001// License: GPL. For details, see LICENSE file.
002package org.openstreetmap.josm.tools;
003
004import static org.openstreetmap.josm.tools.I18n.tr;
005import static org.openstreetmap.josm.tools.I18n.trn;
006
007import java.awt.GridBagLayout;
008import java.util.Arrays;
009import java.util.HashMap;
010import java.util.Map;
011import java.util.Map.Entry;
012import java.util.regex.Matcher;
013import java.util.regex.Pattern;
014
015import javax.swing.JLabel;
016import javax.swing.JOptionPane;
017import javax.swing.JPanel;
018
019import org.openstreetmap.josm.Main;
020import org.openstreetmap.josm.gui.ExtendedDialog;
021import org.openstreetmap.josm.gui.datatransfer.ClipboardUtils;
022import org.openstreetmap.josm.gui.help.HelpUtil;
023import org.openstreetmap.josm.gui.widgets.UrlLabel;
024import org.openstreetmap.josm.io.XmlWriter;
025import org.openstreetmap.josm.tools.LanguageInfo.LocaleType;
026
027/**
028 * Class that helps to parse tags from arbitrary text
029 */
030public final class TextTagParser {
031
032    // properties need JOSM restart to apply, modified rarely enough
033    private static final int MAX_KEY_LENGTH = Main.pref.getInteger("tags.paste.max-key-length", 50);
034    private static final int MAX_KEY_COUNT = Main.pref.getInteger("tags.paste.max-key-count", 30);
035    private static final String KEY_PATTERN = Main.pref.get("tags.paste.tag-pattern", "[0-9a-zA-Z:_]*");
036    private static final int MAX_VALUE_LENGTH = 255;
037
038    private TextTagParser() {
039        // Hide default constructor for utils classes
040    }
041
042    public static class TextAnalyzer {
043        private boolean quotesStarted;
044        private boolean esc;
045        private final StringBuilder s = new StringBuilder(200);
046        private int pos;
047        private final String data;
048        private final int n;
049
050        public TextAnalyzer(String text) {
051            pos = 0;
052            data = text;
053            n = data.length();
054        }
055
056        /**
057         * Read tags from "Free format"
058         * @return map of tags
059         */
060        private Map<String, String> getFreeParsedTags() {
061            String k, v;
062            Map<String, String> tags = new HashMap<>();
063
064            while (true) {
065                skipEmpty();
066                if (pos == n) {
067                    break;
068                }
069                k = parseString("\n\r\t= ");
070                if (pos == n) {
071                    tags.clear();
072                    break;
073                }
074                skipSign();
075                if (pos == n) {
076                    tags.clear();
077                    break;
078                }
079                v = parseString("\n\r\t ");
080                tags.put(k, v);
081            }
082            return tags;
083        }
084
085        private String parseString(String stopChars) {
086            char[] stop = stopChars.toCharArray();
087            Arrays.sort(stop);
088            char c;
089            while (pos < n) {
090                c = data.charAt(pos);
091                if (esc) {
092                    esc = false;
093                    s.append(c); //  \" \\
094                } else if (c == '\\') {
095                    esc = true;
096                } else if (c == '\"' && !quotesStarted) { // opening "
097                    if (!s.toString().trim().isEmpty()) { // we had   ||some text"||
098                        s.append(c); // just add ", not open
099                    } else {
100                        s.delete(0, s.length()); // forget that empty characthers and start reading "....
101                        quotesStarted = true;
102                    }
103                } else if (c == '\"' && quotesStarted) {  // closing "
104                    quotesStarted = false;
105                    pos++;
106                    break;
107                } else if (!quotesStarted && (Arrays.binarySearch(stop, c) >= 0)) {
108                    // stop-symbol found
109                    pos++;
110                    break;
111                } else {
112                    // skip non-printable characters
113                    if (c >= 32) s.append(c);
114                }
115                pos++;
116            }
117
118            String res = s.toString();
119            s.delete(0, s.length());
120            return res.trim();
121        }
122
123        private void skipSign() {
124            char c;
125            boolean signFound = false;
126            while (pos < n) {
127                c = data.charAt(pos);
128                if (c == '\t' || c == '\n' || c == ' ') {
129                    pos++;
130                } else if (c == '=') {
131                    if (signFound) break; // a  =  =qwerty means "a"="=qwerty"
132                    signFound = true;
133                    pos++;
134                } else {
135                    break;
136                }
137            }
138        }
139
140        private void skipEmpty() {
141            char c;
142            while (pos < n) {
143                c = data.charAt(pos);
144                if (c == '\t' || c == '\n' || c == '\r' || c == ' ') {
145                    pos++;
146                } else {
147                    break;
148                }
149            }
150        }
151    }
152
153    static String unescape(String k) {
154        if (!(k.startsWith("\"") && k.endsWith("\""))) {
155            if (k.contains("=")) {
156                // '=' not in quotes will be treated as an error!
157                return null;
158            } else {
159                return k;
160            }
161        }
162        String text = k.substring(1, k.length()-1);
163        return (new TextAnalyzer(text)).parseString("\r\t\n");
164    }
165
166    /**
167     * Try to find tag-value pairs in given text
168     * @param text - text in which tags are looked for
169     * @param splitRegex - text is splitted into parts with this delimiter
170     * @param tagRegex - each part is matched against this regex
171     * @param unescapeTextInQuotes - if true, matched tag and value will be analyzed more thoroughly
172     * @return map of tags
173     */
174    public static Map<String, String> readTagsByRegexp(String text, String splitRegex, String tagRegex, boolean unescapeTextInQuotes) {
175         String[] lines = text.split(splitRegex);
176         Pattern p = Pattern.compile(tagRegex);
177         Map<String, String> tags = new HashMap<>();
178         String k;
179         String v;
180         for (String line: lines) {
181            if (line.trim().isEmpty()) continue; // skip empty lines
182            Matcher m = p.matcher(line);
183            if (m.matches()) {
184                 k = m.group(1).trim();
185                 v = m.group(2).trim();
186                 if (unescapeTextInQuotes) {
187                     k = unescape(k);
188                     v = unescape(v);
189                     if (k == null || v == null) return null;
190                 }
191                 tags.put(k, v);
192            } else {
193                return null;
194            }
195         }
196         if (!tags.isEmpty()) {
197            return tags;
198         } else {
199            return null;
200         }
201    }
202
203    public static Map<String, String> getValidatedTagsFromText(String buf) {
204        Map<String, String> tags = readTagsFromText(buf);
205        return validateTags(tags) ? tags : null;
206    }
207
208    /**
209     * Apply different methods to extract tag-value pairs from arbitrary text
210     * @param buf buffer
211     * @return null if no format is suitable
212     */
213    public static Map<String, String> readTagsFromText(String buf) {
214        Map<String, String> tags;
215
216        // Format
217        // tag1\tval1\ntag2\tval2\n
218        tags = readTagsByRegexp(buf, "[\\r\\n]+", ".*?([a-zA-Z0-9:_]+).*\\t(.*?)", false);
219        // try "tag\tvalue\n" format
220        if (tags != null) return tags;
221
222        // Format
223        // a=b \n c=d \n "a b"=hello
224        // SORRY: "a=b" = c is not supported fror now, only first = will be considered
225        // a = "b=c" is OK
226        // a = b=c  - this method of parsing fails intentionally
227        tags = readTagsByRegexp(buf, "[\\n\\t\\r]+", "(.*?)=(.*?)", true);
228        // try format  t1=v1\n t2=v2\n ...
229        if (tags != null) return tags;
230
231        // JSON-format
232        String bufJson = buf.trim();
233        // trim { }, if there are any
234        if (bufJson.startsWith("{") && bufJson.endsWith("}"))
235            bufJson = bufJson.substring(1, bufJson.length()-1);
236        tags = readTagsByRegexp(bufJson, "[\\s]*,[\\s]*",
237                "[\\s]*(\\\".*?[^\\\\]\\\")"+"[\\s]*:[\\s]*"+"(\\\".*?[^\\\\]\\\")[\\s]*", true);
238        if (tags != null) return tags;
239
240        // Free format
241        // a 1 "b" 2 c=3 d 4 e "5"
242        return new TextAnalyzer(buf).getFreeParsedTags();
243    }
244
245    /**
246     * Check tags for correctness and display warnings if needed
247     * @param tags - map key-&gt;value to check
248     * @return true if the tags should be pasted
249     */
250    public static boolean validateTags(Map<String, String> tags) {
251        int r;
252        int s = tags.size();
253        if (s > MAX_KEY_COUNT) {
254            // Use trn() even if for english it makes no sense, as s > 30
255            r = warning(trn("There was {0} tag found in the buffer, it is suspicious!",
256            "There were {0} tags found in the buffer, it is suspicious!", s,
257            s), "", "tags.paste.toomanytags");
258            if (r == 2 || r == 3) return false; if (r == 4) return true;
259        }
260        for (Entry<String, String> entry : tags.entrySet()) {
261            String key = entry.getKey();
262            String value = entry.getValue();
263            if (key.length() > MAX_KEY_LENGTH) {
264                r = warning(tr("Key is too long (max {0} characters):", MAX_KEY_LENGTH), key+'='+value, "tags.paste.keytoolong");
265                if (r == 2 || r == 3) return false; if (r == 4) return true;
266            }
267            if (!key.matches(KEY_PATTERN)) {
268                r = warning(tr("Suspicious characters in key:"), key, "tags.paste.keydoesnotmatch");
269                if (r == 2 || r == 3) return false; if (r == 4) return true;
270            }
271            if (value.length() > MAX_VALUE_LENGTH) {
272                r = warning(tr("Value is too long (max {0} characters):", MAX_VALUE_LENGTH), value, "tags.paste.valuetoolong");
273                if (r == 2 || r == 3) return false; if (r == 4) return true;
274            }
275        }
276        return true;
277    }
278
279    private static int warning(String text, String data, String code) {
280        ExtendedDialog ed = new ExtendedDialog(
281                    Main.parent,
282                    tr("Do you want to paste these tags?"),
283                    new String[]{tr("Ok"), tr("Cancel"), tr("Clear buffer"), tr("Ignore warnings")});
284        ed.setButtonIcons(new String[]{"ok", "cancel", "dialogs/delete", "pastetags"});
285        ed.setContent("<html><b>"+text + "</b><br/><br/><div width=\"300px\">"+XmlWriter.encode(data, true)+"</html>");
286        ed.setDefaultButton(2);
287        ed.setCancelButton(2);
288        ed.setIcon(JOptionPane.WARNING_MESSAGE);
289        ed.toggleEnable(code);
290        ed.showDialog();
291        int r = ed.getValue();
292        if (r == 0) r = 2;
293        // clean clipboard if user asked
294        if (r == 3) ClipboardUtils.copyString("");
295        return r;
296    }
297
298    /**
299     * Shows message that the buffer can not be pasted, allowing user to clean the buffer
300     * @param helpTopic the help topic of the parent action
301     * TODO: Replace by proper HelpAwareOptionPane instead of self-made help link
302     */
303    public static void showBadBufferMessage(String helpTopic) {
304        String msg = tr("<html><p> Sorry, it is impossible to paste tags from buffer. It does not contain any JOSM object"
305            + " or suitable text. </p></html>");
306        JPanel p = new JPanel(new GridBagLayout());
307        p.add(new JLabel(msg), GBC.eop());
308        String helpUrl = HelpUtil.getHelpTopicUrl(HelpUtil.buildAbsoluteHelpTopic(helpTopic, LocaleType.DEFAULT));
309        if (helpUrl != null) {
310            p.add(new UrlLabel(helpUrl), GBC.eop());
311        }
312
313        ExtendedDialog ed = new ExtendedDialog(
314                    Main.parent,
315                    tr("Warning"),
316                    new String[]{tr("Ok"), tr("Clear buffer")});
317
318        ed.setButtonIcons(new String[]{"ok", "dialogs/delete"});
319
320        ed.setContent(p);
321        ed.setDefaultButton(1);
322        ed.setCancelButton(1);
323        ed.setIcon(JOptionPane.WARNING_MESSAGE);
324        ed.toggleEnable("tags.paste.cleanbadbuffer");
325        ed.showDialog();
326
327        int r = ed.getValue();
328        // clean clipboard if user asked
329        if (r == 2) ClipboardUtils.copyString("");
330    }
331}