ambers 0.3.8

/// Subtype 18: Variable Attributes Record.
///
/// Text-based format (NOT pascal-string binary like subtypes 21/22).
/// Grammar per PSPP Developer Guide:
///
///   record     = var_set ("/" var_set)*
///   var_set    = var_name ":" attribute+
///   attribute  = attr_name "(" value+ ")"
///   value      = "'" text "'" "\n"
///
/// Example: `age:$@Role('0'\n)/income:$@Role('1'\n)`
/// A parsed variable attribute set.
#[derive(Debug, Clone)]
pub struct VarAttributeSet {
    /// Variable name (may be short name — resolved in dictionary.rs).
    pub var_name: String,
    /// (attribute_name, [values]) pairs.
    pub attributes: Vec<(String, Vec<String>)>,
}

/// Parse a subtype 18 text blob into variable attribute sets.
pub fn parse_var_attributes(data: &[u8]) -> Vec<VarAttributeSet> {
    let text = String::from_utf8_lossy(data);
    let text = text.trim_end_matches('\0');

    let mut result = Vec::new();

    // Split on '/' to get individual variable attribute sets.
    // Be careful: '/' can appear inside quoted values, so we need to
    // track whether we're inside quotes.
    for var_chunk in split_var_sets(text) {
        let var_chunk = var_chunk.trim();
        if var_chunk.is_empty() {
            continue;
        }

        // Split on first ':' to get var_name and attribute text
        let colon_pos = match var_chunk.find(':') {
            Some(p) => p,
            None => continue,
        };

        let var_name = var_chunk[..colon_pos].trim().to_string();
        let attr_text = &var_chunk[colon_pos + 1..];

        let attributes = parse_attributes(attr_text);
        if !attributes.is_empty() {
            result.push(VarAttributeSet {
                var_name,
                attributes,
            });
        }
    }

    result
}

/// Split the record text on '/' delimiters, respecting single-quoted values.
fn split_var_sets(text: &str) -> Vec<&str> {
    let mut result = Vec::new();
    let mut start = 0;
    let mut in_quote = false;
    let bytes = text.as_bytes();

    for i in 0..bytes.len() {
        match bytes[i] {
            b'\'' => in_quote = !in_quote,
            b'/' if !in_quote => {
                result.push(&text[start..i]);
                start = i + 1;
            }
            _ => {}
        }
    }
    // Last segment
    if start < text.len() {
        result.push(&text[start..]);
    }

    result
}

/// Parse the attribute portion: `AttrName1('val1'\n)AttrName2('val2'\n)`
fn parse_attributes(text: &str) -> Vec<(String, Vec<String>)> {
    let mut result = Vec::new();
    let mut pos = 0;
    let bytes = text.as_bytes();

    while pos < bytes.len() {
        // Skip whitespace
        while pos < bytes.len() && bytes[pos].is_ascii_whitespace() {
            pos += 1;
        }
        if pos >= bytes.len() {
            break;
        }

        // Read attribute name (up to '(')
        let name_start = pos;
        while pos < bytes.len() && bytes[pos] != b'(' {
            pos += 1;
        }
        if pos >= bytes.len() {
            break;
        }
        let attr_name = text[name_start..pos].trim().to_string();
        pos += 1; // skip '('

        // Read values until closing ')'
        let mut values = Vec::new();
        while pos < bytes.len() && bytes[pos] != b')' {
            // Skip whitespace/newlines between values
            while pos < bytes.len()
                && (bytes[pos] == b'\n' || bytes[pos] == b'\r' || bytes[pos] == b' ')
            {
                pos += 1;
            }
            if pos >= bytes.len() || bytes[pos] == b')' {
                break;
            }

            // Expect opening quote
            if bytes[pos] == b'\'' {
                pos += 1;
                // Read until closing quote
                let val_start = pos;
                while pos < bytes.len() && bytes[pos] != b'\'' {
                    pos += 1;
                }
                let val = text[val_start..pos].to_string();
                values.push(val);
                if pos < bytes.len() {
                    pos += 1; // skip closing quote
                }
                // Skip trailing newline (part of the format spec)
                if pos < bytes.len() && bytes[pos] == b'\n' {
                    pos += 1;
                }
            } else {
                // Unexpected character — skip to avoid infinite loop
                pos += 1;
            }
        }
        if pos < bytes.len() {
            pos += 1; // skip ')'
        }

        if !attr_name.is_empty() {
            result.push((attr_name, values));
        }
    }

    result
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn test_parse_single_role() {
        let data = b"age:$@Role('0'\n)";
        let sets = parse_var_attributes(data);
        assert_eq!(sets.len(), 1);
        assert_eq!(sets[0].var_name, "age");
        assert_eq!(sets[0].attributes.len(), 1);
        assert_eq!(sets[0].attributes[0].0, "$@Role");
        assert_eq!(sets[0].attributes[0].1, vec!["0"]);
    }

    #[test]
    fn test_parse_multiple_vars() {
        let data = b"age:$@Role('0'\n)/income:$@Role('1'\n)/region:$@Role('4'\n)";
        let sets = parse_var_attributes(data);
        assert_eq!(sets.len(), 3);
        assert_eq!(sets[0].var_name, "age");
        assert_eq!(sets[0].attributes[0].1, vec!["0"]);
        assert_eq!(sets[1].var_name, "income");
        assert_eq!(sets[1].attributes[0].1, vec!["1"]);
        assert_eq!(sets[2].var_name, "region");
        assert_eq!(sets[2].attributes[0].1, vec!["4"]);
    }

    #[test]
    fn test_parse_multiple_attrs() {
        let data = b"age:$@Role('0'\n)CustomNote('hello'\n)";
        let sets = parse_var_attributes(data);
        assert_eq!(sets.len(), 1);
        assert_eq!(sets[0].attributes.len(), 2);
        assert_eq!(sets[0].attributes[0].0, "$@Role");
        assert_eq!(sets[0].attributes[0].1, vec!["0"]);
        assert_eq!(sets[0].attributes[1].0, "CustomNote");
        assert_eq!(sets[0].attributes[1].1, vec!["hello"]);
    }

    #[test]
    fn test_parse_empty() {
        let sets = parse_var_attributes(b"");
        assert!(sets.is_empty());
    }

    #[test]
    fn test_parse_all_roles() {
        let data = b"v1:$@Role('0'\n)/v2:$@Role('1'\n)/v3:$@Role('2'\n)/v4:$@Role('3'\n)/v5:$@Role('4'\n)/v6:$@Role('5'\n)";
        let sets = parse_var_attributes(data);
        assert_eq!(sets.len(), 6);
        for (i, set) in sets.iter().enumerate() {
            assert_eq!(set.attributes[0].1[0], i.to_string());
        }
    }
}