1package tools
2
3import (
4 "bufio"
5 "context"
6 "encoding/json"
7 "fmt"
8 "io"
9 "os"
10 "os/exec"
11 "path/filepath"
12 "regexp"
13 "sort"
14 "strconv"
15 "strings"
16 "sync"
17 "time"
18
19 "github.com/charmbracelet/crush/internal/config"
20 "github.com/charmbracelet/crush/internal/fsext"
21)
22
23// regexCache provides thread-safe caching of compiled regex patterns
24type regexCache struct {
25 cache map[string]*regexp.Regexp
26 mu sync.RWMutex
27}
28
29// newRegexCache creates a new regex cache
30func newRegexCache() *regexCache {
31 return ®exCache{
32 cache: make(map[string]*regexp.Regexp),
33 }
34}
35
36// get retrieves a compiled regex from cache or compiles and caches it
37func (rc *regexCache) get(pattern string) (*regexp.Regexp, error) {
38 // Try to get from cache first (read lock)
39 rc.mu.RLock()
40 if regex, exists := rc.cache[pattern]; exists {
41 rc.mu.RUnlock()
42 return regex, nil
43 }
44 rc.mu.RUnlock()
45
46 // Compile the regex (write lock)
47 rc.mu.Lock()
48 defer rc.mu.Unlock()
49
50 // Double-check in case another goroutine compiled it while we waited
51 if regex, exists := rc.cache[pattern]; exists {
52 return regex, nil
53 }
54
55 // Compile and cache the regex
56 regex, err := regexp.Compile(pattern)
57 if err != nil {
58 return nil, err
59 }
60
61 rc.cache[pattern] = regex
62 return regex, nil
63}
64
65// Global regex cache instances
66var (
67 searchRegexCache = newRegexCache()
68 globRegexCache = newRegexCache()
69 // Pre-compiled regex for glob conversion (used frequently)
70 globBraceRegex = regexp.MustCompile(`\{([^}]+)\}`)
71)
72
73type GrepParams struct {
74 Pattern string `json:"pattern"`
75 Path string `json:"path"`
76 Include string `json:"include"`
77 LiteralText bool `json:"literal_text"`
78}
79
80type grepMatch struct {
81 path string
82 modTime time.Time
83 lineNum int
84 lineText string
85}
86
87type GrepResponseMetadata struct {
88 NumberOfMatches int `json:"number_of_matches"`
89 Truncated bool `json:"truncated"`
90}
91
92type grepTool struct{}
93
94const (
95 GrepToolName = "grep"
96 grepDescription = `Fast content search tool that finds files containing specific text or patterns, returning matching file paths sorted by modification time (newest first).
97
98WHEN TO USE THIS TOOL:
99- Use when you need to find files containing specific text or patterns
100- Great for searching code bases for function names, variable declarations, or error messages
101- Useful for finding all files that use a particular API or pattern
102
103HOW TO USE:
104- Provide a regex pattern to search for within file contents
105- Set literal_text=true if you want to search for the exact text with special characters (recommended for non-regex users)
106- Optionally specify a starting directory (defaults to current working directory)
107- Optionally provide an include pattern to filter which files to search
108- Results are sorted with most recently modified files first
109
110REGEX PATTERN SYNTAX (when literal_text=false):
111- Supports standard regular expression syntax
112- 'function' searches for the literal text "function"
113- 'log\..*Error' finds text starting with "log." and ending with "Error"
114- 'import\s+.*\s+from' finds import statements in JavaScript/TypeScript
115
116COMMON INCLUDE PATTERN EXAMPLES:
117- '*.js' - Only search JavaScript files
118- '*.{ts,tsx}' - Only search TypeScript files
119- '*.go' - Only search Go files
120
121LIMITATIONS:
122- Results are limited to 100 files (newest first)
123- Performance depends on the number of files being searched
124- Very large binary files may be skipped
125- Hidden files (starting with '.') are skipped
126
127WINDOWS NOTES:
128- Uses ripgrep (rg) command if available for better performance
129- On Windows, install ripgrep via: winget install BurntSushi.ripgrep.MSVC
130- Falls back to built-in Go implementation if ripgrep is not available
131- File paths are normalized automatically for Windows compatibility
132
133TIPS:
134- For faster, more targeted searches, first use Glob to find relevant files, then use Grep
135- When doing iterative exploration that may require multiple rounds of searching, consider using the Agent tool instead
136- Always check if results are truncated and refine your search pattern if needed
137- Use literal_text=true when searching for exact text containing special characters like dots, parentheses, etc.`
138)
139
140func NewGrepTool() BaseTool {
141 return &grepTool{}
142}
143
144func (g *grepTool) Info() ToolInfo {
145 return ToolInfo{
146 Name: GrepToolName,
147 Description: grepDescription,
148 Parameters: map[string]any{
149 "pattern": map[string]any{
150 "type": "string",
151 "description": "The regex pattern to search for in file contents",
152 },
153 "path": map[string]any{
154 "type": "string",
155 "description": "The directory to search in. Defaults to the current working directory.",
156 },
157 "include": map[string]any{
158 "type": "string",
159 "description": "File pattern to include in the search (e.g. \"*.js\", \"*.{ts,tsx}\")",
160 },
161 "literal_text": map[string]any{
162 "type": "boolean",
163 "description": "If true, the pattern will be treated as literal text with special regex characters escaped. Default is false.",
164 },
165 },
166 Required: []string{"pattern"},
167 }
168}
169
170// escapeRegexPattern escapes special regex characters so they're treated as literal characters
171func escapeRegexPattern(pattern string) string {
172 specialChars := []string{"\\", ".", "+", "*", "?", "(", ")", "[", "]", "{", "}", "^", "$", "|"}
173 escaped := pattern
174
175 for _, char := range specialChars {
176 escaped = strings.ReplaceAll(escaped, char, "\\"+char)
177 }
178
179 return escaped
180}
181
182func (g *grepTool) Run(ctx context.Context, call ToolCall) (ToolResponse, error) {
183 var params GrepParams
184 if err := json.Unmarshal([]byte(call.Input), ¶ms); err != nil {
185 return NewTextErrorResponse(fmt.Sprintf("error parsing parameters: %s", err)), nil
186 }
187
188 if params.Pattern == "" {
189 return NewTextErrorResponse("pattern is required"), nil
190 }
191
192 // If literal_text is true, escape the pattern
193 searchPattern := params.Pattern
194 if params.LiteralText {
195 searchPattern = escapeRegexPattern(params.Pattern)
196 }
197
198 searchPath := params.Path
199 if searchPath == "" {
200 searchPath = config.WorkingDirectory()
201 }
202
203 matches, truncated, err := searchFiles(searchPattern, searchPath, params.Include, 100)
204 if err != nil {
205 return ToolResponse{}, fmt.Errorf("error searching files: %w", err)
206 }
207
208 var output strings.Builder
209 if len(matches) == 0 {
210 output.WriteString("No files found")
211 } else {
212 fmt.Fprintf(&output, "Found %d matches\n", len(matches))
213
214 currentFile := ""
215 for _, match := range matches {
216 if currentFile != match.path {
217 if currentFile != "" {
218 output.WriteString("\n")
219 }
220 currentFile = match.path
221 fmt.Fprintf(&output, "%s:\n", match.path)
222 }
223 if match.lineNum > 0 {
224 fmt.Fprintf(&output, " Line %d: %s\n", match.lineNum, match.lineText)
225 } else {
226 fmt.Fprintf(&output, " %s\n", match.path)
227 }
228 }
229
230 if truncated {
231 output.WriteString("\n(Results are truncated. Consider using a more specific path or pattern.)")
232 }
233 }
234
235 return WithResponseMetadata(
236 NewTextResponse(output.String()),
237 GrepResponseMetadata{
238 NumberOfMatches: len(matches),
239 Truncated: truncated,
240 },
241 ), nil
242}
243
244func searchFiles(pattern, rootPath, include string, limit int) ([]grepMatch, bool, error) {
245 matches, err := searchWithRipgrep(pattern, rootPath, include)
246 if err != nil {
247 matches, err = searchFilesWithRegex(pattern, rootPath, include)
248 if err != nil {
249 return nil, false, err
250 }
251 }
252
253 sort.Slice(matches, func(i, j int) bool {
254 return matches[i].modTime.After(matches[j].modTime)
255 })
256
257 truncated := len(matches) > limit
258 if truncated {
259 matches = matches[:limit]
260 }
261
262 return matches, truncated, nil
263}
264
265func searchWithRipgrep(pattern, path, include string) ([]grepMatch, error) {
266 cmd := fsext.GetRgSearchCmd(pattern, path, include)
267 if cmd == nil {
268 return nil, fmt.Errorf("ripgrep not found in $PATH")
269 }
270
271 output, err := cmd.Output()
272 if err != nil {
273 if exitErr, ok := err.(*exec.ExitError); ok && exitErr.ExitCode() == 1 {
274 return []grepMatch{}, nil
275 }
276 return nil, err
277 }
278
279 lines := strings.Split(strings.TrimSpace(string(output)), "\n")
280 matches := make([]grepMatch, 0, len(lines))
281
282 for _, line := range lines {
283 if line == "" {
284 continue
285 }
286
287 // Parse ripgrep output format: file:line:content
288 parts := strings.SplitN(line, ":", 3)
289 if len(parts) < 3 {
290 continue
291 }
292
293 filePath := parts[0]
294 lineNum, err := strconv.Atoi(parts[1])
295 if err != nil {
296 continue
297 }
298 lineText := parts[2]
299
300 fileInfo, err := os.Stat(filePath)
301 if err != nil {
302 continue // Skip files we can't access
303 }
304
305 matches = append(matches, grepMatch{
306 path: filePath,
307 modTime: fileInfo.ModTime(),
308 lineNum: lineNum,
309 lineText: lineText,
310 })
311 }
312
313 return matches, nil
314}
315
316func searchFilesWithRegex(pattern, rootPath, include string) ([]grepMatch, error) {
317 matches := []grepMatch{}
318
319 // Use cached regex compilation
320 regex, err := searchRegexCache.get(pattern)
321 if err != nil {
322 return nil, fmt.Errorf("invalid regex pattern: %w", err)
323 }
324
325 var includePattern *regexp.Regexp
326 if include != "" {
327 regexPattern := globToRegex(include)
328 includePattern, err = globRegexCache.get(regexPattern)
329 if err != nil {
330 return nil, fmt.Errorf("invalid include pattern: %w", err)
331 }
332 }
333
334 err = filepath.Walk(rootPath, func(path string, info os.FileInfo, err error) error {
335 if err != nil {
336 return nil // Skip errors
337 }
338
339 if info.IsDir() {
340 return nil // Skip directories
341 }
342
343 if fsext.SkipHidden(path) {
344 return nil
345 }
346
347 if includePattern != nil && !includePattern.MatchString(path) {
348 return nil
349 }
350
351 match, lineNum, lineText, err := fileContainsPattern(path, regex)
352 if err != nil {
353 return nil // Skip files we can't read
354 }
355
356 if match {
357 matches = append(matches, grepMatch{
358 path: path,
359 modTime: info.ModTime(),
360 lineNum: lineNum,
361 lineText: lineText,
362 })
363
364 if len(matches) >= 200 {
365 return filepath.SkipAll
366 }
367 }
368
369 return nil
370 })
371 if err != nil {
372 return nil, err
373 }
374
375 return matches, nil
376}
377
378func fileContainsPattern(filePath string, pattern *regexp.Regexp) (bool, int, string, error) {
379 // Quick binary file detection
380 if isBinaryFile(filePath) {
381 return false, 0, "", nil
382 }
383
384 file, err := os.Open(filePath)
385 if err != nil {
386 return false, 0, "", err
387 }
388 defer file.Close()
389
390 scanner := bufio.NewScanner(file)
391 lineNum := 0
392 for scanner.Scan() {
393 lineNum++
394 line := scanner.Text()
395 if pattern.MatchString(line) {
396 return true, lineNum, line, nil
397 }
398 }
399
400 return false, 0, "", scanner.Err()
401}
402
403var binaryExts = map[string]struct{}{
404 ".exe": {}, ".dll": {}, ".so": {}, ".dylib": {},
405 ".bin": {}, ".obj": {}, ".o": {}, ".a": {},
406 ".zip": {}, ".tar": {}, ".gz": {}, ".bz2": {},
407 ".jpg": {}, ".jpeg": {}, ".png": {}, ".gif": {},
408 ".pdf": {}, ".doc": {}, ".docx": {}, ".xls": {},
409 ".mp3": {}, ".mp4": {}, ".avi": {}, ".mov": {},
410}
411
412// isBinaryFile performs a quick check to determine if a file is binary
413func isBinaryFile(filePath string) bool {
414 // Check file extension first (fastest)
415 ext := strings.ToLower(filepath.Ext(filePath))
416 if _, isBinary := binaryExts[ext]; isBinary {
417 return true
418 }
419
420 // Quick content check for files without clear extensions
421 file, err := os.Open(filePath)
422 if err != nil {
423 return false // If we can't open it, let the caller handle the error
424 }
425 defer file.Close()
426
427 // Read first 512 bytes to check for null bytes
428 buffer := make([]byte, 512)
429 n, err := file.Read(buffer)
430 if err != nil && err != io.EOF {
431 return false
432 }
433
434 // Check for null bytes (common in binary files)
435 for i := range n {
436 if buffer[i] == 0 {
437 return true
438 }
439 }
440
441 return false
442}
443
444func globToRegex(glob string) string {
445 regexPattern := strings.ReplaceAll(glob, ".", "\\.")
446 regexPattern = strings.ReplaceAll(regexPattern, "*", ".*")
447 regexPattern = strings.ReplaceAll(regexPattern, "?", ".")
448
449 // Use pre-compiled regex instead of compiling each time
450 regexPattern = globBraceRegex.ReplaceAllStringFunc(regexPattern, func(match string) string {
451 inner := match[1 : len(match)-1]
452 return "(" + strings.ReplaceAll(inner, ",", "|") + ")"
453 })
454
455 return regexPattern
456}