Refactor ansi stripping into nu-utils functions (#6966)

Allows use of slightly optimized variants that check if they have to use
the heavier vte parser. Tries to avoid unnnecessary allocations. Initial
performance characteristics proven out in #4378.

Also reduces boilerplate with right-ward drift.

crates/nu-utils/src/deansi.rs

@@ -0,0 +1,91 @@
+use std::borrow::Cow;
+
+/// Removes ANSI escape codes and some ASCII control characters
+///
+/// Optimized for strings that rarely contain ANSI control chars.
+/// Uses fast search to avoid reallocations.
+///
+/// Keeps `\n` removes `\r`, `\t` etc.
+///
+/// If parsing fails silently returns the input string
+pub fn strip_ansi_unlikely(string: &str) -> Cow<str> {
+    // Check if any ascii control character except LF(0x0A = 10) is present,
+    // which will be stripped. Includes the primary start of ANSI sequences ESC
+    // (0x1B = decimal 27)
+    if string.bytes().any(|x| matches!(x, 0..=9 | 11..=31)) {
+        if let Ok(stripped) = strip_ansi_escapes::strip(string) {
+            if let Ok(new_string) = String::from_utf8(stripped) {
+                return Cow::Owned(new_string);
+            }
+        }
+    }
+    // Else case includes failures to parse!
+    Cow::Borrowed(string)
+}
+
+/// Removes ANSI escape codes and some ASCII control characters
+///
+/// Optimized for strings that likely contain ANSI control chars.
+///
+/// Keeps `\n` removes `\r`, `\t` etc.
+///
+/// If parsing fails silently returns the input string
+pub fn strip_ansi_likely(string: &str) -> Cow<str> {
+    // Check if any ascii control character except LF(0x0A = 10) is present,
+    // which will be stripped. Includes the primary start of ANSI sequences ESC
+    // (0x1B = decimal 27)
+    if let Ok(stripped) = strip_ansi_escapes::strip(string) {
+        if let Ok(new_string) = String::from_utf8(stripped) {
+            return Cow::Owned(new_string);
+        }
+    }
+    // Else case includes failures to parse!
+    Cow::Borrowed(string)
+}
+
+/// Removes ANSI escape codes and some ASCII control characters
+///
+/// Optimized for strings that rarely contain ANSI control chars.
+/// Uses fast search to avoid reallocations.
+///
+/// Keeps `\n` removes `\r`, `\t` etc.
+///
+/// If parsing fails silently returns the input string
+pub fn strip_ansi_string_unlikely(string: String) -> String {
+    // Check if any ascii control character except LF(0x0A = 10) is present,
+    // which will be stripped. Includes the primary start of ANSI sequences ESC
+    // (0x1B = decimal 27)
+    if string
+        .as_str()
+        .bytes()
+        .any(|x| matches!(x, 0..=9 | 11..=31))
+    {
+        if let Ok(stripped) = strip_ansi_escapes::strip(&string) {
+            if let Ok(new_string) = String::from_utf8(stripped) {
+                return new_string;
+            }
+        }
+    }
+    // Else case includes failures to parse!
+    string
+}
+
+/// Removes ANSI escape codes and some ASCII control characters
+///
+/// Optimized for strings that likely contain ANSI control chars.
+///
+/// Keeps `\n` removes `\r`, `\t` etc.
+///
+/// If parsing fails silently returns the input string
+pub fn strip_ansi_string_likely(string: String) -> String {
+    // Check if any ascii control character except LF(0x0A = 10) is present,
+    // which will be stripped. Includes the primary start of ANSI sequences ESC
+    // (0x1B = decimal 27)
+    if let Ok(stripped) = strip_ansi_escapes::strip(&string) {
+        if let Ok(new_string) = String::from_utf8(stripped) {
+            return new_string;
+        }
+    }
+    // Else case includes failures to parse!
+    string
+}

crates/nu-utils/src/lib.rs

@@ -1,3 +1,4 @@
+mod deansi;
 pub mod locale;
 pub mod utils;
 
@@ -6,3 +7,7 @@ pub use utils::{
     enable_vt_processing, get_default_config, get_default_env, get_ls_colors,
     stderr_write_all_and_flush, stdout_write_all_and_flush,
 };
+
+pub use deansi::{
+    strip_ansi_likely, strip_ansi_string_likely, strip_ansi_string_unlikely, strip_ansi_unlikely,
+};