Files
gochat/backend/internal/service/captain_document_page_parser_backend.go

67 lines
2.0 KiB
Go

package service
import (
"context"
"fmt"
"io"
"net/http"
"strings"
"time"
)
// captainDocumentPageParserBackendImpl is the production implementation of
// CaptainDocumentPageParserBackend. It fetches a single web page and
// extracts its title and visible text content — the same logic used by
// the sync backend for individual page URLs.
type captainDocumentPageParserBackendImpl struct {
httpClient *http.Client
}
// NewCaptainDocumentPageParserBackend creates the production page parser backend.
func NewCaptainDocumentPageParserBackend() CaptainDocumentPageParserBackend {
return &captainDocumentPageParserBackendImpl{
httpClient: &http.Client{Timeout: 30 * time.Second},
}
}
func (b *captainDocumentPageParserBackendImpl) ParseCaptainDocumentPage(ctx context.Context, pageLink string) (*CaptainDocumentSyncResult, error) {
if pageLink == "" {
return &CaptainDocumentSyncResult{ErrorCode: "not_found"}, nil
}
req, err := http.NewRequestWithContext(ctx, http.MethodGet, pageLink, nil)
if err != nil {
return &CaptainDocumentSyncResult{ErrorCode: "fetch_failed"}, fmt.Errorf("create request: %w", err)
}
req.Header.Set("User-Agent", "GoChat-Captain/1.0")
resp, err := b.httpClient.Do(req)
if err != nil {
return &CaptainDocumentSyncResult{ErrorCode: "fetch_failed"}, fmt.Errorf("fetch url: %w", err)
}
defer resp.Body.Close()
if resp.StatusCode != http.StatusOK {
return &CaptainDocumentSyncResult{ErrorCode: "fetch_failed"}, fmt.Errorf("HTTP %d for %s", resp.StatusCode, pageLink)
}
body, err := io.ReadAll(io.LimitReader(resp.Body, 5<<20)) // 5MB max
if err != nil {
return &CaptainDocumentSyncResult{ErrorCode: "fetch_failed"}, fmt.Errorf("read body: %w", err)
}
title, content := extractHTMLText(string(body))
content = strings.TrimSpace(content)
if content == "" {
return &CaptainDocumentSyncResult{ErrorCode: "content_empty"}, nil
}
if title == "" {
title = pageLink
}
return &CaptainDocumentSyncResult{
Content: content,
Title: title,
}, nil
}