mirror of
https://github.com/Dvorinka/Devour.git
synced 2026-06-03 20:13:03 +00:00
215 lines
5.2 KiB
Go
215 lines
5.2 KiB
Go
package scraper
|
|
|
|
import (
|
|
"context"
|
|
"crypto/sha256"
|
|
"encoding/hex"
|
|
"fmt"
|
|
"io"
|
|
"net/http"
|
|
"strings"
|
|
"time"
|
|
|
|
"github.com/yourorg/devour/pkg/reactdocs"
|
|
)
|
|
|
|
type ReactDocsScraper struct {
|
|
config *Config
|
|
parser *reactdocs.Parser
|
|
client *http.Client
|
|
}
|
|
|
|
func NewReactDocsScraper(config *Config) *ReactDocsScraper {
|
|
return &ReactDocsScraper{
|
|
config: config,
|
|
parser: reactdocs.NewParser(),
|
|
client: &http.Client{
|
|
Timeout: config.Timeout,
|
|
},
|
|
}
|
|
}
|
|
|
|
func (s *ReactDocsScraper) Scrape(ctx context.Context, source *Source) ([]*Document, error) {
|
|
var documents []*Document
|
|
|
|
if source.URL == "" {
|
|
return nil, fmt.Errorf("URL is required for React docs scraper")
|
|
}
|
|
|
|
html, err := s.fetchPage(ctx, source.URL)
|
|
if err != nil {
|
|
return nil, fmt.Errorf("failed to fetch page: %w", err)
|
|
}
|
|
|
|
ref, err := s.parser.ParseReferencePage(html, source.URL)
|
|
if err != nil {
|
|
return nil, fmt.Errorf("failed to parse reference: %w", err)
|
|
}
|
|
|
|
mainDoc := s.referenceToDocument(ref, source.Name)
|
|
documents = append(documents, mainDoc)
|
|
|
|
for _, hook := range ref.Hooks {
|
|
doc := s.hookToDocument(hook, source.Name)
|
|
documents = append(documents, doc)
|
|
}
|
|
|
|
for _, comp := range ref.Components {
|
|
doc := s.componentToDocument(comp, source.Name)
|
|
documents = append(documents, doc)
|
|
}
|
|
|
|
for _, api := range ref.APIs {
|
|
doc := s.apiToDocument(api, source.Name)
|
|
documents = append(documents, doc)
|
|
}
|
|
|
|
return documents, nil
|
|
}
|
|
|
|
func (s *ReactDocsScraper) 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 *ReactDocsScraper) 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 *ReactDocsScraper) generateHash(content string) string {
|
|
hash := sha256.Sum256([]byte(content))
|
|
return hex.EncodeToString(hash[:])
|
|
}
|
|
|
|
func (s *ReactDocsScraper) referenceToDocument(ref *reactdocs.Reference, sourceName string) *Document {
|
|
var content strings.Builder
|
|
fmt.Fprintf(&content, "# React API Reference\n\n")
|
|
fmt.Fprintf(&content, "Hooks: %d, Components: %d, APIs: %d\n", len(ref.Hooks), len(ref.Components), len(ref.APIs))
|
|
|
|
return &Document{
|
|
ID: generateDocID(ref.DocURL),
|
|
Source: sourceName,
|
|
Type: "react-reference",
|
|
Title: "React API Reference",
|
|
Content: content.String(),
|
|
URL: ref.DocURL,
|
|
Metadata: map[string]interface{}{"doc_type": "react-reference"},
|
|
Hash: s.generateHash(content.String()),
|
|
Timestamp: time.Now(),
|
|
}
|
|
}
|
|
|
|
func (s *ReactDocsScraper) hookToDocument(hook *reactdocs.Hook, sourceName string) *Document {
|
|
var content strings.Builder
|
|
fmt.Fprintf(&content, "# %s\n\n", hook.Name)
|
|
if hook.Signature != "" {
|
|
fmt.Fprintf(&content, "```javascript\n%s\n```\n\n", hook.Signature)
|
|
}
|
|
fmt.Fprintf(&content, "%s\n", hook.Doc)
|
|
|
|
metadata := map[string]interface{}{
|
|
"name": hook.Name,
|
|
"category": hook.Category,
|
|
"doc_url": hook.DocURL,
|
|
"doc_type": "react-hook",
|
|
}
|
|
|
|
return &Document{
|
|
ID: generateDocID(hook.DocURL),
|
|
Source: sourceName,
|
|
Type: "react-hook",
|
|
Title: hook.Name,
|
|
Content: content.String(),
|
|
URL: hook.DocURL,
|
|
Metadata: metadata,
|
|
Hash: s.generateHash(content.String()),
|
|
Timestamp: time.Now(),
|
|
}
|
|
}
|
|
|
|
func (s *ReactDocsScraper) componentToDocument(comp *reactdocs.Component, sourceName string) *Document {
|
|
var content strings.Builder
|
|
fmt.Fprintf(&content, "# <%s />\n\n", comp.Name)
|
|
fmt.Fprintf(&content, "%s\n", comp.Doc)
|
|
|
|
if len(comp.Props) > 0 {
|
|
fmt.Fprintf(&content, "\n## Props\n")
|
|
for _, p := range comp.Props {
|
|
fmt.Fprintf(&content, "- `%s: %s`\n", p.Name, p.Type)
|
|
}
|
|
}
|
|
|
|
metadata := map[string]interface{}{
|
|
"name": comp.Name,
|
|
"doc_url": comp.DocURL,
|
|
"doc_type": "react-component",
|
|
}
|
|
|
|
return &Document{
|
|
ID: generateDocID(comp.DocURL),
|
|
Source: sourceName,
|
|
Type: "react-component",
|
|
Title: comp.Name,
|
|
Content: content.String(),
|
|
URL: comp.DocURL,
|
|
Metadata: metadata,
|
|
Hash: s.generateHash(content.String()),
|
|
Timestamp: time.Now(),
|
|
}
|
|
}
|
|
|
|
func (s *ReactDocsScraper) apiToDocument(api *reactdocs.API, sourceName string) *Document {
|
|
var content strings.Builder
|
|
fmt.Fprintf(&content, "# %s()\n\n", api.Name)
|
|
if api.Signature != "" {
|
|
fmt.Fprintf(&content, "```javascript\n%s\n```\n\n", api.Signature)
|
|
}
|
|
fmt.Fprintf(&content, "%s\n", api.Doc)
|
|
|
|
metadata := map[string]interface{}{
|
|
"name": api.Name,
|
|
"doc_url": api.DocURL,
|
|
"doc_type": "react-api",
|
|
}
|
|
|
|
return &Document{
|
|
ID: generateDocID(api.DocURL),
|
|
Source: sourceName,
|
|
Type: "react-api",
|
|
Title: api.Name,
|
|
Content: content.String(),
|
|
URL: api.DocURL,
|
|
Metadata: metadata,
|
|
Hash: s.generateHash(content.String()),
|
|
Timestamp: time.Now(),
|
|
}
|
|
}
|