path: root/src/main/java/com/juick/util/MessageUtils.java

/*
 * Copyright (C) 2008-2024, Juick
 *
 * This program is free software: you can redistribute it and/or modify
 * it under the terms of the GNU Affero General Public License as
 * published by the Free Software Foundation, either version 3 of the
 * License, or (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU Affero General Public License for more details.
 *
 * You should have received a copy of the GNU Affero General Public License
 * along with this program.  If not, see <http://www.gnu.org/licenses/>.
 */

package com.juick.util;

import com.juick.model.Entity;
import com.juick.model.Message;
import com.juick.model.Tag;
import com.juick.model.User;
import org.apache.commons.lang3.StringUtils;
import org.springframework.web.util.UriComponentsBuilder;

import java.net.URI;
import java.net.URLEncoder;
import java.nio.charset.StandardCharsets;
import java.util.*;
import java.util.function.Function;
import java.util.regex.Matcher;
import java.util.regex.Pattern;
import java.util.stream.Collectors;

/**
 * Created by aalexeev on 11/13/16.
 */
public class MessageUtils {
    private MessageUtils() {
        throw new IllegalStateException();
    }

    public static String formatQuote(final String quote) {
        String result = quote;

        if (quote != null) {
            if (quote.length() > 50) {
                result = ">" + StringUtils.abbreviate(quote, "…", 47).replace('\n', ' ') + "\n";
            } else if (!quote.isEmpty()) {
                result = ">" + quote.replace('\n', ' ') + "\n";
            }
        }

        return result;
    }

    private final static String urlWhiteSpacePrefix = "((?<=\\s)|(?<=\\A))";

    private final static String urlRegex = "((?:(?:ht|f)tps?://(?:www\\.)?([^\\s()<>/?#]+)([^\\s()<>«»]*)?))";

    private final static String urlWithWhitespacesRegex =
            urlWhiteSpacePrefix + urlRegex;

    private final static Pattern regexLinks2 = Pattern.compile("((?<=\\s)|(?<=\\A))([\\[{])((?:ht|f)tps?://(?:www\\.)?([^/\\s\")!&gt;]*)/?(?:[^]}](?<!&gt;))*)([]}])");

    private final static String replyNumberRegex = "((?<=\\s)|(?<=\\A))\\/(\\d+)((?=\\s)|(?=\\Z)|(?=\\p{Punct}))";

    private final static String usernameRegex = "((?<=\\s)|(?<=\\A))@([\\w\\-]{2,16})((?=\\s)|(?=\\Z)|(?=\\p{Punct}))";
    private final static Pattern usernamePattern = Pattern.compile(usernameRegex);

    private final static String jidRegex = "((?<=\\s)|(?<=\\A))@([\\w\\-\\.]+@[\\w\\-\\.]+)((?=\\s)|(?=\\Z)|(?=\\p{Punct}))";
    private final static Pattern jidPattern = Pattern.compile(jidRegex);

    private final static String midRegex = "((?<=\\s)|(?<=\\A)|(?<=\\p{Punct}))#(\\d+)((?=\\s)|(?=\\Z)|(?=\\))|(?=\\.)|(?=\\,))";

    private final static String ridRegex = "((?<=\\s)|(?<=\\A)|(?<=\\p{Punct}))#(\\d+)/(\\d+)((?=\\s)|(?=\\Z)|(?=\\p{Punct}))";

    private final static String textUrlRegex = "\\[([^]]+)]\\[((?:ht|f)tps?://[^]]+)]";

    private final static String textUrlRegex2 = "\\[([^]]+)]\\(((?:ht|f)tps?://[^)]+)\\)";
    private final static String pidginUrlRegex = "(\\S+)\\s&lt;((?:ht|f)tps?://[^&]+)&gt;";
    private final static String boldRegex = "((?<=\\s)|(?<=\\A)|(?<=\\p{Punct}))\\*([^\\*\\n<>]+)\\*((?=\\s)|(?=\\Z)|(?=\\p{Punct}))";

    private final static String italicRegex = "((?<=\\s)|(?<=\\A))/([^\\/\\n<>]+)/((?=\\s)|(?=\\Z)|(?=\\p{Punct}))";

    private final static String underlineRegex = "((?<=\\s)|(?<=\\A))_([^\\_\\n<>]+)_((?=\\s)|(?=\\Z)|(?=\\p{Punct}))";

    private final static String citateRegex = "(?:(?<=\\n)|(?<=\\A))(?:&gt;|>) *(.*)?(\\r?\\n|(?=\\Z))";

    public static List<Entity> getEntities(Message msg) {

        String txt = msg.getText();

        // http://juick.com/last?page=2
        List<Entity> result = new ArrayList<>(entitiesForType("a", txt, urlWithWhitespacesRegex, matcher -> matcher.group(3), matcher -> Optional.of(matcher.group(2)), 0));
        // [link text][http://juick.com/last?page=2]
        result.addAll(entitiesForType("a", txt, textUrlRegex, matcher -> matcher.group(1), matcher -> Optional.of(matcher.group(2)), 0));
        // [link text](http://juick.com/last?page=2)
        result.addAll(entitiesForType("a", txt, textUrlRegex2, matcher -> matcher.group(1), matcher -> Optional.of(matcher.group(2)), 0));
        // #12345
        result.addAll(entitiesForType("a", txt, midRegex, matcher -> String.format("#%s", matcher.group(2)), matcher -> Optional.of("https://juick.com/m/" + matcher.group(2)), 0));
        // #12345/65
        result.addAll(entitiesForType("a", txt, ridRegex, matcher -> String.format("#%s/%s", matcher.group(2), matcher.group(3)), matcher -> Optional.of(String.format("https://juick.com/m/%s#%s", matcher.group(2), matcher.group(3))), 0));
        // /12
        result.addAll(entitiesForType("a", txt, replyNumberRegex, matcher -> "/" + matcher.group(2), matcher -> Optional.of(String.format("https://juick.com/m/%d#%s", msg.getMid(), matcher.group(2))), 0));
        // @username@jabber.org
        result.addAll(entitiesForType("a", txt, jidRegex, matcher -> matcher.group(2), matcher -> Optional.of(String.format("https://juick.com/%s", matcher.group(2))), 0));
        // @username
        result.addAll(entitiesForType("a", txt, usernameRegex, matcher -> matcher.group(2), matcher -> Optional.of(String.format("https://juick.com/%s", matcher.group(2))), 0));
        // *bold*
        result.addAll(entitiesForType("b", txt, boldRegex, matcher -> matcher.group(2), matcher -> Optional.empty(), 0));
        // /italic/
        result.addAll(entitiesForType("i", txt, italicRegex, matcher -> matcher.group(2), matcher -> Optional.empty(), 0));
        // _underline_
        result.addAll(entitiesForType("u", txt, underlineRegex, matcher -> matcher.group(2), matcher -> Optional.empty(), 0));
        // > citate
        result.addAll(entitiesForType("q", txt, citateRegex, matcher -> matcher.group(1), matcher -> Optional.empty(), 1));
        return result;
    }

    public static String formatMessageCode(String msg) {
        msg = msg.replaceAll("&", "&amp;");
        msg = msg.replaceAll("<", "&lt;");
        msg = msg.replaceAll(">", "&gt;");

        // http://juick.com/last?page=2
        // <a href="http://juick.com/last?page=2" rel="nofollow">http://juick.com/last?page=2</a>
        msg = msg.replaceAll(urlWithWhitespacesRegex, "$1<a href=\"$2\" rel=\"nofollow\">$2</a>");

        // (http://juick.com/last?page=2)
        // (<a href="http://juick.com/last?page=2" rel="nofollow">http://juick.com/last?page=2</a>)
        Matcher m = regexLinks2.matcher(msg);
        StringBuffer sb = new StringBuffer();
        while (m.find()) {
            String url = m.group(3).replace(" ", "%20").replaceAll("\\s+", StringUtils.EMPTY);
            m.appendReplacement(sb, "$1$2<a href=\"" + url + "\" rel=\"nofollow\">" + url + "</a>$5");
        }
        m.appendTail(sb);
        msg = sb.toString();

        return "<pre>" + msg + "</pre>";
    }

    public static String formatMessage(String msg) {
        return formatMessage(msg, false);
    }
    public static String formatMessage(String msg, boolean compatibleWithDurov) {

        msg = msg.replaceAll("&", "&amp;");
        msg = msg.replaceAll("<", "&lt;");
        msg = msg.replaceAll(">", "&gt;");

        // --
        // &mdash;
        if (!compatibleWithDurov) {
            msg = msg.replaceAll("((?<=\\s)|(?<=\\A))\\-\\-?((?=\\s)|(?=\\Z))", "$1&mdash;$2");
        }

        // http://juick.com/last?page=2
        // <a href="http://juick.com/last?page=2" rel="nofollow">juick.com</a>
        msg = msg.replaceAll(urlWithWhitespacesRegex, "$1<a href=\"$2\" rel=\"nofollow\">$3</a>");

        // [link text][http://juick.com/last?page=2]
        // [link text](http://juick.com/last?page=2)
        // www.juick.com <http://juick.com/>
        // <a href="http://juick.com/last?page=2" rel="nofollow">link text</a>
        msg = msg.replaceAll(textUrlRegex, "<a href=\"$2\" rel=\"nofollow\">$1</a>");
        msg = msg.replaceAll(textUrlRegex2, "<a href=\"$2\" rel=\"nofollow\">$1</a>");
        msg = msg.replaceAll(pidginUrlRegex, "<a href=\"$2\" rel=\"nofollow\">$1</a>");

        // #12345
        // <a href="http://juick.com/12345">#12345</a>
        msg = msg.replaceAll(midRegex, "$1<a href=\"https://juick.com/m/$2\">#$2</a>$3");

        // #12345/65
        // <a href="http://juick.com/12345#65">#12345/65</a>
        msg = msg.replaceAll(ridRegex, "$1<a href=\"https://juick.com/m/$2#$3\">#$2/$3</a>$4");

        // *bold*
        // <b>bold</b>
        msg = msg.replaceAll(boldRegex, "$1<b>$2</b>$3");

        // /italic/
        // <i>italic</i>
        msg = msg.replaceAll(italicRegex, "$1<i>$2</i>$3");

        // _underline_
        // <span class="u">underline</span>
        msg = msg.replaceAll(underlineRegex, "$1<u>$2</u>$3");

        // /12
        // <a href="#12">/12</a>
        msg = msg.replaceAll(replyNumberRegex, "$1<a href=\"#$2\">/$2</a>$3");


        // @username@mastodon.social
        // <a href="http://juick.com/mention?username=username@mastodon.social/">@username@mastodon.social</a>
        if (compatibleWithDurov) {
            msg = msg.replaceAll(jidRegex, "$1<a href=\"http://juick.com/mention?username=$2\">@$2</a>$3");
        } else {
            msg = msg.replaceAll(jidRegex, "$1<span class=\"h-card\"><a class=\"u-url\" href=\"http://juick.com/mention?username=$2\">@$2</a></span>$3");
        }

        // @username
        // <a href="http://juick.com/username/">@username</a>
        if (compatibleWithDurov) {
            msg = msg.replaceAll(usernameRegex, "$1<a href=\"https://juick.com/$2/\">@$2</a>$3");
        } else {
            msg = msg.replaceAll(usernameRegex, "$1<span class=\"h-card\"><a class=\"u-url\" href=\"https://juick.com/$2/\">@$2</a></span>$3");
        }
        // (http://juick.com/last?page=2)
        // (<a href="http://juick.com/last?page=2" rel="nofollow">juick.com</a>)
        Matcher m = regexLinks2.matcher(msg);
        StringBuffer sb = new StringBuffer();
        while (m.find()) {
            String url = m.group(3).replace(" ", "%20").replace("$", "%36").replaceAll("\\s+", StringUtils.EMPTY);
            m.appendReplacement(sb, "$1$2<a href=\"" + url + "\" rel=\"nofollow\">$4</a>$5");
        }
        m.appendTail(sb);
        msg = sb.toString();

        if (!compatibleWithDurov) {
            // > citate
            msg = msg.replaceAll(citateRegex, "<blockquote>$1</blockquote>");
            msg = msg.replaceAll("</blockquote><blockquote>", "\n");

            msg = msg.replaceAll("\n", "<br/>\n");
        }
        return msg;
    }

    public static String formatHtml(Message jmsg) {
        StringBuilder sb = new StringBuilder();
        boolean isReply = jmsg.getRid() > 0;
        String title = isReply ? "<b>Reply by @"  : "<b>@";
        String subtitle = isReply ? "<blockquote>" + jmsg.getReplyQuote() + "</blockquote>" : "<i>" + getTagsString(jmsg) + "</i>";
        boolean isCode = jmsg.getTags().stream().anyMatch(t -> t.getName().equals("code"));

        sb.append(title).append(jmsg.getUser().getName()).append(":</b></br/>")
                .append(subtitle).append("<br/>")
                .append(isCode ? formatMessageCode(StringUtils.defaultString(jmsg.getText()))
                        : formatMessage(StringUtils.defaultString(jmsg.getText()))).append("<br />");
        if (StringUtils.isNotEmpty(jmsg.getAttachmentType())) {
            // FIXME: attachment does not serialized to xml
            if (jmsg.getAttachment() == null) {
                if (jmsg.getRid() > 0) {
                    sb.append(String.format("<img src=\"http://i.juick.com/photos-1024/%d-%d.%s\" />", jmsg.getMid(),
                            jmsg.getRid(), jmsg.getAttachmentType()));
                } else {
                    sb.append(String.format("<img src=\"http://i.juick.com/photos-1024/%d.%s\" />", jmsg.getMid(),
                            jmsg.getAttachmentType()));
                }
            } else {
                sb.append("<img src=\"").append(jmsg.getAttachment().getMedium().getUrl()).append("\" />");
            }
        }
        return sb.toString();
    }

    public static String getMessageHashTags(final Message jmsg) {
        StringBuilder hashtags = new StringBuilder();
        for (Tag tag : jmsg.getTags()) {
            hashtags.append("#").append(tag).append(" ");
        }
        return hashtags.toString();
    }
    public static String getMarkdownTags(final Message jmsg) {
        return jmsg.getTags().stream().map(t -> String.format("[%s](http://juick.com/tag/%s)", t.getName(), percentEncode(t.getName())))
                .collect(Collectors.joining(", "));
    }

    public static String getUserHtmlLink(final User user, final String webDomain) {
        if (user.getUri().toASCIIString().length() > 0) {
            return String.format("<a href=\"%s\">%s</a>", user.getUri(), user.getName());
        } else {
            return String.format("<a href=\"https://%s/%s\">%s</a>", webDomain, user.getName(), user.getName());
        }
    }

    // TODO: check if it is really needed
    public static String percentEncode(final String s) {
        return URLEncoder.encode(s, StandardCharsets.UTF_8).replace("+", "%20")
                .replace("*", "%2A").replace("%7E", "~");
    }
    public static String formatMarkdownText(final Message msg) {
        return StringUtils.defaultString(msg.getText())
                .replaceAll(replyNumberRegex, String.format("$1[/$2](https://juick.com/m/%d#$2)$3", msg.getMid()));
    }
    public static String attachmentUrl(final Message jmsg) {
        if (StringUtils.isEmpty(jmsg.getAttachmentType())) {
            return StringUtils.EMPTY;
        }
        // FIXME: attachment does not serialized to xml
        if (jmsg.getAttachment() == null) {
            if (jmsg.getRid() > 0) {
                return String.format("http://i.juick.com/photos-1024/%d-%d.%s", jmsg.getMid(),
                        jmsg.getRid(), jmsg.getAttachmentType());
            } else {
                return String.format("http://i.juick.com/photos-1024/%d.%s", jmsg.getMid(),
                        jmsg.getAttachmentType());
            }
        } else {
            return jmsg.getAttachment().getMedium().getUrl();
        }
    }
    public static boolean replyStartsWithQuote(Message msg) {
        return msg.getRid() > 0 && StringUtils.defaultString(msg.getText()).startsWith(">");
    }
    public static Set<Tag> parseTags(String strTags) {
        return StringUtils.isEmpty(strTags) ? Collections.emptySet()
                : Arrays.stream(strTags.split(" ")).map(Tag::new)
                .collect(Collectors.toCollection(LinkedHashSet::new));
    }
    public static String getTagsString(Message msg) {
        StringBuilder builder = new StringBuilder();
        Set<Tag> tags = msg.getTags();

        if (!tags.isEmpty()) {
            for (Tag Tag : tags)
                builder.append(" *").append(Tag.getName());

            if (msg.isFriendsOnly())
                builder.append(" *friends");
            if (msg.ReadOnly)
                builder.append(" *readonly");
        }
        return builder.toString();
    }
    public static boolean isPM(Message message) {
        return message.getMid() == 0;
    }
    public static boolean isReply(Message message) {
        return message.getRid() > 0;
    }

    public static String stripNonSafeUrls(String input) {
        // strip login urls
        try {
            Matcher urlMatcher = Pattern.compile(MessageUtils.urlRegex).matcher(input);
            while (urlMatcher.find()) {
                URI uri = URI.create(urlMatcher.group(0));
                if (uri.getHost().equals("juick.com")) {
                    UriComponentsBuilder uriComponentsBuilder = UriComponentsBuilder.fromUri(uri);
                    uriComponentsBuilder.replaceQueryParam("hash");
                    input = input.replace(urlMatcher.group(0), uriComponentsBuilder.build().toUriString());
                }
            }
        } catch (IllegalArgumentException | NullPointerException e) {
            return input;
        }
        return input;
    }
    private static List<String> collectMatches(Pattern pattern, String input) {
        Matcher matcher = pattern.matcher(input);
        List<String> result = new ArrayList<>();
        while (matcher.find()) {
            result.add(matcher.group());
        }
        return result;
    }
    public static List<String> getMentions(Message msg) {
        return collectMatches(usernamePattern, msg.getText());
    }
    public static List<String> getGlobalMentions(Message msg) {
        return collectMatches(jidPattern, msg.getText());
    }

    /**
     *
     * @param type Name of the entity
     * @param input data to find matches
     * @param patternText pattern to match
     * @param textGroup function which return text representation
     * @param linkGroup function which return link address
     * @param endGroupId group id used to set end of entity (e.g. do not count linebreak as part of quote entity)
     * @return list of entities
     */
    private static List<Entity> entitiesForType(String type, String input, String patternText,
                                                Function<Matcher, String> textGroup,
                                                Function<Matcher, Optional<String>> linkGroup,
                                                int endGroupId) {
        List<Entity> result = new ArrayList<>();
        Pattern pattern = Pattern.compile(patternText);
        Matcher matcher = pattern.matcher(input);
        while (matcher.find()) {
            Entity entity = new Entity();
            entity.setType(type);
            entity.setText(textGroup.apply(matcher));
            Optional<String> link = linkGroup.apply(matcher);
            link.ifPresent(entity::setUrl);
            entity.setStart(matcher.start());
            entity.setEnd(matcher.end(endGroupId));
            result.add(entity);
        }
        return result;
    }

    public static boolean isSensitive(Message msg) {
        return msg.getTags().stream().anyMatch((t) -> t.getName().equals("NSFW"));
    }
}