feat: add pattern-based rules and optimize dump processing

Cargo.toml

@@ -1,25 +1,34 @@
 [package]
 name = "pg_stage_rs"
-version = "0.1.3"
+version = "0.2.0"
 edition = "2021"
 description = "PostgreSQL dump anonymizer - streaming obfuscation for plain and custom formats"
 
+[features]
+default = ["mimalloc-allocator", "parallel"]
+mimalloc-allocator = ["dep:mimalloc"]
+# Opt-in: requires cmake/cc in build env; swaps flate2 backend to zlib-ng for ~1.5-2x zlib throughput.
+zlib-ng = ["flate2/zlib-ng"]
+parallel = ["dep:crossbeam-channel", "dep:num_cpus"]
+
 [dependencies]
 clap = { version = "4", features = ["derive"] }
 serde = { version = "1", features = ["derive"] }
 serde_json = "1"
 flate2 = "1"
 zstd = { version = "0.13", features = ["zstdmt"] }
 memchr = "2"
-crossbeam-channel = "0.5"
-num_cpus = "1.16"
 uuid = { version = "1", features = ["v4", "v5"] }
 regex = "1"
 rand = "0.8"
 chrono = { version = "0.4", default-features = false, features = ["clock", "std"] }
 hmac = "0.12"
 sha2 = "0.10"
 thiserror = "2"
+ahash = { version = "0.8", features = ["serde"] }
+mimalloc = { version = "0.1", default-features = false, optional = true }
+crossbeam-channel = { version = "0.5", optional = true }
+num_cpus = { version = "1.16", optional = true }
 
 [profile.release]
 opt-level = 3

src/conditions.rs

@@ -1,46 +1,42 @@
-use regex::Regex;
+use std::sync::Arc;
 
-use crate::types::Condition;
+use crate::types::{CompiledCondition, CondOp};
+use crate::FastMap;
 
-/// Check if conditions are met for a given row.
-/// Returns true if at least one condition matches.
-/// Returns true if conditions list is empty.
-pub fn check_conditions<S: AsRef<str>>(
-    conditions: &[Condition],
-    values: &[S],
-    column_indices: &std::collections::HashMap<String, usize>,
+/// Trait giving condition-evaluation access to a row's current (possibly already
+/// mutated) values by column index.
+pub trait RowRead {
+    fn len(&self) -> usize;
+    fn value_at(&self, idx: usize) -> &str;
+}
+
+/// Check if a compiled condition list matches the current row.
+/// Returns true if the list is empty, or if at least one condition matches.
+pub fn check_conditions(
+    conditions: &[CompiledCondition],
+    row: &dyn RowRead,
+    column_indices: &FastMap<Arc<str>, usize>,
 ) -> bool {
     if conditions.is_empty() {
         return true;
     }
-
     for condition in conditions {
-        let col_idx = match column_indices.get(&condition.column_name) {
-            Some(idx) => *idx,
+        let col_idx = match column_indices.get(condition.column_name.as_ref()) {
+            Some(&idx) => idx,
             None => continue,
         };
-        if col_idx >= values.len() {
+        if col_idx >= row.len() {
             continue;
         }
-        let col_value = values[col_idx].as_ref();
-
-        let matched = match condition.operation.as_str() {
-            "equal" => col_value == condition.value,
-            "not_equal" => col_value != condition.value,
-            "by_pattern" => {
-                if let Ok(re) = Regex::new(&condition.value) {
-                    re.is_match(col_value)
-                } else {
-                    false
-                }
-            }
-            _ => false,
+        let col_value = row.value_at(col_idx);
+        let matched = match &condition.op {
+            CondOp::Equal(v) => col_value == v.as_str(),
+            CondOp::NotEqual(v) => col_value != v.as_str(),
+            CondOp::ByPattern(re) => re.is_match(col_value),
         };
-
         if matched {
             return true;
         }
     }
-
     false
 }