refactor: split processor package into smaller files

internal/reader/processor/bilibili.go

@@ -0,0 +1,92 @@
+// SPDX-FileCopyrightText: Copyright The Miniflux Authors. All rights reserved.
+// SPDX-License-Identifier: Apache-2.0
+
+package processor
+
+import (
+	"encoding/json"
+	"fmt"
+	"log/slog"
+	"regexp"
+
+	"miniflux.app/v2/internal/config"
+	"miniflux.app/v2/internal/model"
+	"miniflux.app/v2/internal/reader/fetcher"
+)
+
+var (
+	bilibiliURLRegex     = regexp.MustCompile(`bilibili\.com/video/(.*)$`)
+	bilibiliVideoIdRegex = regexp.MustCompile(`/video/(?:av(\d+)|BV([a-zA-Z0-9]+))`)
+)
+
+func shouldFetchBilibiliWatchTime(entry *model.Entry) bool {
+	if !config.Opts.FetchBilibiliWatchTime() {
+		return false
+	}
+	matches := bilibiliURLRegex.FindStringSubmatch(entry.URL)
+	urlMatchesBilibiliPattern := len(matches) == 2
+	return urlMatchesBilibiliPattern
+}
+
+func extractBilibiliVideoID(websiteURL string) (string, string, error) {
+	matches := bilibiliVideoIdRegex.FindStringSubmatch(websiteURL)
+	if matches == nil {
+		return "", "", fmt.Errorf("no video ID found in URL: %s", websiteURL)
+	}
+	if matches[1] != "" {
+		return "aid", matches[1], nil
+	}
+	if matches[2] != "" {
+		return "bvid", matches[2], nil
+	}
+	return "", "", fmt.Errorf("unexpected regex match result for URL: %s", websiteURL)
+}
+
+func fetchBilibiliWatchTime(websiteURL string) (int, error) {
+	requestBuilder := fetcher.NewRequestBuilder()
+	requestBuilder.WithTimeout(config.Opts.HTTPClientTimeout())
+	requestBuilder.WithProxy(config.Opts.HTTPClientProxy())
+
+	idType, videoID, extractErr := extractBilibiliVideoID(websiteURL)
+	if extractErr != nil {
+		return 0, extractErr
+	}
+	bilibiliApiURL := fmt.Sprintf("https://api.bilibili.com/x/web-interface/view?%s=%s", idType, videoID)
+
+	responseHandler := fetcher.NewResponseHandler(requestBuilder.ExecuteRequest(bilibiliApiURL))
+	defer responseHandler.Close()
+
+	if localizedError := responseHandler.LocalizedError(); localizedError != nil {
+		slog.Warn("Unable to fetch Bilibili API",
+			slog.String("website_url", websiteURL),
+			slog.String("api_url", bilibiliApiURL),
+			slog.Any("error", localizedError.Error()))
+		return 0, localizedError.Error()
+	}
+
+	var result map[string]interface{}
+	doc := json.NewDecoder(responseHandler.Body(config.Opts.HTTPClientMaxBodySize()))
+	if docErr := doc.Decode(&result); docErr != nil {
+		return 0, fmt.Errorf("failed to decode API response: %v", docErr)
+	}
+
+	if code, ok := result["code"].(float64); !ok || code != 0 {
+		return 0, fmt.Errorf("API returned error code: %v", result["code"])
+	}
+
+	data, ok := result["data"].(map[string]interface{})
+	if !ok {
+		return 0, fmt.Errorf("data field not found or not an object")
+	}
+
+	duration, ok := data["duration"].(float64)
+	if !ok {
+		return 0, fmt.Errorf("duration not found or not a number")
+	}
+	intDuration := int(duration)
+	durationMin := intDuration / 60
+	if intDuration%60 != 0 {
+		durationMin++
+	}
+	return durationMin, nil
+}

internal/reader/processor/nebula.go

@@ -0,0 +1,60 @@
+// SPDX-FileCopyrightText: Copyright The Miniflux Authors. All rights reserved.
+// SPDX-License-Identifier: Apache-2.0
+
+package processor
+
+import (
+	"errors"
+	"fmt"
+	"log/slog"
+	"regexp"
+	"strconv"
+
+	"github.com/PuerkitoBio/goquery"
+
+	"miniflux.app/v2/internal/config"
+	"miniflux.app/v2/internal/model"
+	"miniflux.app/v2/internal/reader/fetcher"
+)
+
+var nebulaRegex = regexp.MustCompile(`^https://nebula\.tv`)
+
+func shouldFetchNebulaWatchTime(entry *model.Entry) bool {
+	if !config.Opts.FetchNebulaWatchTime() {
+		return false
+	}
+	matches := nebulaRegex.FindStringSubmatch(entry.URL)
+	return matches != nil
+}
+
+func fetchNebulaWatchTime(websiteURL string) (int, error) {
+	requestBuilder := fetcher.NewRequestBuilder()
+	requestBuilder.WithTimeout(config.Opts.HTTPClientTimeout())
+	requestBuilder.WithProxy(config.Opts.HTTPClientProxy())
+
+	responseHandler := fetcher.NewResponseHandler(requestBuilder.ExecuteRequest(websiteURL))
+	defer responseHandler.Close()
+
+	if localizedError := responseHandler.LocalizedError(); localizedError != nil {
+		slog.Warn("Unable to fetch Nebula watch time", slog.String("website_url", websiteURL), slog.Any("error", localizedError.Error()))
+		return 0, localizedError.Error()
+	}
+
+	doc, docErr := goquery.NewDocumentFromReader(responseHandler.Body(config.Opts.HTTPClientMaxBodySize()))
+	if docErr != nil {
+		return 0, docErr
+	}
+
+	durs, exists := doc.Find(`meta[property="video:duration"]`).First().Attr("content")
+	// durs contains video watch time in seconds
+	if !exists {
+		return 0, errors.New("duration has not found")
+	}
+
+	dur, err := strconv.ParseInt(durs, 10, 64)
+	if err != nil {
+		return 0, fmt.Errorf("unable to parse duration %s: %v", durs, err)
+	}
+
+	return int(dur / 60), nil
+}

internal/reader/processor/odysee.go

@@ -0,0 +1,60 @@
+// SPDX-FileCopyrightText: Copyright The Miniflux Authors. All rights reserved.
+// SPDX-License-Identifier: Apache-2.0
+
+package processor
+
+import (
+	"errors"
+	"fmt"
+	"log/slog"
+	"regexp"
+	"strconv"
+
+	"github.com/PuerkitoBio/goquery"
+
+	"miniflux.app/v2/internal/config"
+	"miniflux.app/v2/internal/model"
+	"miniflux.app/v2/internal/reader/fetcher"
+)
+
+var odyseeRegex = regexp.MustCompile(`^https://odysee\.com`)
+
+func shouldFetchOdyseeWatchTime(entry *model.Entry) bool {
+	if !config.Opts.FetchOdyseeWatchTime() {
+		return false
+	}
+	matches := odyseeRegex.FindStringSubmatch(entry.URL)
+	return matches != nil
+}
+
+func fetchOdyseeWatchTime(websiteURL string) (int, error) {
+	requestBuilder := fetcher.NewRequestBuilder()
+	requestBuilder.WithTimeout(config.Opts.HTTPClientTimeout())
+	requestBuilder.WithProxy(config.Opts.HTTPClientProxy())
+
+	responseHandler := fetcher.NewResponseHandler(requestBuilder.ExecuteRequest(websiteURL))
+	defer responseHandler.Close()
+
+	if localizedError := responseHandler.LocalizedError(); localizedError != nil {
+		slog.Warn("Unable to fetch Odysee watch time", slog.String("website_url", websiteURL), slog.Any("error", localizedError.Error()))
+		return 0, localizedError.Error()
+	}
+
+	doc, docErr := goquery.NewDocumentFromReader(responseHandler.Body(config.Opts.HTTPClientMaxBodySize()))
+	if docErr != nil {
+		return 0, docErr
+	}
+
+	durs, exists := doc.Find(`meta[property="og:video:duration"]`).First().Attr("content")
+	// durs contains video watch time in seconds
+	if !exists {
+		return 0, errors.New("duration has not found")
+	}
+
+	dur, err := strconv.ParseInt(durs, 10, 64)
+	if err != nil {
+		return 0, fmt.Errorf("unable to parse duration %s: %v", durs, err)
+	}
+
+	return int(dur / 60), nil
+}