Devour/internal/scraper/local.go

package scraper

import (
	"context"
	"crypto/sha256"
	"encoding/hex"
	"fmt"
	"io/fs"
	"os"
	"path/filepath"
	"regexp"
	"strings"
	"time"
)

var (
	reLocalBlankLines  = regexp.MustCompile(`\n{3,}`)
	reFrontMatterBlock = regexp.MustCompile(`(?s)\A---\s*\n.*?\n---\s*\n`)
)

// LocalScraper scrapes documentation from local filesystem.
type LocalScraper struct {
	config *Config
}

// NewLocalScraper creates a new local scraper.
func NewLocalScraper(config *Config) *LocalScraper {
	return &LocalScraper{config: config}
}

// Scrape scans and parses documents from a local directory.
func (s *LocalScraper) Scrape(ctx context.Context, source *Source) ([]*Document, error) {
	if source == nil {
		return nil, fmt.Errorf("source is required")
	}

	root := strings.TrimSpace(source.Path)
	if root == "" {
		root = strings.TrimSpace(source.URL)
	}
	if root == "" {
		return nil, fmt.Errorf("path or url is required for local source")
	}

	info, err := os.Stat(root)
	if err != nil {
		return nil, err
	}

	docs := make([]*Document, 0)
	if !info.IsDir() {
		doc, err := s.fileToDocument(root, source)
		if err != nil {
			return nil, err
		}
		return []*Document{doc}, nil
	}

	web := NewWebScraper(s.config)
	err = filepath.WalkDir(root, func(path string, d fs.DirEntry, err error) error {
		if err != nil {
			return err
		}
		select {
		case <-ctx.Done():
			return ctx.Err()
		default:
		}

		if d.IsDir() {
			name := d.Name()
			if strings.HasPrefix(name, ".") || name == "node_modules" || name == "vendor" || name == "dist" || name == "build" {
				return filepath.SkipDir
			}
			return nil
		}

		relPath := path
		if rel, relErr := filepath.Rel(root, path); relErr == nil {
			relPath = rel
		}
		normalized := filepath.ToSlash(relPath)
		if !web.shouldInclude(normalized, source.Include, source.Exclude) {
			return nil
		}
		if !isDocumentationFile(path) {
			return nil
		}

		doc, err := s.fileToDocument(path, source)
		if err != nil {
			return nil
		}
		docs = append(docs, doc)
		return nil
	})
	if err != nil {
		return nil, err
	}

	return docs, nil
}

// DetectChanges checks if files have been modified.
func (s *LocalScraper) DetectChanges(ctx context.Context, source *Source, lastHash string) (bool, string, error) {
	if source == nil {
		return false, "", fmt.Errorf("source is required")
	}

	root := strings.TrimSpace(source.Path)
	if root == "" {
		root = strings.TrimSpace(source.URL)
	}
	if root == "" {
		return false, "", fmt.Errorf("path or url is required for local source")
	}

	h := sha256.New()
	err := filepath.WalkDir(root, func(path string, d fs.DirEntry, err error) error {
		if err != nil {
			return err
		}
		if d.IsDir() {
			name := d.Name()
			if strings.HasPrefix(name, ".") || name == "node_modules" || name == "vendor" || name == "dist" || name == "build" {
				return filepath.SkipDir
			}
			return nil
		}
		if !isDocumentationFile(path) {
			return nil
		}

		info, infoErr := d.Info()
		if infoErr != nil {
			return infoErr
		}
		fmt.Fprintf(h, "%s|%d|%d\n", path, info.Size(), info.ModTime().UnixNano())
		return nil
	})
	if err != nil {
		return false, "", err
	}

	hash := hex.EncodeToString(h.Sum(nil))
	return hash != lastHash, hash, nil
}

func (s *LocalScraper) fileToDocument(path string, source *Source) (*Document, error) {
	b, err := os.ReadFile(path)
	if err != nil {
		return nil, err
	}

	ext := strings.ToLower(filepath.Ext(path))
	content := normalizeLocalContent(string(b), ext)
	if content == "" {
		return nil, fmt.Errorf("empty file")
	}

	title := strings.TrimSuffix(filepath.Base(path), filepath.Ext(path))
	hash := sha256.Sum256(b)
	uri := "file://" + filepath.ToSlash(path)

	docType := "local-document"
	switch ext {
	case ".md", ".mdx":
		docType = "local-markdown"
	case ".txt":
		docType = "local-text"
	case ".json", ".yaml", ".yml":
		docType = "local-data"
	case ".go", ".js", ".ts", ".tsx", ".py", ".java", ".rs", ".rb", ".php":
		docType = "local-code"
	}

	name := source.Name
	if strings.TrimSpace(name) == "" {
		name = filepath.Base(filepath.Dir(path))
	}

	return &Document{
		ID:      generateDocID(uri),
		Source:  name,
		Type:    docType,
		Title:   title,
		Content: content,
		URL:     uri,
		Metadata: map[string]interface{}{
			"path": path,
			"size": len(b),
		},
		Hash:      hex.EncodeToString(hash[:]),
		Timestamp: time.Now(),
	}, nil
}

func normalizeLocalContent(content, ext string) string {
	content = strings.TrimSpace(content)
	if content == "" {
		return ""
	}

	switch ext {
	case ".md", ".mdx":
		content = stripMarkdownFrontmatter(content)
		content = stripMDXPreamble(content)
	}

	// Collapse excessive blank lines to reduce indexing noise.
	content = reLocalBlankLines.ReplaceAllString(content, "\n\n")
	return strings.TrimSpace(content)
}

func stripMarkdownFrontmatter(content string) string {
	if !strings.HasPrefix(content, "---\n") && !strings.HasPrefix(content, "---\r\n") {
		return content
	}

	trimmed := reFrontMatterBlock.ReplaceAllString(content, "")
	return trimmed
}

func stripMDXPreamble(content string) string {
	lines := strings.Split(content, "\n")
	i := 0
	for i < len(lines) {
		line := strings.TrimSpace(lines[i])
		if line == "" {
			i++
			continue
		}
		if strings.HasPrefix(line, "import ") || strings.HasPrefix(line, "export ") {
			i++
			continue
		}
		break
	}
	return strings.Join(lines[i:], "\n")
}

func isDocumentationFile(path string) bool {
	ext := strings.ToLower(filepath.Ext(path))
	switch ext {
	case ".md", ".mdx", ".txt", ".rst", ".adoc", ".json", ".yaml", ".yml", ".go", ".py", ".js", ".ts", ".tsx", ".java", ".rs", ".rb", ".php", ".html":
		return true
	default:
		return false
	}
}