Fixed: Unicode Segmentation Beep Bop Beep! ⚡

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "flpc"
-version = "0.1.0"
+version = "0.2.0"
 edition = "2021"
 
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
@@ -9,6 +9,7 @@ name = "flpc"
 crate-type = ["cdylib"]
 
 [dependencies]
-pyo3 = "0.21.1"
+pyo3 = "0.22.0"
 regex = "1.10.5"
-lazy_static = "1.4.0"
+lazy_static = "1.5.0"
+unicode-segmentation = "1.11.0"

README.md

@@ -16,6 +16,7 @@
 </div>
 
 ## 🌟 Why flpc?
+[`CURRENTLY IN EXPERIMENTAL STAGE`]
 
 flpc is a powerful Python library that wraps the blazing-fast [Rust regex crate](https://crates.io/crates/regex), bringing enhanced speed to your regular expression operations. It's designed to be a drop-in replacement for Python's native `re` module, with some minor syntax differences.
 

examples/unicodes.py

@@ -0,0 +1,4 @@
+import flpc as re
+re.match = re.fmatch
+compiled_regex = re.compile('.*')
+print(re.match(compiled_regex,'hello \N{EARTH GLOBE AMERICAS}').span(0))

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "flpc"
-version = "0.2.1"
+version = "0.2.2"
 description = "A Lightning Fast ⚡ Rust-based regex crate wrapper for Python3 to get faster performance. 👾"
 maintainers = [{ name = "Adarsh Gourab Mahalik", email = "gourabmahalikadarsh@gmail.com" }]
 readme = "README.md"

src/lib.rs

@@ -5,6 +5,7 @@ use regex::{Captures, Regex, RegexBuilder};
 use std::collections::HashMap;
 use std::sync::Mutex;
 use lazy_static::lazy_static;
+use unicode_segmentation::UnicodeSegmentation;
 
 #[pyclass]
 struct Pattern {
@@ -55,19 +56,29 @@ impl Match {
     }
 
     fn start(&self, idx: usize) -> Option<usize> {
-        self.captures.get(idx).map(|m| m.start())
+        self.captures.get(idx).map(|m| {
+            self.captures.get(0).unwrap().as_str()[..m.start()].graphemes(true).count()
+        })
     }
 
     fn end(&self, idx: usize) -> Option<usize> {
-        self.captures.get(idx).map(|m| m.end())
+        self.captures.get(idx).map(|m| {
+            self.captures.get(0).unwrap().as_str()[..m.end()].graphemes(true).count()
+        })
     }
 
     fn span(&self, idx: usize) -> Option<(usize, usize)> {
-        self.captures.get(idx).map(|m| (m.start(), m.end()))
+        self.captures.get(idx).map(|m| {
+            let full_match = self.captures.get(0).unwrap().as_str();
+            let start = full_match[..m.start()].graphemes(true).count();
+            let end = full_match[..m.end()].graphemes(true).count();
+            (start, end)
+        })
     }
 }
 
 #[pyfunction]
+#[pyo3(signature = (pattern, flags=None))]
 fn compile(pattern: &str, flags: Option<u32>) -> PyResult<Pattern> {
     let flags = flags.unwrap_or(0);
     let mut cache = REGEX_CACHE.lock().unwrap();