doks/internal/search/text_search.go

package search

import (
	"bufio"
	"bytes"
	"os"
	"os/exec"
	"path/filepath"
	"strconv"
	"strings"

	"github.com/yeho/doks/internal/registry"
)

type TextSearcher struct {
	filesDir     string
	registry     *registry.Registry
	hasRipgrep   bool
	contextLines int
}

func NewTextSearcher(filesDir string, reg *registry.Registry) *TextSearcher {
	_, err := exec.LookPath("rg")
	return &TextSearcher{
		filesDir:   filesDir,
		registry:   reg,
		hasRipgrep: err == nil,
	}
}

func (t *TextSearcher) Search(query string, opts Options) ([]Result, error) {
	if t.hasRipgrep {
		return t.searchWithRipgrep(query, opts)
	}
	return t.searchPureGo(query, opts)
}

func (t *TextSearcher) searchWithRipgrep(query string, opts Options) ([]Result, error) {
	contextArg := strconv.Itoa(opts.ContextLines)
	cmd := exec.Command("rg",
		"--line-number",
		"--with-filename",
		"--context", contextArg,
		"--max-count", strconv.Itoa(opts.MaxResults*2), // Get more than needed, will dedupe
		"-i", // Case insensitive
		query,
		t.filesDir,
	)

	var stdout, stderr bytes.Buffer
	cmd.Stdout = &stdout
	cmd.Stderr = &stderr

	err := cmd.Run()
	if err != nil {
		// Exit code 1 means no matches, which is fine
		if exitErr, ok := err.(*exec.ExitError); ok && exitErr.ExitCode() == 1 {
			return []Result{}, nil
		}
		// Other errors might be okay too (e.g., some files couldn't be read)
		if stdout.Len() == 0 {
			return []Result{}, nil
		}
	}

	return t.parseRipgrepOutput(stdout.String(), opts)
}

func (t *TextSearcher) parseRipgrepOutput(output string, opts Options) ([]Result, error) {
	results := []Result{}
	seen := make(map[string]bool) // Track unique file:line combinations

	lines := strings.Split(output, "\n")
	for _, line := range lines {
		if line == "" || line == "--" {
			continue
		}

		// ripgrep output format: filename:linenum:text or filename-linenum-text (context)
		// We want the matching lines (with :)
		parts := strings.SplitN(line, ":", 3)
		if len(parts) < 3 {
			continue
		}

		filename := filepath.Base(parts[0])
		lineNum, err := strconv.Atoi(parts[1])
		if err != nil {
			continue
		}
		lineText := parts[2]

		// Find the key for this file
		key := t.findKeyByFilename(filename)
		if key == "" {
			continue
		}

		// Deduplicate
		dedupKey := filename + ":" + strconv.Itoa(lineNum)
		if seen[dedupKey] {
			continue
		}
		seen[dedupKey] = true

		results = append(results, Result{
			Key:        key,
			Filename:   filename,
			FilePath:   parts[0],
			LineNumber: lineNum,
			LineText:   strings.TrimSpace(lineText),
		})

		if len(results) >= opts.MaxResults {
			break
		}
	}

	return results, nil
}

func (t *TextSearcher) searchPureGo(query string, opts Options) ([]Result, error) {
	results := []Result{}
	queryLower := strings.ToLower(query)

	entries := t.registry.List()
	for _, entry := range entries {
		filePath := filepath.Join(t.filesDir, entry.Filename)
		file, err := os.Open(filePath)
		if err != nil {
			continue
		}

		scanner := bufio.NewScanner(file)
		lineNum := 0
		for scanner.Scan() {
			lineNum++
			line := scanner.Text()
			if strings.Contains(strings.ToLower(line), queryLower) {
				results = append(results, Result{
					Key:        entry.Key,
					Filename:   entry.Filename,
					FilePath:   filePath,
					LineNumber: lineNum,
					LineText:   strings.TrimSpace(line),
				})

				if len(results) >= opts.MaxResults {
					file.Close()
					return results, nil
				}
			}
		}
		file.Close()
	}

	return results, nil
}

func (t *TextSearcher) findKeyByFilename(filename string) string {
	entries := t.registry.List()
	for _, entry := range entries {
		if entry.Filename == filename {
			return entry.Key
		}
	}
	return ""
}