001// License: GPL. For details, see LICENSE file.
002package org.openstreetmap.josm.io;
003
004import static org.openstreetmap.josm.tools.I18n.tr;
005
006import java.io.BufferedInputStream;
007import java.io.BufferedReader;
008import java.io.Closeable;
009import java.io.File;
010import java.io.FileInputStream;
011import java.io.IOException;
012import java.io.InputStream;
013import java.net.HttpURLConnection;
014import java.net.MalformedURLException;
015import java.net.URL;
016import java.nio.charset.StandardCharsets;
017import java.nio.file.Files;
018import java.nio.file.StandardCopyOption;
019import java.util.ArrayList;
020import java.util.Arrays;
021import java.util.Enumeration;
022import java.util.List;
023import java.util.Map;
024import java.util.concurrent.ConcurrentHashMap;
025import java.util.zip.ZipEntry;
026import java.util.zip.ZipFile;
027
028import org.openstreetmap.josm.Main;
029import org.openstreetmap.josm.tools.HttpClient;
030import org.openstreetmap.josm.tools.Pair;
031import org.openstreetmap.josm.tools.Utils;
032
033/**
034 * Downloads a file and caches it on disk in order to reduce network load.
035 *
036 * Supports URLs, local files, and a custom scheme (<code>resource:</code>) to get
037 * resources from the current *.jar file. (Local caching is only done for URLs.)
038 * <p>
039 * The mirrored file is only downloaded if it has been more than 7 days since
040 * last download. (Time can be configured.)
041 * <p>
042 * The file content is normally accessed with {@link #getInputStream()}, but
043 * you can also get the mirrored copy with {@link #getFile()}.
044 */
045public class CachedFile implements Closeable {
046
047    /**
048     * Caching strategy.
049     */
050    public enum CachingStrategy {
051        /**
052         * If cached file on disk is older than a certain time (7 days by default),
053         * consider the cache stale and try to download the file again.
054         */
055        MaxAge,
056        /**
057         * Similar to MaxAge, considers the cache stale when a certain age is
058         * exceeded. In addition, a If-Modified-Since HTTP header is added.
059         * When the server replies "304 Not Modified", this is considered the same
060         * as a full download.
061         */
062        IfModifiedSince
063    }
064
065    protected String name;
066    protected long maxAge;
067    protected String destDir;
068    protected String httpAccept;
069    protected CachingStrategy cachingStrategy;
070
071    private boolean fastFail;
072    private HttpClient activeConnection;
073    protected File cacheFile;
074    protected boolean initialized;
075
076    public static final long DEFAULT_MAXTIME = -1L;
077    public static final long DAYS = 24*60*60; // factor to get caching time in days
078
079    private final Map<String, String> httpHeaders = new ConcurrentHashMap<>();
080
081    /**
082     * Constructs a CachedFile object from a given filename, URL or internal resource.
083     *
084     * @param name can be:<ul>
085     *  <li>relative or absolute file name</li>
086     *  <li>{@code file:///SOME/FILE} the same as above</li>
087     *  <li>{@code http://...} a URL. It will be cached on disk.</li>
088     *  <li>{@code resource://SOME/FILE} file from the classpath (usually in the current *.jar)</li>
089     *  <li>{@code josmdir://SOME/FILE} file inside josm user data directory (since r7058)</li>
090     *  <li>{@code josmplugindir://SOME/FILE} file inside josm plugin directory (since r7834)</li></ul>
091     */
092    public CachedFile(String name) {
093        this.name = name;
094    }
095
096    /**
097     * Set the name of the resource.
098     * @param name can be:<ul>
099     *  <li>relative or absolute file name</li>
100     *  <li>{@code file:///SOME/FILE} the same as above</li>
101     *  <li>{@code http://...} a URL. It will be cached on disk.</li>
102     *  <li>{@code resource://SOME/FILE} file from the classpath (usually in the current *.jar)</li>
103     *  <li>{@code josmdir://SOME/FILE} file inside josm user data directory (since r7058)</li>
104     *  <li>{@code josmplugindir://SOME/FILE} file inside josm plugin directory (since r7834)</li></ul>
105     * @return this object
106     */
107    public CachedFile setName(String name) {
108        this.name = name;
109        return this;
110    }
111
112    /**
113     * Set maximum age of cache file. Only applies to URLs.
114     * When this time has passed after the last download of the file, the
115     * cache is considered stale and a new download will be attempted.
116     * @param maxAge the maximum cache age in seconds
117     * @return this object
118     */
119    public CachedFile setMaxAge(long maxAge) {
120        this.maxAge = maxAge;
121        return this;
122    }
123
124    /**
125     * Set the destination directory for the cache file. Only applies to URLs.
126     * @param destDir the destination directory
127     * @return this object
128     */
129    public CachedFile setDestDir(String destDir) {
130        this.destDir = destDir;
131        return this;
132    }
133
134    /**
135     * Set the accepted MIME types sent in the HTTP Accept header. Only applies to URLs.
136     * @param httpAccept the accepted MIME types
137     * @return this object
138     */
139    public CachedFile setHttpAccept(String httpAccept) {
140        this.httpAccept = httpAccept;
141        return this;
142    }
143
144    /**
145     * Set the caching strategy. Only applies to URLs.
146     * @param cachingStrategy caching strategy
147     * @return this object
148     */
149    public CachedFile setCachingStrategy(CachingStrategy cachingStrategy) {
150        this.cachingStrategy = cachingStrategy;
151        return this;
152    }
153
154    /**
155     * Sets the http headers. Only applies to URL pointing to http or https resources
156     * @param headers that should be sent together with request
157     * @return this object
158     */
159    public CachedFile setHttpHeaders(Map<String, String> headers) {
160        this.httpHeaders.putAll(headers);
161        return this;
162    }
163
164    /**
165     * Sets whether opening HTTP connections should fail fast, i.e., whether a
166     * {@link HttpClient#setConnectTimeout(int) low connect timeout} should be used.
167     * @param fastFail whether opening HTTP connections should fail fast
168     */
169    public void setFastFail(boolean fastFail) {
170        this.fastFail = fastFail;
171    }
172
173    public String getName() {
174        return name;
175    }
176
177    public long getMaxAge() {
178        return maxAge;
179    }
180
181    public String getDestDir() {
182        return destDir;
183    }
184
185    public String getHttpAccept() {
186        return httpAccept;
187    }
188
189    public CachingStrategy getCachingStrategy() {
190        return cachingStrategy;
191    }
192
193    /**
194     * Get InputStream to the requested resource.
195     * @return the InputStream
196     * @throws IOException when the resource with the given name could not be retrieved
197     */
198    public InputStream getInputStream() throws IOException {
199        File file = getFile();
200        if (file == null) {
201            if (name.startsWith("resource://")) {
202                InputStream is = getClass().getResourceAsStream(
203                        name.substring("resource:/".length()));
204                if (is == null)
205                    throw new IOException(tr("Failed to open input stream for resource ''{0}''", name));
206                return is;
207            } else {
208                throw new IOException("No file found for: "+name);
209            }
210        }
211        return new FileInputStream(file);
212    }
213
214    /**
215     * Returns {@link #getInputStream()} wrapped in a buffered reader.
216     * <p>
217     * Detects Unicode charset in use utilizing {@link UTFInputStreamReader}.
218     *
219     * @return buffered reader
220     * @throws IOException if any I/O error occurs
221     * @since 9411
222     */
223    public BufferedReader getContentReader() throws IOException {
224        return new BufferedReader(UTFInputStreamReader.create(getInputStream()));
225    }
226
227    /**
228     * Get local file for the requested resource.
229     * @return The local cache file for URLs. If the resource is a local file,
230     * returns just that file.
231     * @throws IOException when the resource with the given name could not be retrieved
232     */
233    public synchronized File getFile() throws IOException {
234        if (initialized)
235            return cacheFile;
236        initialized = true;
237        URL url;
238        try {
239            url = new URL(name);
240            if ("file".equals(url.getProtocol())) {
241                cacheFile = new File(name.substring("file:/".length() - 1));
242                if (!cacheFile.exists()) {
243                    cacheFile = new File(name.substring("file://".length() - 1));
244                }
245            } else {
246                cacheFile = checkLocal(url);
247            }
248        } catch (MalformedURLException e) {
249            if (name.startsWith("resource://")) {
250                return null;
251            } else if (name.startsWith("josmdir://")) {
252                cacheFile = new File(Main.pref.getUserDataDirectory(), name.substring("josmdir://".length()));
253            } else if (name.startsWith("josmplugindir://")) {
254                cacheFile = new File(Main.pref.getPluginsDirectory(), name.substring("josmplugindir://".length()));
255            } else {
256                cacheFile = new File(name);
257            }
258        }
259        if (cacheFile == null)
260            throw new IOException("Unable to get cache file for "+name);
261        return cacheFile;
262    }
263
264    /**
265     * Looks for a certain entry inside a zip file and returns the entry path.
266     *
267     * Replies a file in the top level directory of the ZIP file which has an
268     * extension <code>extension</code>. If more than one files have this
269     * extension, the last file whose name includes <code>namepart</code>
270     * is opened.
271     *
272     * @param extension  the extension of the file we're looking for
273     * @param namepart the name part
274     * @return The zip entry path of the matching file. Null if this cached file
275     * doesn't represent a zip file or if there was no matching
276     * file in the ZIP file.
277     */
278    public String findZipEntryPath(String extension, String namepart) {
279        Pair<String, InputStream> ze = findZipEntryImpl(extension, namepart);
280        if (ze == null) return null;
281        return ze.a;
282    }
283
284    /**
285     * Like {@link #findZipEntryPath}, but returns the corresponding InputStream.
286     * @param extension  the extension of the file we're looking for
287     * @param namepart the name part
288     * @return InputStream to the matching file. Null if this cached file
289     * doesn't represent a zip file or if there was no matching
290     * file in the ZIP file.
291     * @since 6148
292     */
293    public InputStream findZipEntryInputStream(String extension, String namepart) {
294        Pair<String, InputStream> ze = findZipEntryImpl(extension, namepart);
295        if (ze == null) return null;
296        return ze.b;
297    }
298
299    private Pair<String, InputStream> findZipEntryImpl(String extension, String namepart) {
300        File file = null;
301        try {
302            file = getFile();
303        } catch (IOException ex) {
304            Main.warn(ex, false);
305        }
306        if (file == null)
307            return null;
308        Pair<String, InputStream> res = null;
309        try {
310            ZipFile zipFile = new ZipFile(file, StandardCharsets.UTF_8);
311            ZipEntry resentry = null;
312            Enumeration<? extends ZipEntry> entries = zipFile.entries();
313            while (entries.hasMoreElements()) {
314                ZipEntry entry = entries.nextElement();
315                if (entry.getName().endsWith('.' + extension)) {
316                    /* choose any file with correct extension. When more than
317                        one file, prefer the one which matches namepart */
318                    if (resentry == null || entry.getName().indexOf(namepart) >= 0) {
319                        resentry = entry;
320                    }
321                }
322            }
323            if (resentry != null) {
324                InputStream is = zipFile.getInputStream(resentry);
325                res = Pair.create(resentry.getName(), is);
326            } else {
327                Utils.close(zipFile);
328            }
329        } catch (Exception e) {
330            if (file.getName().endsWith(".zip")) {
331                Main.warn(tr("Failed to open file with extension ''{2}'' and namepart ''{3}'' in zip file ''{0}''. Exception was: {1}",
332                        file.getName(), e.toString(), extension, namepart));
333            }
334        }
335        return res;
336    }
337
338    /**
339     * Clear the cache for the given resource.
340     * This forces a fresh download.
341     * @param name the URL
342     */
343    public static void cleanup(String name) {
344        cleanup(name, null);
345    }
346
347    /**
348     * Clear the cache for the given resource.
349     * This forces a fresh download.
350     * @param name the URL
351     * @param destDir the destination directory (see {@link #setDestDir(java.lang.String)})
352     */
353    public static void cleanup(String name, String destDir) {
354        URL url;
355        try {
356            url = new URL(name);
357            if (!"file".equals(url.getProtocol())) {
358                String prefKey = getPrefKey(url, destDir);
359                List<String> localPath = new ArrayList<>(Main.pref.getCollection(prefKey));
360                if (localPath.size() == 2) {
361                    File lfile = new File(localPath.get(1));
362                    if (lfile.exists()) {
363                        Utils.deleteFile(lfile);
364                    }
365                }
366                Main.pref.putCollection(prefKey, null);
367            }
368        } catch (MalformedURLException e) {
369            Main.warn(e);
370        }
371    }
372
373    /**
374     * Get preference key to store the location and age of the cached file.
375     * 2 resources that point to the same url, but that are to be stored in different
376     * directories will not share a cache file.
377     * @param url URL
378     * @param destDir destination directory
379     * @return Preference key
380     */
381    private static String getPrefKey(URL url, String destDir) {
382        StringBuilder prefKey = new StringBuilder("mirror.");
383        if (destDir != null) {
384            prefKey.append(destDir).append('.');
385        }
386        prefKey.append(url.toString());
387        return prefKey.toString().replaceAll("=", "_");
388    }
389
390    private File checkLocal(URL url) throws IOException {
391        String prefKey = getPrefKey(url, destDir);
392        String urlStr = url.toExternalForm();
393        long age = 0L;
394        long lMaxAge = maxAge;
395        Long ifModifiedSince = null;
396        File localFile = null;
397        List<String> localPathEntry = new ArrayList<>(Main.pref.getCollection(prefKey));
398        boolean offline = false;
399        try {
400            checkOfflineAccess(urlStr);
401        } catch (OfflineAccessException e) {
402            offline = true;
403        }
404        if (localPathEntry.size() == 2) {
405            localFile = new File(localPathEntry.get(1));
406            if (!localFile.exists()) {
407                localFile = null;
408            } else {
409                if (maxAge == DEFAULT_MAXTIME
410                        || maxAge <= 0 // arbitrary value <= 0 is deprecated
411                ) {
412                    lMaxAge = Main.pref.getInteger("mirror.maxtime", 7*24*60*60); // one week
413                }
414                age = System.currentTimeMillis() - Long.parseLong(localPathEntry.get(0));
415                if (offline || age < lMaxAge*1000) {
416                    return localFile;
417                }
418                if (cachingStrategy == CachingStrategy.IfModifiedSince) {
419                    ifModifiedSince = Long.valueOf(localPathEntry.get(0));
420                }
421            }
422        }
423        if (destDir == null) {
424            destDir = Main.pref.getCacheDirectory().getPath();
425        }
426
427        File destDirFile = new File(destDir);
428        if (!destDirFile.exists()) {
429            Utils.mkDirs(destDirFile);
430        }
431
432        // No local file + offline => nothing to do
433        if (offline) {
434            return null;
435        }
436
437        String a = urlStr.replaceAll("[^A-Za-z0-9_.-]", "_");
438        String localPath = "mirror_" + a;
439        destDirFile = new File(destDir, localPath + ".tmp");
440        try {
441            activeConnection = HttpClient.create(url)
442                    .setAccept(httpAccept)
443                    .setIfModifiedSince(ifModifiedSince == null ? 0L : ifModifiedSince)
444                    .setHeaders(httpHeaders);
445            if (fastFail) {
446                activeConnection.setReadTimeout(1000);
447            }
448            final HttpClient.Response con = activeConnection.connect();
449            if (ifModifiedSince != null && con.getResponseCode() == HttpURLConnection.HTTP_NOT_MODIFIED) {
450                if (Main.isDebugEnabled()) {
451                    Main.debug("304 Not Modified ("+urlStr+')');
452                }
453                if (localFile == null)
454                    throw new AssertionError();
455                Main.pref.putCollection(prefKey,
456                        Arrays.asList(Long.toString(System.currentTimeMillis()), localPathEntry.get(1)));
457                return localFile;
458            }
459            try (InputStream bis = new BufferedInputStream(con.getContent())) {
460                Files.copy(bis, destDirFile.toPath(), StandardCopyOption.REPLACE_EXISTING);
461            }
462            activeConnection = null;
463            localFile = new File(destDir, localPath);
464            if (Main.platform.rename(destDirFile, localFile)) {
465                Main.pref.putCollection(prefKey,
466                        Arrays.asList(Long.toString(System.currentTimeMillis()), localFile.toString()));
467            } else {
468                Main.warn(tr("Failed to rename file {0} to {1}.",
469                destDirFile.getPath(), localFile.getPath()));
470            }
471        } catch (IOException e) {
472            if (age >= lMaxAge*1000 && age < lMaxAge*1000*2) {
473                Main.warn(tr("Failed to load {0}, use cached file and retry next time: {1}", urlStr, e));
474                return localFile;
475            } else {
476                throw e;
477            }
478        }
479
480        return localFile;
481    }
482
483    private static void checkOfflineAccess(String urlString) {
484        OnlineResource.JOSM_WEBSITE.checkOfflineAccess(urlString, Main.getJOSMWebsite());
485        OnlineResource.OSM_API.checkOfflineAccess(urlString, OsmApi.getOsmApi().getServerUrl());
486    }
487
488    /**
489     * Attempts to disconnect an URL connection.
490     * @see HttpClient#disconnect()
491     * @since 9411
492     */
493    @Override
494    public void close() {
495        if (activeConnection != null) {
496            activeConnection.disconnect();
497        }
498    }
499}