feat: first cut on M0 - base structure of project

crates/bds-core/src/util/checksum.rs

@@ -0,0 +1,48 @@
+use sha2::{Digest, Sha256};
+
+/// Compute a hex-encoded SHA-256 hash of the given content.
+pub fn content_hash(content: &[u8]) -> String {
+    let mut hasher = Sha256::new();
+    hasher.update(content);
+    let result = hasher.finalize();
+    hex::encode(result)
+}
+
+// sha2 doesn't include hex encoding, so we use a tiny inline helper.
+mod hex {
+    const HEX_CHARS: &[u8; 16] = b"0123456789abcdef";
+
+    pub fn encode(bytes: impl AsRef<[u8]>) -> String {
+        let bytes = bytes.as_ref();
+        let mut s = String::with_capacity(bytes.len() * 2);
+        for &b in bytes {
+            s.push(HEX_CHARS[(b >> 4) as usize] as char);
+            s.push(HEX_CHARS[(b & 0xf) as usize] as char);
+        }
+        s
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn known_hash() {
+        // SHA-256 of "hello"
+        let hash = content_hash(b"hello");
+        assert_eq!(
+            hash,
+            "2cf24dba5fb0a30e26e83b2ac5b9e29e1b161e5c1fa7425e73043362938b9824"
+        );
+    }
+
+    #[test]
+    fn empty_hash() {
+        let hash = content_hash(b"");
+        assert_eq!(
+            hash,
+            "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855"
+        );
+    }
+}

crates/bds-core/src/util/mod.rs

@@ -0,0 +1,5 @@
+mod slug;
+mod checksum;
+
+pub use slug::slugify;
+pub use checksum::content_hash;

crates/bds-core/src/util/slug.rs

@@ -0,0 +1,61 @@
+use deunicode::deunicode;
+
+/// Generate a URL-safe slug from a title string.
+///
+/// Transliterates Unicode to ASCII, lowercases, replaces non-alphanumeric
+/// chars with hyphens, and collapses/trims hyphens.
+pub fn slugify(input: &str) -> String {
+    let ascii = deunicode(input);
+    let lowered = ascii.to_lowercase();
+    let mut slug = String::with_capacity(lowered.len());
+    let mut prev_hyphen = true; // avoid leading hyphen
+    for c in lowered.chars() {
+        if c.is_ascii_alphanumeric() {
+            slug.push(c);
+            prev_hyphen = false;
+        } else if !prev_hyphen {
+            slug.push('-');
+            prev_hyphen = true;
+        }
+    }
+    // Trim trailing hyphen
+    if slug.ends_with('-') {
+        slug.pop();
+    }
+    slug
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn basic_slug() {
+        assert_eq!(slugify("Hello World"), "hello-world");
+    }
+
+    #[test]
+    fn unicode_slug() {
+        assert_eq!(slugify("Über die Brücke"), "uber-die-brucke");
+    }
+
+    #[test]
+    fn special_chars() {
+        assert_eq!(slugify("What's up? (2024)"), "what-s-up-2024");
+    }
+
+    #[test]
+    fn already_clean() {
+        assert_eq!(slugify("already-clean"), "already-clean");
+    }
+
+    #[test]
+    fn empty_input() {
+        assert_eq!(slugify(""), "");
+    }
+
+    #[test]
+    fn consecutive_special_chars() {
+        assert_eq!(slugify("a --- b"), "a-b");
+    }
+}