This commit is contained in:
Tomas Dvorak
2026-02-22 15:41:27 +01:00
parent 0b88627e54
commit 409acd2e08
84 changed files with 65382 additions and 27475 deletions
+221
View File
@@ -0,0 +1,221 @@
package scraper
import (
"context"
"crypto/sha256"
"encoding/hex"
"fmt"
"io"
"net/http"
"strings"
"time"
"github.com/yourorg/devour/pkg/springdocs"
)
type SpringDocsScraper struct {
config *Config
parser *springdocs.Parser
client *http.Client
}
func NewSpringDocsScraper(config *Config) *SpringDocsScraper {
return &SpringDocsScraper{
config: config,
parser: springdocs.NewParser(),
client: &http.Client{
Timeout: config.Timeout,
},
}
}
func (s *SpringDocsScraper) Scrape(ctx context.Context, source *Source) ([]*Document, error) {
var documents []*Document
if source.URL == "" {
return nil, fmt.Errorf("URL is required for Spring docs scraper")
}
html, err := s.fetchPage(ctx, source.URL)
if err != nil {
return nil, fmt.Errorf("failed to fetch page: %w", err)
}
module, err := s.parser.ParseModulePage(html, source.URL)
if err != nil {
return nil, fmt.Errorf("failed to parse module: %w", err)
}
mainDoc := s.moduleToDocument(module, source.Name)
documents = append(documents, mainDoc)
for _, class := range module.Classes {
doc := s.classToDocument(class, module, source.Name)
documents = append(documents, doc)
}
for _, prop := range module.Properties {
doc := s.propertyToDocument(prop, source.Name)
documents = append(documents, doc)
}
for _, guide := range module.Guides {
doc := s.guideToDocument(guide, source.Name)
documents = append(documents, doc)
}
return documents, nil
}
func (s *SpringDocsScraper) 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 *SpringDocsScraper) 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 *SpringDocsScraper) generateHash(content string) string {
hash := sha256.Sum256([]byte(content))
return hex.EncodeToString(hash[:])
}
func (s *SpringDocsScraper) moduleToDocument(module *springdocs.Module, sourceName string) *Document {
var content strings.Builder
fmt.Fprintf(&content, "# %s\n\n", module.Name)
fmt.Fprintf(&content, "%s\n", module.Doc)
metadata := map[string]interface{}{
"module": module.Name,
"version": module.Version,
"doc_url": module.DocURL,
"doc_type": "spring-module",
}
return &Document{
ID: generateDocID(module.DocURL),
Source: sourceName,
Type: "spring-module",
Title: module.Name,
Content: content.String(),
URL: module.DocURL,
Metadata: metadata,
Hash: s.generateHash(content.String()),
Timestamp: time.Now(),
}
}
func (s *SpringDocsScraper) classToDocument(class *springdocs.Class, module *springdocs.Module, sourceName string) *Document {
var content strings.Builder
fmt.Fprintf(&content, "# %s\n\n", class.QualifiedName)
fmt.Fprintf(&content, "%s\n", class.Doc)
if len(class.Methods) > 0 {
fmt.Fprintf(&content, "\n## Methods\n")
for _, m := range class.Methods {
fmt.Fprintf(&content, "- `%s`\n", m.Signature)
}
}
metadata := map[string]interface{}{
"module": module.Name,
"qualified_name": class.QualifiedName,
"kind": class.Kind,
"doc_url": class.DocURL,
}
return &Document{
ID: generateDocID(class.DocURL),
Source: sourceName,
Type: "spring-class",
Title: class.QualifiedName,
Content: content.String(),
URL: class.DocURL,
Metadata: metadata,
Hash: s.generateHash(content.String()),
Timestamp: time.Now(),
}
}
func (s *SpringDocsScraper) propertyToDocument(prop *springdocs.Property, sourceName string) *Document {
var content strings.Builder
fmt.Fprintf(&content, "# %s\n\n", prop.Name)
fmt.Fprintf(&content, "Type: %s\n", prop.Type)
if prop.Default != "" {
fmt.Fprintf(&content, "Default: `%s`\n", prop.Default)
}
fmt.Fprintf(&content, "\n%s\n", prop.Doc)
metadata := map[string]interface{}{
"property": prop.Name,
"type": prop.Type,
"default": prop.Default,
"doc_url": prop.DocURL,
}
return &Document{
ID: generateDocID(prop.Name),
Source: sourceName,
Type: "spring-property",
Title: prop.Name,
Content: content.String(),
URL: prop.DocURL,
Metadata: metadata,
Hash: s.generateHash(content.String()),
Timestamp: time.Now(),
}
}
func (s *SpringDocsScraper) guideToDocument(guide *springdocs.Guide, sourceName string) *Document {
var content strings.Builder
fmt.Fprintf(&content, "# %s\n\n", guide.Title)
fmt.Fprintf(&content, "%s\n", guide.Description)
metadata := map[string]interface{}{
"title": guide.Title,
"doc_url": guide.DocURL,
"level": guide.Level,
"doc_type": "spring-guide",
}
return &Document{
ID: generateDocID(guide.DocURL),
Source: sourceName,
Type: "spring-guide",
Title: guide.Title,
Content: content.String(),
URL: guide.DocURL,
Metadata: metadata,
Hash: s.generateHash(content.String()),
Timestamp: time.Now(),
}
}