Fix string id edge cases (#1460)

* fix number detection edge cases for string to id conversion and add some tests

* fix the test that was testing broken behaviour

* custom parsing function instead of regex for speed

python/tests/test_graphdb.py

@@ -1970,8 +1970,8 @@ def test_leading_zeroes_ids():
     g = Graph()
     g.add_node(0, 0)
     g.add_node(1, "0")
-    assert g.node(0).history() == [0]
-    assert g.nodes.name.collect() == ["0", "0"]
+    assert g.node(0).history() == [0, 1]
+    assert g.nodes.name.collect() == ["0"]
 
     # g = Graph()
     # g.add_node(0, 1)

raphtory/src/core/entities/nodes/input_node.rs

@@ -5,6 +5,47 @@
 //! `u64`, `&str`, and `String`.
 
 use crate::core::utils::hashing;
+use regex::Regex;
+const MAX_U64_BYTES: [u8; 20] = [
+    49, 56, 52, 52, 54, 55, 52, 52, 48, 55, 51, 55, 48, 57, 53, 53, 49, 54, 49, 53,
+];
+
+fn parse_u64_strict(input: &str) -> Option<u64> {
+    if input.len() > 20 {
+        /// a u64 string has at most 20 bytes
+        return None;
+    }
+    let byte_0 = b'0';
+    let byte_1 = b'1';
+    let byte_9 = b'9';
+    let mut input_iter = input.bytes();
+    let first = input_iter.next()?;
+    if first == byte_0 {
+        return input_iter.next().is_none().then_some(0);
+    }
+    if input.len() == 20 && (byte_1..=MAX_U64_BYTES[0]).contains(&first) {
+        let mut result = (first - byte_0) as u64;
+        for (next_byte, max_byte) in input_iter.zip(MAX_U64_BYTES[1..].iter().copied()) {
+            if !(byte_0..=max_byte).contains(&next_byte) {
+                return None;
+            }
+            result = result * 10 + (next_byte - byte_0) as u64;
+        }
+        return Some(result);
+    }
+    if (byte_1..=byte_9).contains(&first) {
+        let mut result = (first - byte_0) as u64;
+        for next_byte in input_iter {
+            if !(byte_0..=byte_9).contains(&next_byte) {
+                return None;
+            }
+            result = result * 10 + (next_byte - byte_0) as u64;
+        }
+        return Some(result);
+    }
+
+    None
+}
 
 pub trait InputNode: Clone {
     fn id(&self) -> u64;
@@ -23,11 +64,7 @@ impl InputNode for u64 {
 
 impl<'a> InputNode for &'a str {
     fn id(&self) -> u64 {
-        if &self.chars().next().unwrap_or('0') != &'0' {
-            self.parse().unwrap_or(hashing::calculate_hash(self))
-        } else {
-            hashing::calculate_hash(self)
-        }
+        parse_u64_strict(self).unwrap_or_else(|| hashing::calculate_hash(self))
     }
 
     fn id_str(&self) -> Option<&str> {
@@ -45,3 +82,17 @@ impl InputNode for String {
         Some(self)
     }
 }
+
+#[cfg(test)]
+mod test {
+    use crate::core::entities::nodes::input_node::InputNode;
+    use regex::Regex;
+
+    #[test]
+    fn test_weird_num_edge_cases() {
+        assert_ne!("+3".id(), "3".id());
+        assert_eq!(3.id(), "3".id());
+        assert_ne!("00".id(), "0".id());
+        assert_eq!("0".id(), 0.id());
+    }
+}