updage

internal/scraper/external/astrodocs.go

@@ -0,0 +1,156 @@
+package scraper
+
+import (
+	"context"
+	"crypto/sha256"
+	"encoding/hex"
+	"fmt"
+	"io"
+	"net/http"
+	"strings"
+	"time"
+
+	"github.com/yourorg/devour/pkg/astrodocs"
+)
+
+type AstroDocsScraper struct {
+	config *Config
+	parser *astrodocs.Parser
+	client *http.Client
+}
+
+func NewAstroDocsScraper(config *Config) *AstroDocsScraper {
+	return &AstroDocsScraper{
+		config: config,
+		parser: astrodocs.NewParser(),
+		client: &http.Client{
+			Timeout: config.Timeout,
+		},
+	}
+}
+
+func (s *AstroDocsScraper) Scrape(ctx context.Context, source *Source) ([]*Document, error) {
+	var documents []*Document
+
+	if source.URL == "" {
+		return nil, fmt.Errorf("URL is required for Astro docs scraper")
+	}
+
+	html, err := s.fetchPage(ctx, source.URL)
+	if err != nil {
+		return nil, fmt.Errorf("failed to fetch page: %w", err)
+	}
+
+	page, err := s.parser.ParsePage(html, source.URL)
+	if err != nil {
+		return nil, fmt.Errorf("failed to parse Astro docs page: %w", err)
+	}
+
+	mainDoc := s.pageToDocument(page, source.Name)
+	documents = append(documents, mainDoc)
+
+	for _, section := range page.Sections {
+		doc := s.sectionToDocument(section, page, source.Name)
+		documents = append(documents, doc)
+	}
+
+	return documents, nil
+}
+
+func (s *AstroDocsScraper) DetectChanges(ctx context.Context, source *Source, lastHash string) (bool, string, error) {
+	html, err := s.fetchPage(ctx, source.URL)
+	if err != nil {
+		return false, "", err
+	}
+
+	hash := s.generateHash(html)
+	changed := hash != lastHash
+
+	return changed, hash, nil
+}
+
+func (s *AstroDocsScraper) fetchPage(ctx context.Context, url string) (string, error) {
+	req, err := http.NewRequestWithContext(ctx, "GET", url, nil)
+	if err != nil {
+		return "", err
+	}
+
+	req.Header.Set("User-Agent", s.config.UserAgent)
+
+	resp, err := s.client.Do(req)
+	if err != nil {
+		return "", err
+	}
+	defer resp.Body.Close()
+
+	if resp.StatusCode != http.StatusOK {
+		return "", fmt.Errorf("HTTP %d", resp.StatusCode)
+	}
+
+	body, err := io.ReadAll(resp.Body)
+	if err != nil {
+		return "", err
+	}
+	return string(body), nil
+}
+
+func (s *AstroDocsScraper) generateHash(content string) string {
+	hash := sha256.Sum256([]byte(content))
+	return hex.EncodeToString(hash[:])
+}
+
+func (s *AstroDocsScraper) pageToDocument(page *astrodocs.Page, sourceName string) *Document {
+	var content strings.Builder
+	fmt.Fprintf(&content, "# %s\n\n", page.Title)
+	fmt.Fprintf(&content, "%s\n", page.Description)
+
+	if len(page.CodeBlocks) > 0 {
+		fmt.Fprintf(&content, "\n## Code Examples\n")
+		for _, cb := range page.CodeBlocks {
+			fmt.Fprintf(&content, "\n```%s\n%s\n```\n", cb.Language, cb.Code)
+		}
+	}
+
+	metadata := map[string]interface{}{
+		"title":    page.Title,
+		"doc_url":  page.URL,
+		"doc_type": "astro-docs",
+	}
+
+	return &Document{
+		ID:        generateDocID(page.URL),
+		Source:    sourceName,
+		Type:      "astro-docs",
+		Title:     page.Title,
+		Content:   content.String(),
+		URL:       page.URL,
+		Metadata:  metadata,
+		Hash:      s.generateHash(content.String()),
+		Timestamp: time.Now(),
+	}
+}
+
+func (s *AstroDocsScraper) sectionToDocument(section *astrodocs.Section, page *astrodocs.Page, sourceName string) *Document {
+	var content strings.Builder
+	fmt.Fprintf(&content, "# %s\n\n", section.Title)
+	fmt.Fprintf(&content, "%s\n", section.Content)
+
+	metadata := map[string]interface{}{
+		"page_title": page.Title,
+		"section_id": section.ID,
+		"doc_url":    section.DocURL,
+		"doc_type":   "astro-section",
+	}
+
+	return &Document{
+		ID:        generateDocID(section.DocURL),
+		Source:    sourceName,
+		Type:      "astro-section",
+		Title:     fmt.Sprintf("%s - %s", page.Title, section.Title),
+		Content:   content.String(),
+		URL:       section.DocURL,
+		Metadata:  metadata,
+		Hash:      s.generateHash(content.String()),
+		Timestamp: time.Now(),
+	}
+}