bone-agent 1.4.0 → 2.0.0

package/src/tools/helpers/file_helpers.py

@@ -1,189 +0,0 @@
-"""Shared utilities for file operations."""
-
-import time
-from functools import lru_cache
-from pathlib import Path
-from typing import Optional
-
-_GITIGNORE_SPEC_REGISTRY = {}
-
-# Performance metrics for gitignore filtering
-_gitignore_filter_times = []
-_gitignore_spec_hits = 0
-
-
-def _register_gitignore_spec(gitignore_spec) -> int:
-    """Register a PathSpec for cached lookups and return its key.
-
-    Args:
-        gitignore_spec: PathSpec object to register
-
-    Returns:
-        Registry key for the PathSpec object
-    """
-    if gitignore_spec is None:
-        return 0
-    key = id(gitignore_spec)
-    _GITIGNORE_SPEC_REGISTRY[key] = gitignore_spec
-    return key
-
-
-@lru_cache(maxsize=1000)
-def _is_ignored_cached(path_str: str, repo_root_str: str, spec_key: int) -> bool:
-    """Cached version of gitignore check.
-
-    Args:
-        path_str: String representation of path to check
-        repo_root_str: String representation of repository root
-        spec_key: Registry key for the PathSpec object
-
-    Returns:
-        True if path is ignored by gitignore spec
-    """
-    gitignore_spec = _GITIGNORE_SPEC_REGISTRY.get(spec_key)
-    if gitignore_spec is None:
-        return False
-
-    from utils.gitignore_filter import is_path_ignored
-
-    path = Path(path_str)
-    repo_root = Path(repo_root_str)
-    is_ignored, _ = is_path_ignored(path, repo_root, gitignore_spec)
-    return is_ignored
-
-
-def _is_reserved_windows_name(name: str) -> bool:
-    """Check if filename is a reserved Windows device name.
-
-    Args:
-        name: Filename to check (without path)
-
-    Returns:
-        True if name is reserved (e.g., CON, PRN, NUL)
-    """
-    if not name:
-        return False
-    base = name.upper().split('.')[0]
-    return base in {
-        'CON', 'PRN', 'AUX', 'NUL',
-        'COM1', 'COM2', 'COM3', 'COM4', 'COM5', 'COM6', 'COM7', 'COM8', 'COM9',
-        'LPT1', 'LPT2', 'LPT3', 'LPT4', 'LPT5', 'LPT6', 'LPT7', 'LPT8', 'LPT9'
-    }
-
-
-class GitignoreFilter:
-    """Centralized gitignore filtering logic.
-
-    This class provides a single interface for checking if paths should be
-    excluded based on .gitignore rules, combining fast-path checks with
-    full gitignore spec evaluation.
-
-    Example:
-        filter = GitignoreFilter(repo_root=Path("/project"), gitignore_spec=spec)
-        if filter.is_ignored(path):
-            continue  # Skip this path
-    """
-
-    def __init__(
-        self,
-        repo_root: Path,
-        gitignore_spec = None
-    ):
-        """Initialize the gitignore filter.
-
-        Args:
-            repo_root: Repository root directory
-            gitignore_spec: Optional PathSpec for .gitignore filtering
-        """
-        self.repo_root = repo_root
-        self.gitignore_spec = gitignore_spec
-        self._spec_key = _register_gitignore_spec(gitignore_spec) if gitignore_spec else None
-
-    def is_ignored(self, path: Path) -> bool:
-        """Check if a path should be ignored by gitignore rules.
-
-        Args:
-            path: Path object to check
-
-        Returns:
-            True if path should be ignored, False otherwise
-        """
-        global _gitignore_spec_hits
-        start_time = time.time()
-
-        # Full gitignore check (only if spec is provided)
-        if self.gitignore_spec is not None and self._spec_key is not None:
-            # Only check paths within the repo
-            try:
-                path.relative_to(self.repo_root)
-                is_ignored = _is_ignored_cached(str(path), str(self.repo_root), self._spec_key)
-                if is_ignored:
-                    _gitignore_spec_hits += 1
-                _gitignore_filter_times.append(time.time() - start_time)
-                return is_ignored
-            except ValueError:
-                # Path is outside repo, don't filter
-                pass
-
-        _gitignore_filter_times.append(time.time() - start_time)
-        return False
-
-    def should_include(self, path: Path) -> bool:
-        """Check if a path should be included (inverse of is_ignored).
-
-        This is provided for readability when filtering:
-            files = [f for f in files if filter.should_include(f)]
-
-        Args:
-            path: Path object to check
-
-        Returns:
-            True if path should be included, False if ignored
-        """
-        return not self.is_ignored(path)
-
-
-def get_gitignore_filter_metrics() -> dict:
-    """Get performance metrics for gitignore filtering operations.
-
-    Returns:
-        Dictionary with metrics:
-        - total_checks: Total number of filter checks
-        - spec_hits: Number of matches by gitignore spec
-        - avg_filter_time: Average filter time in seconds
-        - cache_hit_rate: Estimated LRU cache hit rate
-    """
-    if not _gitignore_filter_times:
-        return {
-            "total_checks": 0,
-            "spec_hits": 0,
-            "avg_filter_time": 0,
-            "cache_hit_rate": 0
-        }
-
-    total_checks = len(_gitignore_filter_times)
-    total_hits = _gitignore_spec_hits
-
-    # Estimate cache hit rate from _is_ignored_cached
-    try:
-        from functools import _is_ignored_cached as cached_func
-        cache_info = cached_func.cache_info() if hasattr(cached_func, 'cache_info') else None
-        cache_hit_rate = cache_info.hits / (cache_info.hits + cache_info.misses) if cache_info and (cache_info.hits + cache_info.misses) > 0 else 0
-    except:
-        cache_hit_rate = 0
-
-    return {
-        "total_checks": total_checks,
-        "spec_hits": _gitignore_spec_hits,
-        "avg_filter_time": sum(_gitignore_filter_times) / total_checks,
-        "cache_hit_rate": cache_hit_rate
-    }
-
-
-def clear_gitignore_filter_metrics():
-    """Clear all accumulated metrics for testing or monitoring reset."""
-    global _gitignore_spec_hits
-    _gitignore_filter_times.clear()
-    _gitignore_spec_hits = 0
-    # Clear LRU cache for _is_ignored_cached
-    _is_ignored_cached.cache_clear()

package/src/tools/helpers/formatters.py

@@ -1,411 +0,0 @@
-"""Result formatting utilities for tool output and diffs."""
-
-import os
-import re
-import difflib
-from rich.text import Text
-
-# Import constants module to access centralized values
-try:
-    from ..constants import (
-        DEFAULT_TERMINAL_WIDTH,
-        FORMATTER_MAX_LINES,
-    )
-except ImportError:
-    # Fallback for standalone usage
-    DEFAULT_TERMINAL_WIDTH = 80
-    FORMATTER_MAX_LINES = 100
-
-# Shell output truncation (lazy import to avoid circular dependency)
-_SHELL_MAX_LINES = None
-
-def _get_shell_max_lines():
-    """Get shell output line limit from settings (lazy import)."""
-    global _SHELL_MAX_LINES
-    if _SHELL_MAX_LINES is None:
-        try:
-            from utils.settings import MAX_SHELL_OUTPUT_LINES
-            _SHELL_MAX_LINES = MAX_SHELL_OUTPUT_LINES
-        except ImportError:
-            _SHELL_MAX_LINES = 200
-    return _SHELL_MAX_LINES
-
-
-def _detect_newline(text):
-    """Detect the newline character used in text."""
-    if "\r\n" in text:
-        return "\r\n"
-    if "\n" in text:
-        return "\n"
-    return os.linesep
-
-
-def _colorize_numbered_lines(lines, file_path=None):
-    """Apply color highlighting to diff lines.
-
-    - Removed lines (-): bold white text on red background, full width
-    - Added lines (+): bold white text on green background, full width
-    - Unchanged lines: dim grey text
-
-    Returns:
-        Rich Text object with styled content
-    """
-    # Get terminal width with fallback for non-TTY environments
-    try:
-        terminal_width = os.get_terminal_size().columns
-    except (OSError, AttributeError):
-        terminal_width = DEFAULT_TERMINAL_WIDTH  # Fallback default
-
-    result = Text()
-    for line in lines:
-        # Check the sign character (7th character, index 6)
-        # Format: "   5 - text" or "   6 + text" or "   7   text"
-        # Where indices 0-4 are line number, 5 is space, 6 is sign
-        if len(line) >= 7:
-            sign = line[6]
-
-            if sign == "-":
-                # Removed line - red background
-                padded = line.ljust(terminal_width)
-                result.append(padded, style="on #870101")
-            elif sign == "+":
-                # Added line - green background
-                padded = line.ljust(terminal_width)
-                result.append(padded, style="on #005f00")
-            else:
-                # Unchanged line - dim grey
-                result.append(line, style="dim")
-        else:
-            result.append(line)
-
-        result.append("\n")
-
-    return result
-
-
-def _build_numbered_diff_lines(original_content, new_content, context_lines):
-    """Build numbered diff lines from original and new content."""
-    if original_content == new_content:
-        return [], 0, 0
-
-    diff_lines = list(difflib.unified_diff(
-        original_content.splitlines(keepends=False),
-        new_content.splitlines(keepends=False),
-        fromfile="old",
-        tofile="new",
-        n=context_lines,
-        lineterm="",
-    ))
-
-    hunk_re = re.compile(r"@@ -(\d+)(?:,(\d+))? \+(\d+)(?:,(\d+))? @@")
-    old_line = None
-    new_line = None
-    removed = 0
-    added = 0
-    formatted_lines = []
-
-    for line in diff_lines:
-        if line.startswith("--- ") or line.startswith("+++ "):
-            continue
-        if line.startswith("@@ "):
-            match = hunk_re.match(line)
-            if match:
-                old_line = int(match.group(1))
-                new_line = int(match.group(3))
-            continue
-
-        if old_line is None or new_line is None:
-            continue
-
-        sign = line[:1]
-        text = line[1:]
-        if sign == " ":
-            line_no = old_line
-            old_line += 1
-            new_line += 1
-        elif sign == "-":
-            line_no = old_line
-            old_line += 1
-            removed += 1
-        elif sign == "+":
-            line_no = new_line
-            new_line += 1
-            added += 1
-        else:
-            continue
-
-        formatted_lines.append(f"{line_no:>5} {sign} {text}")
-
-    return formatted_lines, added, removed
-
-
-def _build_diff(
-    original_content,
-    new_content,
-    file_path,
-    context_lines,
-    show_header=False,
-    repo_root=None
-):
-    """Build a diff with optional header or summary line.
-
-    Args:
-        original_content: Original file content
-        new_content: Modified file content
-        file_path: Path to the file being edited
-        context_lines: Number of context lines for diff
-        show_header: If True, show filename header (used in preview)
-        repo_root: Required when show_header=True to compute relative path
-
-    Returns:
-        Rich Text object with styled diff and optional header/summary
-    """
-    formatted_lines, added, removed = _build_numbered_diff_lines(
-        original_content, new_content, context_lines,
-    )
-
-    # Build header or summary based on mode
-    if show_header:
-        try:
-            rel_path = file_path.relative_to(repo_root)
-        except (ValueError, TypeError):
-            rel_path = file_path
-        header = f"{rel_path} | -{removed} | +{added}"
-    else:
-        header = None
-        summary = f"Changes: +{added}, -{removed}"
-
-    # Handle empty case
-    if not formatted_lines:
-        result = Text()
-        if show_header:
-            result.append(f"\n{header}\n")
-            result.append("(no changes)\n", style="dim")
-        else:
-            result.append("(no changes)\n", style="dim")
-        return result
-
-    # Get colored diff as Text object
-    diff_text = _colorize_numbered_lines(formatted_lines, file_path)
-
-    # Build output based on mode
-    result = Text()
-    if show_header:
-        result.append(f"\n{header}\n")
-        result.append(diff_text)
-    else:
-        result.append(diff_text)
-        result.append(f"{summary}\n")
-
-    return result
-
-
-def _normalize_search_replace_for_newlines(search, replace, newline):
-    """Normalize search/replace text to match file's newline characters."""
-    if newline == "\n":
-        return search, replace, False
-    if "\n" not in search or "\r\n" in search:
-        return search, replace, False
-    normalized_search = search.replace("\n", newline)
-    normalized_replace = replace.replace("\n", newline)
-    return normalized_search, normalized_replace, True
-
-
-def _is_rg_match_line(line):
-    """Return True if an rg output line is a match (not context/separator).
-
-    Match formats:
-      path:line:content   (colon before line number)
-      line:content        (single-file search, line starts with number)
-    Context/separators are excluded (path-line:content uses hyphen, not colon).
-    """
-    if re.search(r':\d+:', line):
-        return True
-    return bool(re.match(r'^\d+:', line))
-
-
-def format_tool_result(result, command=None, is_rg=False, debug_mode=False, max_matches=None):
-    """Format subprocess result for model consumption.
-
-    Args:
-        result: subprocess.CompletedProcess result
-        command: The command that was executed (for display and mode detection)
-        is_rg: Whether this was an rg command (affects empty output and counting)
-        debug_mode: If True, show full output; if False, show summary only
-        max_matches: Maximum number of matches to return (0 = use line-based limit, >0 = match limit)
-
-    Returns:
-        str: Formatted result with exit code
-    """
-    # Number of trailing context lines to include after the last match when truncating
-    _RG_TRAILING_CONTEXT_LINES = 5
-
-    output = (result.stdout or "") + (result.stderr or "")
-    output = output.strip()
-
-    if not output:
-        if is_rg and result.returncode == 1:
-            output = "no matches found"
-        else:
-            output = "(no output)"
-
-    # For rg commands, apply smart truncation to prevent context explosion
-    if is_rg:
-        label = "files" if command and "--files-with-matches" in command.lower() else "matches"
-        MAX_LINES = FORMATTER_MAX_LINES
-
-        # Exit code 0: found matches, Exit code 1: no matches
-        if result.returncode == 1:
-            # No matches found - rg returns 1 in this case
-            count = 0
-        elif result.returncode == 0:
-            # Count actual matches (lines with ':number:' pattern), not context lines
-            if "--files-with-matches" in (command or "").lower():
-                # files-with-matches mode: count lines (each line is a file)
-                lines = [line for line in output.splitlines() if line.strip()]
-                count = len(lines)
-            else:
-                # Normal mode: count match lines (lines with ':number:' pattern), not context
-                count = sum(1 for line in output.splitlines() if _is_rg_match_line(line))
-        else:
-            # Error occurred (exit code 2 or higher)
-            count = 0
-
-        # Handle no matches
-        if result.returncode == 1:
-            return f"exit_code={result.returncode}\n{label}=0\nNo matches found\n\n"
-        elif count == 0:
-            # Exit code 0 but no output - unusual but possible
-            return f"exit_code={result.returncode}\n{output}\n\n"
-
-        # Truncate output
-        output_lines = output.splitlines()
-
-        # Use max_matches directly if positive (>0 means match limit, 0/None means line-based fallback)
-        effective_max_matches = max_matches if max_matches and max_matches > 0 else None
-
-        if effective_max_matches is not None and "--files-with-matches" in (command or "").lower():
-            # files_with_matches mode: truncate by file count (each line is a file)
-            if len(output_lines) > effective_max_matches:
-                truncated = "\n".join(output_lines[:effective_max_matches])
-                omitted = len(output_lines) - effective_max_matches
-                output = f"{truncated}\n\n... ({omitted} more {label} truncated)"
-                return f"exit_code={result.returncode}\n{label}={len(output_lines)}\n{output}\n\n"
-            else:
-                output = "\n".join(output_lines)
-                return f"exit_code={result.returncode}\n{label}={count}\n{output}\n\n"
-
-        if effective_max_matches is not None:
-            # Match-based truncation: find lines that are actual matches (contain :number:)
-            # and cut after the Nth match including its trailing context
-            is_match = [_is_rg_match_line(line) for line in output_lines]
-
-            # Compute total match count once upfront
-            total_matches = sum(is_match)
-
-            # Find cutoff point after Nth match
-            match_count = 0
-            cutoff = len(output_lines)
-            for i, match in enumerate(is_match):
-                if match:
-                    match_count += 1
-                    if match_count >= effective_max_matches:
-                        # Include up to _RG_TRAILING_CONTEXT_LINES trailing context lines after this match
-                        cutoff = i + 1
-                        for j in range(i + 1, min(i + 1 + _RG_TRAILING_CONTEXT_LINES, len(is_match))):
-                            if is_match[j]:
-                                break
-                            cutoff = j + 1
-                        break
-
-            if cutoff < len(output_lines):
-                truncated = "\n".join(output_lines[:cutoff])
-                omitted_matches = max(0, total_matches - effective_max_matches)
-                output = f"{truncated}\n\n... ({omitted_matches} more {label} truncated)"
-                return f"exit_code={result.returncode}\n{label}={total_matches}\n{output}\n\n"
-            else:
-                output = "\n".join(output_lines)
-                return f"exit_code={result.returncode}\n{label}={total_matches}\n{output}\n\n"
-
-        # Fallback: line-based truncation
-        if len(output_lines) > MAX_LINES:
-            truncated = "\n".join(output_lines[:MAX_LINES])
-            omitted = len(output_lines) - MAX_LINES
-            output = f"{truncated}\n\n... ({omitted} more {label} truncated)"
-        else:
-            output = "\n".join(output_lines)
-
-        return f"exit_code={result.returncode}\n{label}={count}\n{output}\n\n"
-
-    # For non-rg shell commands: apply head+tail truncation
-    output_lines = output.splitlines()
-    max_lines = _get_shell_max_lines()
-
-    if len(output_lines) > max_lines:
-        head_count = max_lines // 2
-        tail_count = max_lines - head_count
-        omitted = len(output_lines) - max_lines
-        head = "\n".join(output_lines[:head_count])
-        tail = "\n".join(output_lines[-tail_count:])
-        output = f"{head}\n\n... ({omitted} lines omitted) ...\n\n{tail}"
-    else:
-        output = "\n".join(output_lines)
-
-    return f"exit_code={result.returncode}\n{output}\n\n"
-
-
-
-def format_file_result(exit_code, content=None, error=None, path=None,
-                       lines_read=None, start_line=None, truncated=False, items_count=None,
-                       truncation_info=None):
-    """Format file operation result for model consumption.
-
-    Args:
-        exit_code: Exit code (0 for success, 1 for error)
-        content: Optional content string (for successful reads)
-        error: Optional error message (for failures)
-        path: Path to the file/directory
-        lines_read: Number of lines read (for file reads)
-        start_line: 1-based starting line number for file reads
-        truncated: Whether content was truncated (for file reads)
-        items_count: Number of items (for directory listings)
-        truncation_info: Optional dict with truncation metadata (total, shown, omitted)
-
-    Returns:
-        str: Formatted result with exit code and metadata
-    """
-    metadata_parts = [f"exit_code={exit_code}"]
-
-    if path is not None:
-        metadata_parts.append(f"path={path}")
-
-    if lines_read is not None:
-        metadata_parts.append(f"lines_read={lines_read}")
-
-    if start_line is not None:
-        metadata_parts.append(f"start_line={start_line}")
-
-    if truncated:
-        metadata_parts.append("truncated=true")
-        if truncation_info:
-            metadata_parts.append(
-                f"truncation_info=total:{truncation_info['total']},"
-                f"shown:{truncation_info['shown']},"
-                f"omitted:{truncation_info['omitted']}"
-            )
-
-    if items_count is not None:
-        metadata_parts.append(f"items_count={items_count}")
-
-    metadata = " ".join(metadata_parts)
-
-    if error:
-        return f"{metadata}\nerror: {error}"
-
-    if content is not None:
-        return f"{metadata}\n{content}\n\n"
-
-    return f"{metadata}\n\n"
-
-
-# format_file_preview removed - now using Rich Syntax directly in agentic.py