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.ByteArrayOutputStream;
009import java.io.Closeable;
010import java.io.File;
011import java.io.FileInputStream;
012import java.io.IOException;
013import java.io.InputStream;
014import java.net.HttpURLConnection;
015import java.net.MalformedURLException;
016import java.net.URL;
017import java.nio.charset.StandardCharsets;
018import java.nio.file.Files;
019import java.nio.file.StandardCopyOption;
020import java.util.ArrayList;
021import java.util.Arrays;
022import java.util.Enumeration;
023import java.util.List;
024import java.util.Map;
025import java.util.concurrent.ConcurrentHashMap;
026import java.util.zip.ZipEntry;
027import java.util.zip.ZipFile;
028
029import org.openstreetmap.josm.Main;
030import org.openstreetmap.josm.tools.HttpClient;
031import org.openstreetmap.josm.tools.Pair;
032import org.openstreetmap.josm.tools.Utils;
033
034/**
035 * Downloads a file and caches it on disk in order to reduce network load.
036 *
037 * Supports URLs, local files, and a custom scheme (<code>resource:</code>) to get
038 * resources from the current *.jar file. (Local caching is only done for URLs.)
039 * <p>
040 * The mirrored file is only downloaded if it has been more than 7 days since
041 * last download. (Time can be configured.)
042 * <p>
043 * The file content is normally accessed with {@link #getInputStream()}, but
044 * you can also get the mirrored copy with {@link #getFile()}.
045 */
046public class CachedFile implements Closeable {
047
048    /**
049     * Caching strategy.
050     */
051    public enum CachingStrategy {
052        /**
053         * If cached file on disk is older than a certain time (7 days by default),
054         * consider the cache stale and try to download the file again.
055         */
056        MaxAge,
057        /**
058         * Similar to MaxAge, considers the cache stale when a certain age is
059         * exceeded. In addition, a If-Modified-Since HTTP header is added.
060         * When the server replies "304 Not Modified", this is considered the same
061         * as a full download.
062         */
063        IfModifiedSince
064    }
065
066    protected String name;
067    protected long maxAge;
068    protected String destDir;
069    protected String httpAccept;
070    protected CachingStrategy cachingStrategy;
071
072    private boolean fastFail;
073    private HttpClient activeConnection;
074    protected File cacheFile;
075    protected boolean initialized;
076
077    public static final long DEFAULT_MAXTIME = -1L;
078    public static final long DAYS = 24L*60L*60L; // factor to get caching time in days
079
080    private final Map<String, String> httpHeaders = new ConcurrentHashMap<>();
081
082    /**
083     * Constructs a CachedFile object from a given filename, URL or internal resource.
084     *
085     * @param name can be:<ul>
086     *  <li>relative or absolute file name</li>
087     *  <li>{@code file:///SOME/FILE} the same as above</li>
088     *  <li>{@code http://...} a URL. It will be cached on disk.</li>
089     *  <li>{@code resource://SOME/FILE} file from the classpath (usually in the current *.jar)</li>
090     *  <li>{@code josmdir://SOME/FILE} file inside josm user data directory (since r7058)</li>
091     *  <li>{@code josmplugindir://SOME/FILE} file inside josm plugin directory (since r7834)</li></ul>
092     */
093    public CachedFile(String name) {
094        this.name = name;
095    }
096
097    /**
098     * Set the name of the resource.
099     * @param name can be:<ul>
100     *  <li>relative or absolute file name</li>
101     *  <li>{@code file:///SOME/FILE} the same as above</li>
102     *  <li>{@code http://...} a URL. It will be cached on disk.</li>
103     *  <li>{@code resource://SOME/FILE} file from the classpath (usually in the current *.jar)</li>
104     *  <li>{@code josmdir://SOME/FILE} file inside josm user data directory (since r7058)</li>
105     *  <li>{@code josmplugindir://SOME/FILE} file inside josm plugin directory (since r7834)</li></ul>
106     * @return this object
107     */
108    public CachedFile setName(String name) {
109        this.name = name;
110        return this;
111    }
112
113    /**
114     * Set maximum age of cache file. Only applies to URLs.
115     * When this time has passed after the last download of the file, the
116     * cache is considered stale and a new download will be attempted.
117     * @param maxAge the maximum cache age in seconds
118     * @return this object
119     */
120    public CachedFile setMaxAge(long maxAge) {
121        this.maxAge = maxAge;
122        return this;
123    }
124
125    /**
126     * Set the destination directory for the cache file. Only applies to URLs.
127     * @param destDir the destination directory
128     * @return this object
129     */
130    public CachedFile setDestDir(String destDir) {
131        this.destDir = destDir;
132        return this;
133    }
134
135    /**
136     * Set the accepted MIME types sent in the HTTP Accept header. Only applies to URLs.
137     * @param httpAccept the accepted MIME types
138     * @return this object
139     */
140    public CachedFile setHttpAccept(String httpAccept) {
141        this.httpAccept = httpAccept;
142        return this;
143    }
144
145    /**
146     * Set the caching strategy. Only applies to URLs.
147     * @param cachingStrategy caching strategy
148     * @return this object
149     */
150    public CachedFile setCachingStrategy(CachingStrategy cachingStrategy) {
151        this.cachingStrategy = cachingStrategy;
152        return this;
153    }
154
155    /**
156     * Sets the http headers. Only applies to URL pointing to http or https resources
157     * @param headers that should be sent together with request
158     * @return this object
159     */
160    public CachedFile setHttpHeaders(Map<String, String> headers) {
161        this.httpHeaders.putAll(headers);
162        return this;
163    }
164
165    /**
166     * Sets whether opening HTTP connections should fail fast, i.e., whether a
167     * {@link HttpClient#setConnectTimeout(int) low connect timeout} should be used.
168     * @param fastFail whether opening HTTP connections should fail fast
169     */
170    public void setFastFail(boolean fastFail) {
171        this.fastFail = fastFail;
172    }
173
174    public String getName() {
175        return name;
176    }
177
178    /**
179     * Returns maximum age of cache file. Only applies to URLs.
180     * When this time has passed after the last download of the file, the
181     * cache is considered stale and a new download will be attempted.
182     * @return the maximum cache age in seconds
183     */
184    public long getMaxAge() {
185        return maxAge;
186    }
187
188    public String getDestDir() {
189        return destDir;
190    }
191
192    public String getHttpAccept() {
193        return httpAccept;
194    }
195
196    public CachingStrategy getCachingStrategy() {
197        return cachingStrategy;
198    }
199
200    /**
201     * Get InputStream to the requested resource.
202     * @return the InputStream
203     * @throws IOException when the resource with the given name could not be retrieved
204     */
205    public InputStream getInputStream() throws IOException {
206        File file = getFile();
207        if (file == null) {
208            if (name.startsWith("resource://")) {
209                InputStream is = getClass().getResourceAsStream(
210                        name.substring("resource:/".length()));
211                if (is == null)
212                    throw new IOException(tr("Failed to open input stream for resource ''{0}''", name));
213                return is;
214            } else {
215                throw new IOException("No file found for: "+name);
216            }
217        }
218        return new FileInputStream(file);
219    }
220
221    /**
222     * Get the full content of the requested resource as a byte array.
223     * @return the full content of the requested resource as byte array
224     * @throws IOException in case of an I/O error
225     */
226    public byte[] getByteContent() throws IOException {
227        try (InputStream is = getInputStream()) {
228            ByteArrayOutputStream buffer = new ByteArrayOutputStream();
229            int nRead;
230            byte[] data = new byte[8192];
231            while ((nRead = is.read(data, 0, data.length)) != -1) {
232                buffer.write(data, 0, nRead);
233            }
234            buffer.flush();
235            return buffer.toByteArray();
236        }
237    }
238
239    /**
240     * Returns {@link #getInputStream()} wrapped in a buffered reader.
241     * <p>
242     * Detects Unicode charset in use utilizing {@link UTFInputStreamReader}.
243     *
244     * @return buffered reader
245     * @throws IOException if any I/O error occurs
246     * @since 9411
247     */
248    public BufferedReader getContentReader() throws IOException {
249        return new BufferedReader(UTFInputStreamReader.create(getInputStream()));
250    }
251
252    /**
253     * Get local file for the requested resource.
254     * @return The local cache file for URLs. If the resource is a local file,
255     * returns just that file.
256     * @throws IOException when the resource with the given name could not be retrieved
257     */
258    public synchronized File getFile() throws IOException {
259        if (initialized)
260            return cacheFile;
261        initialized = true;
262        URL url;
263        try {
264            url = new URL(name);
265            if ("file".equals(url.getProtocol())) {
266                cacheFile = new File(name.substring("file:/".length() - 1));
267                if (!cacheFile.exists()) {
268                    cacheFile = new File(name.substring("file://".length() - 1));
269                }
270            } else {
271                cacheFile = checkLocal(url);
272            }
273        } catch (MalformedURLException e) {
274            if (name.startsWith("resource://")) {
275                return null;
276            } else if (name.startsWith("josmdir://")) {
277                cacheFile = new File(Main.pref.getUserDataDirectory(), name.substring("josmdir://".length()));
278            } else if (name.startsWith("josmplugindir://")) {
279                cacheFile = new File(Main.pref.getPluginsDirectory(), name.substring("josmplugindir://".length()));
280            } else {
281                cacheFile = new File(name);
282            }
283        }
284        if (cacheFile == null)
285            throw new IOException("Unable to get cache file for "+name);
286        return cacheFile;
287    }
288
289    /**
290     * Looks for a certain entry inside a zip file and returns the entry path.
291     *
292     * Replies a file in the top level directory of the ZIP file which has an
293     * extension <code>extension</code>. If more than one files have this
294     * extension, the last file whose name includes <code>namepart</code>
295     * is opened.
296     *
297     * @param extension  the extension of the file we're looking for
298     * @param namepart the name part
299     * @return The zip entry path of the matching file. Null if this cached file
300     * doesn't represent a zip file or if there was no matching
301     * file in the ZIP file.
302     */
303    public String findZipEntryPath(String extension, String namepart) {
304        Pair<String, InputStream> ze = findZipEntryImpl(extension, namepart);
305        if (ze == null) return null;
306        return ze.a;
307    }
308
309    /**
310     * Like {@link #findZipEntryPath}, but returns the corresponding InputStream.
311     * @param extension  the extension of the file we're looking for
312     * @param namepart the name part
313     * @return InputStream to the matching file. Null if this cached file
314     * doesn't represent a zip file or if there was no matching
315     * file in the ZIP file.
316     * @since 6148
317     */
318    public InputStream findZipEntryInputStream(String extension, String namepart) {
319        Pair<String, InputStream> ze = findZipEntryImpl(extension, namepart);
320        if (ze == null) return null;
321        return ze.b;
322    }
323
324    private Pair<String, InputStream> findZipEntryImpl(String extension, String namepart) {
325        File file = null;
326        try {
327            file = getFile();
328        } catch (IOException ex) {
329            Main.warn(ex, false);
330        }
331        if (file == null)
332            return null;
333        Pair<String, InputStream> res = null;
334        try {
335            ZipFile zipFile = new ZipFile(file, StandardCharsets.UTF_8);
336            ZipEntry resentry = null;
337            Enumeration<? extends ZipEntry> entries = zipFile.entries();
338            while (entries.hasMoreElements()) {
339                ZipEntry entry = entries.nextElement();
340                if (entry.getName().endsWith('.' + extension)) {
341                    /* choose any file with correct extension. When more than
342                        one file, prefer the one which matches namepart */
343                    if (resentry == null || entry.getName().indexOf(namepart) >= 0) {
344                        resentry = entry;
345                    }
346                }
347            }
348            if (resentry != null) {
349                InputStream is = zipFile.getInputStream(resentry);
350                res = Pair.create(resentry.getName(), is);
351            } else {
352                Utils.close(zipFile);
353            }
354        } catch (IOException e) {
355            if (file.getName().endsWith(".zip")) {
356                Main.warn(tr("Failed to open file with extension ''{2}'' and namepart ''{3}'' in zip file ''{0}''. Exception was: {1}",
357                        file.getName(), e.toString(), extension, namepart));
358            }
359        }
360        return res;
361    }
362
363    /**
364     * Clear the cache for the given resource.
365     * This forces a fresh download.
366     * @param name the URL
367     */
368    public static void cleanup(String name) {
369        cleanup(name, null);
370    }
371
372    /**
373     * Clear the cache for the given resource.
374     * This forces a fresh download.
375     * @param name the URL
376     * @param destDir the destination directory (see {@link #setDestDir(java.lang.String)})
377     */
378    public static void cleanup(String name, String destDir) {
379        URL url;
380        try {
381            url = new URL(name);
382            if (!"file".equals(url.getProtocol())) {
383                String prefKey = getPrefKey(url, destDir);
384                List<String> localPath = new ArrayList<>(Main.pref.getCollection(prefKey));
385                if (localPath.size() == 2) {
386                    File lfile = new File(localPath.get(1));
387                    if (lfile.exists()) {
388                        Utils.deleteFile(lfile);
389                    }
390                }
391                Main.pref.putCollection(prefKey, null);
392            }
393        } catch (MalformedURLException e) {
394            Main.warn(e);
395        }
396    }
397
398    /**
399     * Get preference key to store the location and age of the cached file.
400     * 2 resources that point to the same url, but that are to be stored in different
401     * directories will not share a cache file.
402     * @param url URL
403     * @param destDir destination directory
404     * @return Preference key
405     */
406    private static String getPrefKey(URL url, String destDir) {
407        StringBuilder prefKey = new StringBuilder("mirror.");
408        if (destDir != null) {
409            prefKey.append(destDir).append('.');
410        }
411        prefKey.append(url.toString());
412        return prefKey.toString().replaceAll("=", "_");
413    }
414
415    private File checkLocal(URL url) throws IOException {
416        String prefKey = getPrefKey(url, destDir);
417        String urlStr = url.toExternalForm();
418        long age = 0L;
419        long lMaxAge = maxAge;
420        Long ifModifiedSince = null;
421        File localFile = null;
422        List<String> localPathEntry = new ArrayList<>(Main.pref.getCollection(prefKey));
423        boolean offline = false;
424        try {
425            checkOfflineAccess(urlStr);
426        } catch (OfflineAccessException e) {
427            offline = true;
428        }
429        if (localPathEntry.size() == 2) {
430            localFile = new File(localPathEntry.get(1));
431            if (!localFile.exists()) {
432                localFile = null;
433            } else {
434                if (maxAge == DEFAULT_MAXTIME
435                        || maxAge <= 0 // arbitrary value <= 0 is deprecated
436                ) {
437                    lMaxAge = Main.pref.getInteger("mirror.maxtime", 7*24*60*60); // one week
438                }
439                age = System.currentTimeMillis() - Long.parseLong(localPathEntry.get(0));
440                if (offline || age < lMaxAge*1000) {
441                    return localFile;
442                }
443                if (cachingStrategy == CachingStrategy.IfModifiedSince) {
444                    ifModifiedSince = Long.valueOf(localPathEntry.get(0));
445                }
446            }
447        }
448        if (destDir == null) {
449            destDir = Main.pref.getCacheDirectory().getPath();
450        }
451
452        File destDirFile = new File(destDir);
453        if (!destDirFile.exists()) {
454            Utils.mkDirs(destDirFile);
455        }
456
457        // No local file + offline => nothing to do
458        if (offline) {
459            return null;
460        }
461
462        String a = urlStr.replaceAll("[^A-Za-z0-9_.-]", "_");
463        String localPath = "mirror_" + a;
464        destDirFile = new File(destDir, localPath + ".tmp");
465        try {
466            activeConnection = HttpClient.create(url)
467                    .setAccept(httpAccept)
468                    .setIfModifiedSince(ifModifiedSince == null ? 0L : ifModifiedSince)
469                    .setHeaders(httpHeaders);
470            if (fastFail) {
471                activeConnection.setReadTimeout(1000);
472            }
473            final HttpClient.Response con = activeConnection.connect();
474            if (ifModifiedSince != null && con.getResponseCode() == HttpURLConnection.HTTP_NOT_MODIFIED) {
475                if (Main.isDebugEnabled()) {
476                    Main.debug("304 Not Modified ("+urlStr+')');
477                }
478                if (localFile == null)
479                    throw new AssertionError();
480                Main.pref.putCollection(prefKey,
481                        Arrays.asList(Long.toString(System.currentTimeMillis()), localPathEntry.get(1)));
482                return localFile;
483            }
484            try (InputStream bis = new BufferedInputStream(con.getContent())) {
485                Files.copy(bis, destDirFile.toPath(), StandardCopyOption.REPLACE_EXISTING);
486            }
487            activeConnection = null;
488            localFile = new File(destDir, localPath);
489            if (Main.platform.rename(destDirFile, localFile)) {
490                Main.pref.putCollection(prefKey,
491                        Arrays.asList(Long.toString(System.currentTimeMillis()), localFile.toString()));
492            } else {
493                Main.warn(tr("Failed to rename file {0} to {1}.",
494                destDirFile.getPath(), localFile.getPath()));
495            }
496        } catch (IOException e) {
497            if (age >= lMaxAge*1000 && age < lMaxAge*1000*2) {
498                Main.warn(tr("Failed to load {0}, use cached file and retry next time: {1}", urlStr, e));
499                return localFile;
500            } else {
501                throw e;
502            }
503        }
504
505        return localFile;
506    }
507
508    private static void checkOfflineAccess(String urlString) {
509        OnlineResource.JOSM_WEBSITE.checkOfflineAccess(urlString, Main.getJOSMWebsite());
510        OnlineResource.OSM_API.checkOfflineAccess(urlString, OsmApi.getOsmApi().getServerUrl());
511    }
512
513    /**
514     * Attempts to disconnect an URL connection.
515     * @see HttpClient#disconnect()
516     * @since 9411
517     */
518    @Override
519    public void close() {
520        if (activeConnection != null) {
521            activeConnection.disconnect();
522        }
523    }
524}