Add cbor support

Cargo.toml

@@ -68,6 +68,7 @@ stdlib = [
   "dep:chacha20poly1305",
   "dep:charset",
   "dep:convert_case",
+  "dep:ciborium",
   "dep:cidr-utils",
   "dep:community-id",
   "dep:crypto_secretbox",
@@ -129,6 +130,7 @@ bytes = { version = "1", default-features = false, optional = true }
 charset = { version = "0.1", optional = true }
 chrono = { version = "0.4", default-features = false, features = ["clock", "serde", "wasmbind"], optional = true }
 chrono-tz = { version = "0.10", default-features = false, optional = true }
+ciborium = { version = "0.2.2", default-features = false, optional = true }
 cidr-utils = { version = "0.6", optional = true }
 csv = { version = "1", optional = true }
 clap = { version = "4", features = ["derive"], optional = true }

LICENSE-3rdparty.csv

@@ -44,6 +44,7 @@ chacha20poly1305,https://github.com/RustCrypto/AEADs/tree/master/chacha20poly130
 charset,https://github.com/hsivonen/charset,Apache-2.0 OR MIT,Henri Sivonen <hsivonen@hsivonen.fi>
 chrono,https://github.com/chronotope/chrono,MIT OR Apache-2.0,The chrono Authors
 chrono-tz,https://github.com/chronotope/chrono-tz,MIT OR Apache-2.0,The chrono-tz Authors
+ciborium,https://github.com/enarx/ciborium,Apache-2.0,Nathaniel McCallum <npmccallum@profian.com>
 cidr,https://github.com/stbuehler/rust-cidr,MIT,Stefan Bühler <stbuehler@web.de>
 cidr-utils,https://github.com/magiclen/cidr-utils,MIT,Magic Len <len@magiclen.org>
 cipher,https://github.com/RustCrypto/traits,MIT OR Apache-2.0,RustCrypto Developers
@@ -102,6 +103,7 @@ generic-array,https://github.com/fizyk20/generic-array,MIT,"Bartłomiej Kamińsk
 getrandom,https://github.com/rust-random/getrandom,MIT OR Apache-2.0,The Rand Project Developers
 gimli,https://github.com/gimli-rs/gimli,MIT OR Apache-2.0,The gimli Authors
 grok,https://github.com/daschl/grok,Apache-2.0,Michael Nitschinger <michael@nitschinger.at>
+half,https://github.com/starkat99/half-rs,MIT OR Apache-2.0,Kathryn Long <squeeself@gmail.com>
 hashbrown,https://github.com/rust-lang/hashbrown,MIT OR Apache-2.0,Amanieu d'Antras <amanieu@gmail.com>
 heck,https://github.com/withoutboats/heck,MIT OR Apache-2.0,The heck Authors
 heck,https://github.com/withoutboats/heck,MIT OR Apache-2.0,Without Boats <woboats@gmail.com>

changelog.d/1152.feature.md

@@ -0,0 +1 @@
+Add `parse_cbor` function

src/stdlib/mod.rs

@@ -137,6 +137,7 @@ cfg_if::cfg_if! {
         mod parse_aws_cloudwatch_log_subscription_message;
         mod parse_aws_vpc_flow_log;
         mod parse_cef;
+        mod parse_cbor;
         mod parse_common_log;
         mod parse_csv;
         mod parse_duration;
@@ -316,6 +317,7 @@ cfg_if::cfg_if! {
         pub use parse_aws_alb_log::ParseAwsAlbLog;
         pub use parse_aws_cloudwatch_log_subscription_message::ParseAwsCloudWatchLogSubscriptionMessage;
         pub use parse_aws_vpc_flow_log::ParseAwsVpcFlowLog;
+        pub use parse_cbor::ParseCbor;
         pub use parse_cef::ParseCef;
         pub use parse_common_log::ParseCommonLog;
         pub use parse_csv::ParseCsv;
@@ -500,6 +502,7 @@ pub fn all() -> Vec<Box<dyn Function>> {
         Box::new(ParseAwsAlbLog),
         Box::new(ParseAwsCloudWatchLogSubscriptionMessage),
         Box::new(ParseAwsVpcFlowLog),
+        Box::new(ParseCbor),
         Box::new(ParseCef),
         Box::new(ParseCommonLog),
         Box::new(ParseCsv),

src/stdlib/parse_cbor.rs

@@ -0,0 +1,128 @@
+use crate::compiler::prelude::*;
+use ciborium::de::from_reader;
+use zstd::zstd_safe::WriteBuf;
+
+fn parse_cbor(value: Value) -> Resolved {
+    let bytes = value.try_bytes()?;
+    let value = from_reader(bytes.as_slice()).map_err(|e| format!("unable to parse cbor: {e}"))?;
+    Ok(value)
+}
+
+#[derive(Clone, Copy, Debug)]
+pub struct ParseCbor;
+
+impl Function for ParseCbor {
+    fn identifier(&self) -> &'static str {
+        "parse_cbor"
+    }
+
+    fn summary(&self) -> &'static str {
+        "parse a string to a JSON type"
+    }
+
+    fn usage(&self) -> &'static str {
+        indoc! {"
+            Parses the provided `value` as CBOR.
+
+            Only JSON types are returned. If you need to convert a `string` into a `timestamp`,
+            consider the `parse_timestamp` function.
+        "}
+    }
+
+    fn parameters(&self) -> &'static [Parameter] {
+        &[Parameter {
+            keyword: "value",
+            kind: kind::BYTES,
+            required: true,
+        }]
+    }
+
+    fn examples(&self) -> &'static [Example] {
+        &[Example {
+            title: "object",
+            source: r#"parse_cbor!(decode_base64!("oWVmaWVsZGV2YWx1ZQ=="))"#,
+            result: Ok(r#"{ "field": "value" }"#),
+        }]
+    }
+
+    fn compile(
+        &self,
+        _state: &state::TypeState,
+        _ctx: &mut FunctionCompileContext,
+        arguments: ArgumentList,
+    ) -> Compiled {
+        let value = arguments.required("value");
+        Ok(ParseCborFn { value }.as_expr())
+    }
+}
+
+#[derive(Debug, Clone)]
+struct ParseCborFn {
+    value: Box<dyn Expression>,
+}
+
+impl FunctionExpression for ParseCborFn {
+    fn resolve(&self, ctx: &mut Context) -> Resolved {
+        let value = self.value.resolve(ctx)?;
+        parse_cbor(value)
+    }
+
+    fn type_def(&self, _: &state::TypeState) -> TypeDef {
+        type_def()
+    }
+}
+
+fn inner_kind() -> Kind {
+    Kind::null()
+        | Kind::bytes()
+        | Kind::integer()
+        | Kind::float()
+        | Kind::boolean()
+        | Kind::array(Collection::any())
+        | Kind::object(Collection::any())
+}
+
+fn type_def() -> TypeDef {
+    TypeDef::bytes()
+        .fallible()
+        .or_boolean()
+        .or_integer()
+        .or_float()
+        .add_null()
+        .or_array(Collection::from_unknown(inner_kind()))
+        .or_object(Collection::from_unknown(inner_kind()))
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use crate::value;
+    use nom::AsBytes;
+    use std::env;
+    use std::fs;
+    use std::path::PathBuf;
+
+    fn test_data_dir() -> PathBuf {
+        PathBuf::from(env::var_os("CARGO_MANIFEST_DIR").unwrap()).join("tests/data/cbor")
+    }
+
+    fn read_cbor_file(cbor_bin_message_path: &str) -> Vec<u8> {
+        fs::read(test_data_dir().join(cbor_bin_message_path)).unwrap()
+    }
+
+    test_function![
+        parse_cbor => ParseCbor;
+
+        parses {
+            args: func_args![ value: value!(read_cbor_file("simple.cbor").as_bytes()) ],
+            want: Ok(value!({ field: "value" })),
+            tdef: type_def(),
+        }
+
+        complex_cbor {
+            args: func_args![ value: value!(read_cbor_file("complex.cbor").as_bytes()) ],
+            want: Ok(value!({ object: {string: "value", number: 42, array: ["hello", "world"], boolean: false} })),
+            tdef: type_def(),
+        }
+    ];
+}

tests/data/cbor/complex.cbor

@@ -0,0 +1 @@
+�fobject�fstringevaluefnumber*earray�ehelloeworldgboolean�

tests/data/cbor/simple.cbor

@@ -0,0 +1 @@
+�efieldevalue