miniflux-v2/internal/reader/processor/utils.go

// SPDX-FileCopyrightText: Copyright The Miniflux Authors. All rights reserved.
// SPDX-License-Identifier: Apache-2.0

package processor // import "miniflux.app/v2/internal/reader/processor"

import (
	"errors"
	"fmt"
	"regexp"
	"strconv"
	"time"

	"github.com/tdewolff/minify/v2"
	"github.com/tdewolff/minify/v2/html"
)

// TODO: use something less horrible than a regex to parse ISO 8601 durations.

var (
	iso8601Regex = regexp.MustCompile(`^P((?P<year>\d+)Y)?((?P<month>\d+)M)?((?P<week>\d+)W)?((?P<day>\d+)D)?(T((?P<hour>\d+)H)?((?P<minute>\d+)M)?((?P<second>\d+)S)?)?$`)
)

func parseISO8601(from string) (time.Duration, error) {
	var match []string
	var d time.Duration

	if iso8601Regex.MatchString(from) {
		match = iso8601Regex.FindStringSubmatch(from)
	} else {
		return 0, errors.New("youtube: could not parse duration string")
	}

	for i, name := range iso8601Regex.SubexpNames() {
		part := match[i]
		if i == 0 || name == "" || part == "" {
			continue
		}

		val, err := strconv.ParseInt(part, 10, 64)
		if err != nil {
			return 0, err
		}

		switch name {
		case "hour":
			d += time.Duration(val) * time.Hour
		case "minute":
			d += time.Duration(val) * time.Minute
		case "second":
			d += time.Duration(val) * time.Second
		default:
			return 0, fmt.Errorf("youtube: unknown field %s", name)
		}
	}

	return d, nil
}

func minifyContent(content string) string {
	m := minify.New()

	// Options required to avoid breaking the HTML content.
	m.Add("text/html", &html.Minifier{
		KeepEndTags: true,
		KeepQuotes:  true,
	})

	if minifiedHTML, err := m.String("text/html", content); err == nil {
		content = minifiedHTML
	}

	return content
}

func containsRegexPattern(pattern string, entries []string) bool {
	for _, entry := range entries {
		if matched, _ := regexp.MatchString(pattern, entry); matched {
			return true
		}
	}
	return false
}