo clippy --workspace -- -D warnings -D clippy::unwrap_used -A
clippy::needless_collect -A clippy::result_large_err`: OK
* `cargo test --workspace` : OK
* `cargo run -- -c "use std testing; testing run-tests --path
crates/nu-std"`: OK
# After Submitting
[Command document](https://www.nushell.sh/commands/docs/decode.html) is
auto-generated and requires no action.
---------
Co-authored-by: Horasal
---
Cargo.lock | 12 ++++
crates/nu-command/Cargo.toml | 1 +
.../src/strings/encode_decode/decode.rs | 25 +++++++--
.../src/strings/encode_decode/encoding.rs | 56 +++++++++++++++++++
4 files changed, 89 insertions(+), 5 deletions(-)
diff --git a/Cargo.lock b/Cargo.lock
index 74e6c9220..c6378dded 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -549,6 +549,17 @@ version = "1.0.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "baf1de4339761588bc0619e3cbc0120ee582ebb74b53b4efbf79117bd2da40fd"
+[[package]]
+name = "chardetng"
+version = "0.1.17"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "14b8f0b65b7b08ae3c8187e8d77174de20cb6777864c6b832d8ad365999cf1ea"
+dependencies = [
+ "cfg-if",
+ "encoding_rs",
+ "memchr",
+]
+
[[package]]
name = "chrono"
version = "0.4.26"
@@ -2727,6 +2738,7 @@ dependencies = [
"byteorder",
"bytesize",
"calamine",
+ "chardetng",
"chrono",
"chrono-humanize",
"chrono-tz",
diff --git a/crates/nu-command/Cargo.toml b/crates/nu-command/Cargo.toml
index be05fb437..4583a9a6c 100644
--- a/crates/nu-command/Cargo.toml
+++ b/crates/nu-command/Cargo.toml
@@ -94,6 +94,7 @@ uuid = { version = "1.3", features = ["v4"] }
wax = { version = "0.5" }
which = { version = "4.4", optional = true }
bracoxide = "0.1.2"
+chardetng = "0.1.17"
[target.'cfg(windows)'.dependencies]
winreg = "0.50"
diff --git a/crates/nu-command/src/strings/encode_decode/decode.rs b/crates/nu-command/src/strings/encode_decode/decode.rs
index f99042637..1f1fa4d80 100644
--- a/crates/nu-command/src/strings/encode_decode/decode.rs
+++ b/crates/nu-command/src/strings/encode_decode/decode.rs
@@ -25,7 +25,7 @@ impl Command for Decode {
fn signature(&self) -> nu_protocol::Signature {
Signature::build("decode")
.input_output_types(vec![(Type::Binary, Type::String)])
- .required("encoding", SyntaxShape::String, "the text encoding to use")
+ .optional("encoding", SyntaxShape::String, "the text encoding to use")
.category(Category::Strings)
}
@@ -63,20 +63,35 @@ documentation link at https://docs.rs/encoding_rs/latest/encoding_rs/#statics"#
input: PipelineData,
) -> Result {
let head = call.head;
- let encoding: Spanned = call.req(engine_state, stack, 0)?;
+ let encoding: Option> = call.opt(engine_state, stack, 0)?;
match input {
PipelineData::ExternalStream { stdout: None, .. } => Ok(PipelineData::empty()),
PipelineData::ExternalStream {
stdout: Some(stream),
+ span: input_span,
..
} => {
let bytes: Vec = stream.into_bytes()?.item;
- super::encoding::decode(head, encoding, &bytes).map(|val| val.into_pipeline_data())
+ match encoding {
+ Some(encoding_name) => super::encoding::decode(head, encoding_name, &bytes),
+ None => super::encoding::detect_encoding_name(head, input_span, &bytes)
+ .map(|encoding| encoding.decode(&bytes).0.into_owned())
+ .map(|s| Value::String { val: s, span: head }),
+ }
+ .map(|val| val.into_pipeline_data())
}
PipelineData::Value(v, ..) => match v {
- Value::Binary { val: bytes, .. } => super::encoding::decode(head, encoding, &bytes)
- .map(|val| val.into_pipeline_data()),
+ Value::Binary {
+ val: bytes,
+ span: input_span,
+ } => match encoding {
+ Some(encoding_name) => super::encoding::decode(head, encoding_name, &bytes),
+ None => super::encoding::detect_encoding_name(head, input_span, &bytes)
+ .map(|encoding| encoding.decode(&bytes).0.into_owned())
+ .map(|s| Value::String { val: s, span: head }),
+ }
+ .map(|val| val.into_pipeline_data()),
Value::Error { error, .. } => Err(*error),
_ => Err(ShellError::OnlySupportsThisInputType {
exp_input_type: "binary".into(),
diff --git a/crates/nu-command/src/strings/encode_decode/encoding.rs b/crates/nu-command/src/strings/encode_decode/encoding.rs
index 6c11f1713..1e4f7f748 100644
--- a/crates/nu-command/src/strings/encode_decode/encoding.rs
+++ b/crates/nu-command/src/strings/encode_decode/encoding.rs
@@ -1,6 +1,27 @@
+use chardetng::EncodingDetector;
use encoding_rs::Encoding;
use nu_protocol::{ShellError, Span, Spanned, Value};
+pub fn detect_encoding_name(
+ head: Span,
+ input: Span,
+ bytes: &[u8],
+) -> Result<&'static Encoding, ShellError> {
+ let mut detector = EncodingDetector::new();
+ let _non_ascii = detector.feed(bytes, false);
+ //Guess(TLD=None(usually used in HTML), Allow_UTF8=True)
+ let (encoding, is_certain) = detector.guess_assess(None, true);
+ if !is_certain {
+ return Err(ShellError::UnsupportedInput(
+ "Input contains unknown encoding, try giving a encoding name".into(),
+ "value originates from here".into(),
+ head,
+ input,
+ ));
+ }
+ Ok(encoding)
+}
+
pub fn decode(
head: Span,
encoding_name: Spanned,
@@ -107,4 +128,39 @@ mod test {
assert_eq!(decoded, expected);
}
+
+ #[rstest]
+ #[case::big5(&[186, 251, 176, 242, 164, 106, 168, 229, 161, 93, 87, 105, 107, 105, 112, 101, 100, 105, 97, 161,
+ 94, 170, 204, 161, 65, 186, 244, 184, 244, 172, 176, 194, 166, 161, 70, 182, 176, 164, 209, 164, 85,
+ 170, 190, 161, 66, 165, 124, 174, 252, 168, 165, 161, 66, 178, 179, 164, 72, 167, 211, 161, 65, 174,
+ 209, 166, 202, 172, 236, 178, 106, 161, 67, 169, 108, 167, 64, 170, 204, 161, 65, 186, 251, 176,
+ 242, 180, 67, 197, 233, 176, 242, 170, 247, 183, 124, 164, 93, 161, 67], "Big5")]
+ // FIXME: chardetng fails on this
+ //#[case::shiftjis(&[130, 162, 130, 235, 130, 205, 130, 201, 130, 217, 130, 214, 130, 198, 129, 64, 130,
+ // 191, 130, 232, 130, 202, 130, 233, 130, 240], "SHIFT_JIS")]
+ #[case::eucjp(&[164, 164, 164, 237, 164, 207, 164, 203, 164, 219, 164, 216, 164, 200, 161, 161, 164, 193,
+ 164, 234, 164, 204, 164, 235, 164, 242],"EUC-JP")]
+ #[case::euckr(&[192, 167, 197, 176, 185, 233, 176, 250, 40, 45, 219, 221, 206, 161, 41, 32, 182, 199, 180,
+ 194, 32, 192, 167, 197, 176, 199, 199, 181, 240, 190, 198, 180, 194, 32, 180, 169, 177, 184, 179, 170,
+ 32, 192, 218, 192, 175, 183, 211, 176, 212, 32, 190, 181, 32, 188, 246, 32, 192, 214, 180, 194, 32, 180,
+ 217, 190, 240, 190, 238, 198, 199, 32, 192, 206, 197, 205, 179, 221, 32, 185, 233, 176, 250, 187, 231, 192,
+ 252, 192, 204, 180, 217],"EUC-KR")]
+ #[case::gb2312(&[206, 172, 187, 249, 202, 199, 210, 187, 214, 214, 212, 218, 205, 248, 194, 231, 201, 207, 191,
+ 170, 183, 197, 199, 210, 191, 201, 185, 169, 182, 224, 200, 203, 208, 173, 205, 172, 180, 180, 215, 247,
+ 181, 196, 179, 172, 206, 196, 177, 190, 207, 181, 205, 179, 163, 172, 211, 201, 195, 192, 185, 250, 200,
+ 203, 206, 214, 181, 194, 161, 164, 191, 178, 196, 254, 176, 178, 211, 218, 49, 57, 57, 53, 196, 234, 202,
+ 215, 207, 200, 191, 170, 183, 162], "GB2312")]
+ #[case::tis620(&[199, 212, 185, 226, 180, 199, 202, 236, 45, 49, 50, 53, 50, 224, 187, 231, 185, 195, 203, 209,
+ 202, 205, 209, 161, 162, 195, 208, 225, 186, 186, 203, 185, 214, 232, 167, 228, 186, 181, 236, 183, 213,
+ 232, 227, 170, 233, 161, 209, 186, 205, 209, 161, 201, 195, 197, 208, 181, 212, 185, 32, 193, 209, 161,
+ 182, 217, 161, 227, 170, 233, 227, 185, 205, 167, 164, 236, 187, 195, 208, 161, 205, 186, 195, 216, 232,
+ 185, 224, 161, 232, 210, 227, 185, 228, 193, 226, 164, 195, 171, 205, 191, 183, 236], "TIS-620")]
+ fn smoke_encoding_name(#[case] bytes: &[u8], #[case] expected: &str) {
+ let encoding_name =
+ detect_encoding_name(Span::test_data(), Span::test_data(), bytes).unwrap();
+ assert_eq!(
+ encoding_name,
+ Encoding::for_label(expected.as_bytes()).unwrap()
+ );
+ }
}