// Package urlutil provides helpers for pulling URLs out of arbitrary text.
package urlutil

import (
	"net/url"
	"regexp"
)

// urlPattern matches http(s) URLs embedded in free-form text.
var urlPattern = regexp.MustCompile(`https?://[^\s]+`)

// ExtractURLs returns the http(s) URLs found in text, in order of appearance
// and without duplicates. Trailing punctuation that is unlikely to be part of
// a URL (e.g. a sentence-ending period or a wrapping parenthesis) is trimmed,
// and each candidate is validated with url.Parse.
func ExtractURLs(text string) []string {
	matches := urlPattern.FindAllString(text, -1)

	seen := make(map[string]struct{}, len(matches))
	urls := make([]string, 0, len(matches))

	for _, match := range matches {
		candidate := trimURL(match)

		parsed, err := url.Parse(candidate)
		if err != nil || parsed.Host == "" {
			continue
		}

		if _, ok := seen[candidate]; ok {
			continue
		}
		seen[candidate] = struct{}{}
		urls = append(urls, candidate)
	}

	return urls
}

// trimURL strips common trailing characters that are almost never part of the
// URL itself when it appears inside a sentence.
func trimURL(s string) string {
	for len(s) > 0 {
		last := s[len(s)-1]
		switch last {
		case '.', ',', ';', ':', '!', '?', ')', ']', '}', '"', '\'', '>':
			s = s[:len(s)-1]
		default:
			return s
		}
	}
	return s
}