splice 2.6.4

//! Context extraction for span surroundings.
//!
//! Provides line-based context extraction using ropey for efficient
//! UTF-8 aware line/column calculations.

use crate::error::{Result, SpliceError};
use std::path::Path;

use crate::output::SpanContext;

/// Resolve context counts from -A/-B/-C flags following grep convention.
///
/// The grep convention is: -C (context both) sets the default for both sides,
/// but -A (after) and -B (before) can override -C if larger. This allows:
/// - `-C 3` for 3 lines before and after (default grep behavior)
/// - `-A 5 -B 2` for 5 lines after, 2 lines before (asymmetric)
/// - `-C 10 -A 5` for max(10, 5)=10 before, max(10, 0)=10 after
///
/// # Arguments
///
/// * `context_before` - Value from -B flag (lines before)
/// * `context_after` - Value from -A flag (lines after)
/// * `context_both` - Value from -C flag (lines on both sides)
///
/// # Returns
///
/// * `(before, after)` - Resolved context line counts for before and after
///
/// # Examples
///
/// ```
/// use splice::context::resolve_context_counts;
///
/// // -C 3 (default grep behavior)
/// let (before, after) = resolve_context_counts(0, 0, 3);
/// assert_eq!(before, 3);
/// assert_eq!(after, 3);
///
/// // -A 5 -B 2 (asymmetric)
/// let (before, after) = resolve_context_counts(2, 5, 0);
/// assert_eq!(before, 2);
/// assert_eq!(after, 5);
///
/// // -C 10 -A 5 (C overrides A when larger)
/// let (before, after) = resolve_context_counts(0, 5, 10);
/// assert_eq!(before, 10);
/// assert_eq!(after, 10);
/// ```
pub fn resolve_context_counts(
    context_before: usize,
    context_after: usize,
    context_both: usize,
) -> (usize, usize) {
    let before = context_before.max(context_both);
    let after = context_after.max(context_both);
    (before, after)
}

/// Extract context lines for a byte span with asymmetric before/after counts.
///
/// Given a file path and byte range, extracts lines before, within, and after
/// the span. Allows different amounts of context before vs after the match.
/// Uses UTF-8 byte offsets consistent with span coordinates.
///
/// # Arguments
///
/// * `path` - File path to read
/// * `byte_start` - Start byte offset (must be <= byte_end)
/// * `byte_end` - End byte offset (must be <= file size)
/// * `context_before` - Number of context lines before (default: 0)
/// * `context_after` - Number of context lines after (default: 0)
///
/// # Returns
///
/// * `Ok(SpanContext)` - Extracted context with before/selected/after arrays
/// * `Err(SpliceError)` - If file cannot be read or span is invalid
///
/// # Examples
///
/// ```no_run
/// use splice::context::extract_context_asymmetric;
/// use std::path::Path;
///
/// // 5 lines before, 2 lines after
/// let context = extract_context_asymmetric(Path::new("src/main.rs"), 100, 200, 5, 2)?;
/// println!("Before: {} lines, After: {} lines", context.before.len(), context.after.len());
/// # Ok::<(), splice::error::SpliceError>(())
/// ```
pub fn extract_context_asymmetric(
    path: &Path,
    byte_start: usize,
    byte_end: usize,
    context_before: usize,
    context_after: usize,
) -> Result<SpanContext> {
    use ropey::Rope;

    // Validate byte range
    if byte_start > byte_end {
        return Err(SpliceError::InvalidSpan {
            file: path.to_path_buf(),
            start: byte_start,
            end: byte_end,
            file_size: 0, // Will be updated after read
        });
    }

    // Read file
    let contents = std::fs::read(path).map_err(|e| SpliceError::IoContext {
        context: format!(
            "Failed to read file for context extraction: {}",
            path.display()
        ),
        source: e,
    })?;

    let file_size = contents.len();

    // Validate end is within file
    if byte_end > file_size {
        return Err(SpliceError::InvalidSpan {
            file: path.to_path_buf(),
            start: byte_start,
            end: byte_end,
            file_size,
        });
    }

    // Handle empty file case
    if file_size == 0 {
        return Ok(SpanContext {
            before: vec![],
            selected: vec![],
            after: vec![],
        });
    }

    // Create Rope for efficient line operations (UTF-8 aware)
    let rope =
        Rope::from_str(
            std::str::from_utf8(&contents).map_err(|e| SpliceError::InvalidUtf8 {
                file: path.to_path_buf(),
                source: e,
            })?,
        );

    // Convert byte offsets to line numbers (0-based)
    let start_line = rope.byte_to_line(byte_start);
    let end_line = rope.byte_to_line(byte_end.saturating_sub(1));

    // Calculate context boundaries with asymmetric values
    let context_start = start_line.saturating_sub(context_before);
    let context_end = (end_line + context_after + 1).min(rope.len_lines());

    // Extract before lines
    let before: Vec<String> = (context_start..start_line)
        .map(|i| rope.line(i).to_string())
        .collect();

    // Extract selected lines (the span itself)
    let selected: Vec<String> = (start_line..=end_line)
        .map(|i| rope.line(i).to_string())
        .collect();

    // Extract after lines (filter out empty trailing line from ropey behavior)
    let after: Vec<String> = (end_line + 1..context_end)
        .map(|i| rope.line(i).to_string())
        .filter(|line| !line.is_empty())
        .collect();

    Ok(SpanContext {
        before,
        selected,
        after,
    })
}

/// Extract context lines for a byte span with separate before/after context.
///
/// Given a file path and byte range, extracts lines before, within, and after
/// the span. Uses UTF-8 byte offsets consistent with span coordinates.
///
/// This is a convenience alias for [`extract_context_asymmetric`] with more
/// descriptive parameter names.
///
/// # Arguments
///
/// * `path` - File path to read
/// * `byte_start` - Start byte offset (must be <= byte_end)
/// * `byte_end` - End byte offset (must be <= file size)
/// * `context_lines_before` - Number of context lines before (default: 0)
/// * `context_lines_after` - Number of context lines after (default: 0)
///
/// # Returns
///
/// * `Ok(SpanContext)` - Extracted context with before/selected/after arrays
/// * `Err(SpliceError)` - If file cannot be read or span is invalid
///
/// # Examples
///
/// ```no_run
/// use splice::context::extract_context_with_before_after;
/// use std::path::Path;
///
/// let context = extract_context_with_before_after(Path::new("src/main.rs"), 100, 200, 2, 5)?;
/// println!("Before: {} lines, After: {} lines", context.before.len(), context.after.len());
/// # Ok::<(), splice::error::SpliceError>(())
/// ```
pub fn extract_context_with_before_after(
    path: &Path,
    byte_start: usize,
    byte_end: usize,
    context_lines_before: usize,
    context_lines_after: usize,
) -> Result<SpanContext> {
    extract_context_asymmetric(
        path,
        byte_start,
        byte_end,
        context_lines_before,
        context_lines_after,
    )
}

/// Extract context lines for a byte span.
///
/// Given a file path and byte range, extracts lines before, within, and after
/// the span. Uses UTF-8 byte offsets consistent with span coordinates.
///
/// # Arguments
///
/// * `path` - File path to read
/// * `byte_start` - Start byte offset (must be <= byte_end)
/// * `byte_end` - End byte offset (must be <= file size)
/// * `context_lines` - Number of context lines before/after (default: 3)
///
/// # Returns
///
/// * `Ok(SpanContext)` - Extracted context with before/selected/after arrays
/// * `Err(SpliceError)` - If file cannot be read or span is invalid
///
/// # Examples
///
/// ```no_run
/// use splice::context::extract_context;
/// use std::path::Path;
///
/// let context = extract_context(Path::new("src/main.rs"), 100, 200, 3)?;
/// println!("Before: {} lines", context.before.len());
/// # Ok::<(), splice::error::SpliceError>(())
/// ```
pub fn extract_context(
    path: &Path,
    byte_start: usize,
    byte_end: usize,
    context_lines: usize,
) -> Result<SpanContext> {
    // Delegate to asymmetric version with symmetric context
    extract_context_asymmetric(path, byte_start, byte_end, context_lines, context_lines)
}

#[cfg(test)]
mod tests {
    use super::*;
    use std::io::Write;
    use tempfile::NamedTempFile;

    #[test]
    fn test_extract_context_basic() {
        let mut file = NamedTempFile::new().unwrap();
        writeln!(file, "line 1").unwrap();
        writeln!(file, "line 2").unwrap();
        writeln!(file, "line 3").unwrap();
        writeln!(file, "line 4").unwrap();
        writeln!(file, "line 5").unwrap();

        // Context for lines 2-3
        // "line 1\n" = 7 bytes, "line 2\n" = 7 bytes, "line 3\n" = 7 bytes
        // Bytes 7-20 covers "line 2\nline 3\n"
        let context = extract_context(file.path(), 7, 20, 1).unwrap();

        assert_eq!(context.before.len(), 1); // "line 1"
        assert_eq!(context.selected.len(), 2); // "line 2", "line 3"
        assert_eq!(context.after.len(), 1); // "line 4"
    }

    #[test]
    fn test_extract_context_zero_context() {
        let mut file = NamedTempFile::new().unwrap();
        writeln!(file, "line 1").unwrap();
        writeln!(file, "line 2").unwrap();
        writeln!(file, "line 3").unwrap();

        // "line 1\n" = bytes 0-6, "line 2\n" = bytes 7-13
        let context = extract_context(file.path(), 7, 13, 0).unwrap();

        assert_eq!(context.before.len(), 0);
        assert_eq!(context.selected.len(), 1);
        assert_eq!(context.after.len(), 0);
    }

    #[test]
    fn test_extract_context_start_of_file() {
        let mut file = NamedTempFile::new().unwrap();
        writeln!(file, "line 1").unwrap();
        writeln!(file, "line 2").unwrap();
        writeln!(file, "line 3").unwrap();

        // "line 1\nline 2\n" = bytes 0-13
        let context = extract_context(file.path(), 0, 13, 2).unwrap();

        assert_eq!(context.before.len(), 0); // No lines before start
        assert_eq!(context.selected.len(), 2);
        assert_eq!(context.after.len(), 1); // "line 3"
    }

    #[test]
    fn test_extract_context_end_of_file() {
        let mut file = NamedTempFile::new().unwrap();
        writeln!(file, "line 1").unwrap();
        writeln!(file, "line 2").unwrap();
        writeln!(file, "line 3").unwrap();

        // "line 1\nline 2\n" = bytes 0-13, "line 3\n" = bytes 14-20
        let context = extract_context(file.path(), 14, 20, 2).unwrap();

        assert_eq!(context.before.len(), 2); // "line 1", "line 2"
        assert_eq!(context.selected.len(), 1); // "line 3"
        assert_eq!(context.after.len(), 0); // No lines after end
    }

    #[test]
    fn test_extract_context_utf8_multibyte() {
        let mut file = NamedTempFile::new().unwrap();
        // Use emoji (multi-byte UTF-8)
        writeln!(file, "line 🦀 1").unwrap();
        writeln!(file, "line 🚀 2").unwrap();
        writeln!(file, "line ⭐ 3").unwrap();

        let contents = std::fs::read(file.path()).unwrap();
        // Find the start of "line 🚀 2"
        let rocket_line_start = contents.iter().position(|&b| b == b'2').unwrap();
        // Find the newline after "line 🚀 2"
        let rocket_line_end = contents
            .iter()
            .skip(rocket_line_start)
            .position(|&b| b == b'\n')
            .unwrap()
            + rocket_line_start;

        // Context should still work with multi-byte characters
        let context = extract_context(file.path(), rocket_line_start, rocket_line_end, 1).unwrap();

        assert_eq!(context.selected.len(), 1);
        assert!(context.selected[0].contains("🚀"));
    }

    #[test]
    fn test_extract_context_invalid_span() {
        let mut file = NamedTempFile::new().unwrap();
        writeln!(file, "line 1").unwrap();

        // start > end
        let result = extract_context(file.path(), 10, 5, 1);
        assert!(result.is_err());

        // end beyond file size
        let result = extract_context(file.path(), 0, 1000, 1);
        assert!(result.is_err());
    }

    #[test]
    fn test_extract_context_empty_file() {
        let file = NamedTempFile::new().unwrap();
        let result = extract_context(file.path(), 0, 0, 1);
        assert!(result.is_ok());
        // Empty file should return empty context
        let context = result.unwrap();
        assert_eq!(context.before.len(), 0);
        assert_eq!(context.selected.len(), 0);
        assert_eq!(context.after.len(), 0);
    }

    #[test]
    fn test_extract_context_large_context_request() {
        let mut file = NamedTempFile::new().unwrap();
        writeln!(file, "line 1").unwrap();
        writeln!(file, "line 2").unwrap();
        writeln!(file, "line 3").unwrap();

        // Request more context lines than exist (bytes 7-13 is "line 2\n")
        let context = extract_context(file.path(), 7, 13, 100).unwrap();

        // Should saturate at file boundaries
        assert_eq!(context.before.len(), 1); // Only "line 1"
        assert_eq!(context.selected.len(), 1);
        assert_eq!(context.after.len(), 1); // Only "line 3"
    }

    #[test]
    fn test_extract_context_asymmetric_basic() {
        let mut file = NamedTempFile::new().unwrap();
        writeln!(file, "line 1").unwrap();
        writeln!(file, "line 2").unwrap();
        writeln!(file, "line 3").unwrap();
        writeln!(file, "line 4").unwrap();
        writeln!(file, "line 5").unwrap();
        writeln!(file, "line 6").unwrap();
        writeln!(file, "line 7").unwrap();

        // Context for line 3-4
        // Each line is 7 bytes: "line N\n" = 6 + 1
        // line 1: 0-6, line 2: 7-13, line 3: 14-20, line 4: 21-27
        // Request 2 lines before, 1 line after
        let context = extract_context_asymmetric(file.path(), 14, 28, 2, 1).unwrap();

        assert_eq!(context.before.len(), 2); // "line 1", "line 2"
        assert_eq!(context.selected.len(), 2); // "line 3", "line 4"
        assert_eq!(context.after.len(), 1); // "line 5"
    }

    #[test]
    fn test_extract_context_asymmetric_zero_before() {
        let mut file = NamedTempFile::new().unwrap();
        writeln!(file, "line 1").unwrap();
        writeln!(file, "line 2").unwrap();
        writeln!(file, "line 3").unwrap();
        writeln!(file, "line 4").unwrap();

        // Each line is 7 bytes
        // line 2: 7-13, Request 0 before, 2 after
        let context = extract_context_asymmetric(file.path(), 7, 14, 0, 2).unwrap();

        assert_eq!(context.before.len(), 0);
        assert_eq!(context.selected.len(), 1);
        assert_eq!(context.after.len(), 2); // "line 3", "line 4"
    }

    #[test]
    fn test_extract_context_asymmetric_zero_after() {
        let mut file = NamedTempFile::new().unwrap();
        writeln!(file, "line 1").unwrap();
        writeln!(file, "line 2").unwrap();
        writeln!(file, "line 3").unwrap();
        writeln!(file, "line 4").unwrap();

        // Each line is 7 bytes
        // line 3: 14-20, Request 2 before, 0 after
        let context = extract_context_asymmetric(file.path(), 14, 21, 2, 0).unwrap();

        assert_eq!(context.before.len(), 2); // "line 1", "line 2"
        assert_eq!(context.selected.len(), 1);
        assert_eq!(context.after.len(), 0);
    }
}