Noir ZK Regex

Cargo.toml

@@ -1,6 +1,6 @@
 [workspace]
 members = ["packages/*"]
-exclude = ["packages/circom", "test"]
+exclude = ["packages/circom", "test", "packages/noir"]
 
 # [patch."https://github.com/stalwartlabs/mail-builder"]
 # mail-builder = { version = "0.2.5", git = "https://github.com/stalwartlabs//mail-builder", tag = "0.2.5" }

README.md

@@ -1,6 +1,6 @@
 # zk-regex
 
-A library to compile regex verification in circom. Explained on [our blog post](https://prove.email/blog/zkregex).  You can use regex to specify how to parse an email in a ZK Email proof when defining a new patterm on [the ZK Email SDK registry](https://sdk.prove.email/). Noir coming soon.
+A library to compile regex verification in circom. Explained on [our blog post](https://prove.email/blog/zkregex).  You can use regex to specify how to parse an email in a ZK Email proof when defining a new patterm on [the ZK Email SDK registry](https://sdk.prove.email/). Noir support is also available. 
 
 <!-- We've forked [min-dfa into a UI here](https://mindfa.onrender.com/min_dfa) to create a UI that converts existing regexes with [] support, as well as escapes \_, and the character classes a-z, A-Z, and 0-9. It also shows the DFA states very clearly so you can choose accept states easily. This should make converting regexes into DFA form way cleaner. -->
 

packages/circom/tests/simple_regex.test.js

@@ -68,7 +68,7 @@ describe("Simple Regex", () => {
       for (let idx = 0; idx < 64; ++idx) {
         if (revealedIdx[substr_idx].includes(idx)) {
           expect(BigInt(paddedStr[idx])).toEqual(
-            witness[2 + 64 * substr_idx + idx]
+            witness[2 + 64 * substr_idx +   idx]
           );
         } else {
           expect(0n).toEqual(witness[2 + 64 * substr_idx + idx]);

packages/compiler/Cargo.toml

@@ -36,3 +36,5 @@ regex = "=1.10.6"
 getrandom = { version = "0.2", features = ["js"] }
 wasm-bindgen = "0.2"
 serde-wasm-bindgen = "0.6.5"
+comptime = { git = "https://github.com/jp4g/sparse_array", branch = "feat/comptime-codegen" }
+# comptime = { path = "../../../../aztec/sparse_array/comptime" }

packages/compiler/src/bin/compiler.rs

@@ -60,28 +60,36 @@ enum Commands {
     Decomposed {
         #[arg(short, long)]
         decomposed_regex_path: String,
-        #[arg(short, long)]
+        #[arg(long)]
         halo2_dir_path: Option<String>,
         #[arg(short, long)]
         circom_file_path: Option<String>,
         #[arg(short, long)]
         template_name: Option<String>,
+        #[arg(long)]
+        noir_file_path: Option<String>,
         #[arg(short, long)]
         gen_substrs: Option<bool>,
+        #[arg(short, long)]
+        sparse_array: Option<bool>,
     },
     Raw {
         #[arg(short, long)]
         raw_regex: String,
         #[arg(short, long)]
         substrs_json_path: Option<String>,
-        #[arg(short, long)]
+        #[arg(long)]
         halo2_dir_path: Option<String>,
         #[arg(short, long)]
         circom_file_path: Option<String>,
         #[arg(short, long)]
         template_name: Option<String>,
+        #[arg(long)]
+        noir_file_path: Option<String>,
         #[arg(short, long)]
         gen_substrs: Option<bool>,
+        #[arg(short, long)]
+        sparse_array: Option<bool>,
     },
 }
 
@@ -99,15 +107,19 @@ fn process_decomposed(cli: Cli) {
         halo2_dir_path,
         circom_file_path,
         template_name,
+        noir_file_path,
         gen_substrs,
+        sparse_array,
     } = cli.command
     {
         if let Err(e) = gen_from_decomposed(
             &decomposed_regex_path,
             halo2_dir_path.as_deref(),
             circom_file_path.as_deref(),
             template_name.as_deref(),
+            noir_file_path.as_deref(),
             gen_substrs,
+            sparse_array
         ) {
             eprintln!("Error: {}", e);
             std::process::exit(1);
@@ -122,7 +134,9 @@ fn process_raw(cli: Cli) {
         halo2_dir_path,
         circom_file_path,
         template_name,
+        noir_file_path,
         gen_substrs,
+        sparse_array,
     } = cli.command
     {
         if let Err(e) = gen_from_raw(
@@ -131,7 +145,9 @@ fn process_raw(cli: Cli) {
             halo2_dir_path.as_deref(),
             circom_file_path.as_deref(),
             template_name.as_deref(),
+            noir_file_path.as_deref(),
             gen_substrs,
+            sparse_array
         ) {
             eprintln!("Error: {}", e);
             std::process::exit(1);

packages/compiler/src/lib.rs

@@ -1,14 +1,17 @@
 mod circom;
 mod errors;
 mod halo2;
+mod noir;
 mod regex;
 mod structs;
 mod wasm;
 
+
 use circom::gen_circom_template;
 use errors::CompilerError;
 use halo2::gen_halo2_tables;
 use itertools::Itertools;
+use noir::gen_noir_fn;
 use regex::{create_regex_and_dfa_from_str_and_defs, get_regex_and_dfa};
 use std::{fs::File, path::PathBuf};
 use structs::{DecomposedRegexConfig, RegexAndDFA, SubstringDefinitionsJson};
@@ -46,6 +49,7 @@ fn load_substring_definitions_json(
 /// * `circom_template_name` - An optional name for the Circom template.
 /// * `num_public_parts` - The number of public parts in the regex.
 /// * `gen_substrs` - A boolean indicating whether to generate substrings.
+/// * `sparse_array` - A boolean indicating whether to use a sparse array for the DFA. (TEST UTILITY)
 ///
 /// # Returns
 ///
@@ -55,8 +59,10 @@ fn generate_outputs(
     halo2_dir_path: Option<&str>,
     circom_file_path: Option<&str>,
     circom_template_name: Option<&str>,
+    noir_file_path: Option<&str>,
     num_public_parts: usize,
     gen_substrs: bool,
+    sparse_array: Option<bool>
 ) -> Result<(), CompilerError> {
     if let Some(halo2_dir_path) = halo2_dir_path {
         let halo2_dir_path = PathBuf::from(halo2_dir_path);
@@ -86,6 +92,10 @@ fn generate_outputs(
         )?;
     }
 
+    if let Some(noir_file_path) = noir_file_path {
+        gen_noir_fn(regex_and_dfa, &PathBuf::from(noir_file_path), gen_substrs, sparse_array)?;
+    }
+
     Ok(())
 }
 
@@ -107,7 +117,9 @@ pub fn gen_from_decomposed(
     halo2_dir_path: Option<&str>,
     circom_file_path: Option<&str>,
     circom_template_name: Option<&str>,
+    noir_file_path: Option<&str>,
     gen_substrs: Option<bool>,
+    sparse_array: Option<bool>
 ) -> Result<(), CompilerError> {
     let mut decomposed_regex_config: DecomposedRegexConfig =
         serde_json::from_reader(File::open(decomposed_regex_path)?)?;
@@ -126,8 +138,10 @@ pub fn gen_from_decomposed(
         halo2_dir_path,
         circom_file_path,
         circom_template_name,
+        noir_file_path,
         num_public_parts,
         gen_substrs,
+        sparse_array
     )?;
 
     Ok(())
@@ -153,22 +167,26 @@ pub fn gen_from_raw(
     halo2_dir_path: Option<&str>,
     circom_file_path: Option<&str>,
     template_name: Option<&str>,
+    noir_file_path: Option<&str>,
     gen_substrs: Option<bool>,
+    sparse_array: Option<bool>
 ) -> Result<(), CompilerError> {
     let substrs_defs_json = load_substring_definitions_json(substrs_json_path)?;
     let num_public_parts = substrs_defs_json.transitions.len();
 
     let regex_and_dfa = create_regex_and_dfa_from_str_and_defs(raw_regex, substrs_defs_json)?;
 
-    let gen_substrs = gen_substrs.unwrap_or(true);
+    let gen_substrs = gen_substrs.unwrap_or(false);
 
     generate_outputs(
         &regex_and_dfa,
         halo2_dir_path,
         circom_file_path,
         template_name,
+        noir_file_path,
         num_public_parts,
         gen_substrs,
+        sparse_array
     )?;
 
     Ok(())