LatinIME/java/src/com/android/inputmethod/latin/BinaryDictionaryGetter.java

/*
 * Copyright (C) 2011 The Android Open Source Project
 *
 * Licensed under the Apache License, Version 2.0 (the "License"); you may not
 * use this file except in compliance with the License. You may obtain a copy of
 * the License at
 *
 * http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
 * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
 * License for the specific language governing permissions and limitations under
 * the License.
 */

package com.android.inputmethod.latin;

import com.android.inputmethod.latin.makedict.BinaryDictInputOutput;
import com.android.inputmethod.latin.makedict.FormatSpec;

import android.content.Context;
import android.content.SharedPreferences;
import android.content.pm.PackageManager.NameNotFoundException;
import android.content.res.AssetFileDescriptor;
import android.util.Log;

import java.io.File;
import java.io.FileInputStream;
import java.io.IOException;
import java.nio.BufferUnderflowException;
import java.nio.ByteBuffer;
import java.nio.channels.FileChannel;
import java.util.ArrayList;
import java.util.HashMap;
import java.util.Locale;

/**
 * Helper class to get the address of a mmap'able dictionary file.
 */
class BinaryDictionaryGetter {

    /**
     * Used for Log actions from this class
     */
    private static final String TAG = BinaryDictionaryGetter.class.getSimpleName();

    /**
     * Used to return empty lists
     */
    private static final File[] EMPTY_FILE_ARRAY = new File[0];

    /**
     * Name of the common preferences name to know which word list are on and which are off.
     */
    private static final String COMMON_PREFERENCES_NAME = "LatinImeDictPrefs";

    // Name of the category for the main dictionary
    private static final String MAIN_DICTIONARY_CATEGORY = "main";
    public static final String ID_CATEGORY_SEPARATOR = ":";

    // The key considered to read the version attribute in a dictionary file.
    private static String VERSION_KEY = "version";

    // Prevents this from being instantiated
    private BinaryDictionaryGetter() {}

    /**
     * Returns whether we may want to use this character as part of a file name.
     *
     * This basically only accepts ascii letters and numbers, and rejects everything else.
     */
    private static boolean isFileNameCharacter(int codePoint) {
        if (codePoint >= 0x30 && codePoint <= 0x39) return true; // Digit
        if (codePoint >= 0x41 && codePoint <= 0x5A) return true; // Uppercase
        if (codePoint >= 0x61 && codePoint <= 0x7A) return true; // Lowercase
        return codePoint == '_'; // Underscore
    }

    /**
     * Escapes a string for any characters that may be suspicious for a file or directory name.
     *
     * Concretely this does a sort of URL-encoding except it will encode everything that's not
     * alphanumeric or underscore. (true URL-encoding leaves alone characters like '*', which
     * we cannot allow here)
     */
    // TODO: create a unit test for this method
    private static String replaceFileNameDangerousCharacters(final String name) {
        // This assumes '%' is fully available as a non-separator, normal
        // character in a file name. This is probably true for all file systems.
        final StringBuilder sb = new StringBuilder();
        final int nameLength = name.length();
        for (int i = 0; i < nameLength; i = name.offsetByCodePoints(i, 1)) {
            final int codePoint = name.codePointAt(i);
            if (isFileNameCharacter(codePoint)) {
                sb.appendCodePoint(codePoint);
            } else {
                // 6 digits - unicode is limited to 21 bits
                sb.append(String.format((Locale)null, "%%%1$06x", codePoint));
            }
        }
        return sb.toString();
    }

    /**
     * Reverse escaping done by replaceFileNameDangerousCharacters.
     */
    private static String getWordListIdFromFileName(final String fname) {
        final StringBuilder sb = new StringBuilder();
        final int fnameLength = fname.length();
        for (int i = 0; i < fnameLength; i = fname.offsetByCodePoints(i, 1)) {
            final int codePoint = fname.codePointAt(i);
            if ('%' != codePoint) {
                sb.appendCodePoint(codePoint);
            } else {
                final int encodedCodePoint = Integer.parseInt(fname.substring(i + 1, i + 7), 16);
                i += 6;
                sb.appendCodePoint(encodedCodePoint);
            }
        }
        return sb.toString();
    }

    /**
     * Helper method to get the top level cache directory.
     */
    private static String getWordListCacheDirectory(final Context context) {
        return context.getFilesDir() + File.separator + "dicts";
    }

    /**
     * Find out the cache directory associated with a specific locale.
     */
    private static String getCacheDirectoryForLocale(final String locale, final Context context) {
        final String relativeDirectoryName = replaceFileNameDangerousCharacters(locale);
        final String absoluteDirectoryName = getWordListCacheDirectory(context) + File.separator
                + relativeDirectoryName;
        final File directory = new File(absoluteDirectoryName);
        if (!directory.exists()) {
            if (!directory.mkdirs()) {
                Log.e(TAG, "Could not create the directory for locale" + locale);
            }
        }
        return absoluteDirectoryName;
    }

    /**
     * Generates a file name for the id and locale passed as an argument.
     *
     * In the current implementation the file name returned will always be unique for
     * any id/locale pair, but please do not expect that the id can be the same for
     * different dictionaries with different locales. An id should be unique for any
     * dictionary.
     * The file name is pretty much an URL-encoded version of the id inside a directory
     * named like the locale, except it will also escape characters that look dangerous
     * to some file systems.
     * @param id the id of the dictionary for which to get a file name
     * @param locale the locale for which to get the file name as a string
     * @param context the context to use for getting the directory
     * @return the name of the file to be created
     */
    public static String getCacheFileName(String id, String locale, Context context) {
        final String fileName = replaceFileNameDangerousCharacters(id);
        return getCacheDirectoryForLocale(locale, context) + File.separator + fileName;
    }

    /**
     * Returns a file address from a resource, or null if it cannot be opened.
     */
    private static AssetFileAddress loadFallbackResource(final Context context,
            final int fallbackResId) {
        final AssetFileDescriptor afd = context.getResources().openRawResourceFd(fallbackResId);
        if (afd == null) {
            Log.e(TAG, "Found the resource but cannot read it. Is it compressed? resId="
                    + fallbackResId);
            return null;
        }
        return AssetFileAddress.makeFromFileNameAndOffset(
                context.getApplicationInfo().sourceDir, afd.getStartOffset(), afd.getLength());
    }

    static private class DictPackSettings {
        final SharedPreferences mDictPreferences;
        public DictPackSettings(final Context context) {
            Context dictPackContext = null;
            try {
                final String dictPackName =
                        context.getString(R.string.dictionary_pack_package_name);
                dictPackContext = context.createPackageContext(dictPackName, 0);
            } catch (NameNotFoundException e) {
                // The dictionary pack is not installed...
                // TODO: fallback on the built-in dict, see the TODO above
                Log.e(TAG, "Could not find a dictionary pack");
            }
            mDictPreferences = null == dictPackContext ? null
                    : dictPackContext.getSharedPreferences(COMMON_PREFERENCES_NAME,
                            Context.MODE_WORLD_READABLE | Context.MODE_MULTI_PROCESS);
        }
        public boolean isWordListActive(final String dictId) {
            if (null == mDictPreferences) {
                // If we don't have preferences it basically means we can't find the dictionary
                // pack - either it's not installed, or it's disabled, or there is some strange
                // bug. Either way, a word list with no settings should be on by default: default
                // dictionaries in LatinIME are on if there is no settings at all, and if for some
                // reason some dictionaries have been installed BUT the dictionary pack can't be
                // found anymore it's safer to actually supply installed dictionaries.
                return true;
            } else {
                // The default is true here for the same reasons as above. We got the dictionary
                // pack but if we don't have any settings for it it means the user has never been
                // to the settings yet. So by default, the main dictionaries should be on.
                return mDictPreferences.getBoolean(dictId, true);
            }
        }
    }

    /**
     * Helper method to the list of cache directories, one for each distinct locale.
     */
    private static File[] getCachedDirectoryList(final Context context) {
        return new File(getWordListCacheDirectory(context)).listFiles();
    }

    /**
     * Returns the category for a given file name.
     *
     * This parses the file name, extracts the category, and returns it. See
     * {@link #getMainDictId(Locale)} and {@link #isMainWordListId(String)}.
     * @return The category as a string or null if it can't be found in the file name.
     */
    private static String getCategoryFromFileName(final String fileName) {
        final String id = getWordListIdFromFileName(fileName);
        final String[] idArray = id.split(ID_CATEGORY_SEPARATOR);
        if (2 != idArray.length) return null;
        return idArray[0];
    }

    /**
     * Utility class for the {@link #getCachedWordLists} method
     */
    private static class FileAndMatchLevel {
        final File mFile;
        final int mMatchLevel;
        public FileAndMatchLevel(final File file, final int matchLevel) {
            mFile = file;
            mMatchLevel = matchLevel;
        }
    }

    /**
     * Returns the list of cached files for a specific locale, one for each category.
     *
     * This will return exactly one file for each word list category that matches
     * the passed locale. If several files match the locale for any given category,
     * this returns the file with the closest match to the locale. For example, if
     * the passed word list is en_US, and for a category we have an en and an en_US
     * word list available, we'll return only the en_US one.
     * Thus, the list will contain as many files as there are categories.
     *
     * @param locale the locale to find the dictionary files for, as a string.
     * @param context the context on which to open the files upon.
     * @return an array of binary dictionary files, which may be empty but may not be null.
     */
    private static File[] getCachedWordLists(final String locale,
            final Context context) {
        final File[] directoryList = getCachedDirectoryList(context);
        if (null == directoryList) return EMPTY_FILE_ARRAY;
        final HashMap<String, FileAndMatchLevel> cacheFiles = CollectionUtils.newHashMap();
        for (File directory : directoryList) {
            if (!directory.isDirectory()) continue;
            final String dirLocale = getWordListIdFromFileName(directory.getName());
            final int matchLevel = LocaleUtils.getMatchLevel(dirLocale, locale);
            if (LocaleUtils.isMatch(matchLevel)) {
                final File[] wordLists = directory.listFiles();
                if (null != wordLists) {
                    for (File wordList : wordLists) {
                        final String category = getCategoryFromFileName(wordList.getName());
                        final FileAndMatchLevel currentBestMatch = cacheFiles.get(category);
                        if (null == currentBestMatch || currentBestMatch.mMatchLevel < matchLevel) {
                            cacheFiles.put(category, new FileAndMatchLevel(wordList, matchLevel));
                        }
                    }
                }
            }
        }
        if (cacheFiles.isEmpty()) return EMPTY_FILE_ARRAY;
        final File[] result = new File[cacheFiles.size()];
        int index = 0;
        for (final FileAndMatchLevel entry : cacheFiles.values()) {
            result[index++] = entry.mFile;
        }
        return result;
    }

    /**
     * Remove all files with the passed id, except the passed file.
     *
     * If a dictionary with a given ID has a metadata change that causes it to change
     * path, we need to remove the old version. The only way to do this is to check all
     * installed files for a matching ID in a different directory.
     */
    public static void removeFilesWithIdExcept(final Context context, final String id,
            final File fileToKeep) {
        try {
            final File canonicalFileToKeep = fileToKeep.getCanonicalFile();
            final File[] directoryList = getCachedDirectoryList(context);
            if (null == directoryList) return;
            for (File directory : directoryList) {
                // There is one directory per locale. See #getCachedDirectoryList
                if (!directory.isDirectory()) continue;
                final File[] wordLists = directory.listFiles();
                if (null == wordLists) continue;
                for (File wordList : wordLists) {
                    final String fileId = getWordListIdFromFileName(wordList.getName());
                    if (fileId.equals(id)) {
                        if (!canonicalFileToKeep.equals(wordList.getCanonicalFile())) {
                            wordList.delete();
                        }
                    }
                }
            }
        } catch (java.io.IOException e) {
            Log.e(TAG, "IOException trying to cleanup files : " + e);
        }
    }


    /**
     * Returns the id associated with the main word list for a specified locale.
     *
     * Word lists stored in Android Keyboard's resources are referred to as the "main"
     * word lists. Since they can be updated like any other list, we need to assign a
     * unique ID to them. This ID is just the name of the language (locale-wise) they
     * are for, and this method returns this ID.
     */
    private static String getMainDictId(final Locale locale) {
        // This works because we don't include by default different dictionaries for
        // different countries. This actually needs to return the id that we would
        // like to use for word lists included in resources, and the following is okay.
        return MAIN_DICTIONARY_CATEGORY + ID_CATEGORY_SEPARATOR + locale.getLanguage().toString();
    }

    private static boolean isMainWordListId(final String id) {
        final String[] idArray = id.split(ID_CATEGORY_SEPARATOR);
        if (2 != idArray.length) return false;
        return MAIN_DICTIONARY_CATEGORY.equals(idArray[0]);
    }

    // ## HACK ## we prevent usage of a dictionary before version 18 for English only. The reason
    // for this is, since those do not include whitelist entries, the new code with an old version
    // of the dictionary would lose whitelist functionality.
    private static boolean hackCanUseDictionaryFile(final Locale locale, final File f) {
        // Only for English - other languages didn't have a whitelist, hence this
        // ad-hock ## HACK ##
        if (!Locale.ENGLISH.getLanguage().equals(locale.getLanguage())) return true;

        FileInputStream inStream = null;
        try {
            // Read the version of the file
            inStream = new FileInputStream(f);
            final ByteBuffer buffer = inStream.getChannel().map(
                    FileChannel.MapMode.READ_ONLY, 0, f.length());
            final int magic = buffer.getInt();
            if (magic != FormatSpec.VERSION_2_MAGIC_NUMBER) {
                return false;
            }
            final int formatVersion = buffer.getInt();
            final int headerSize = buffer.getInt();
            final HashMap<String, String> options = CollectionUtils.newHashMap();
            BinaryDictInputOutput.populateOptions(buffer, headerSize, options);

            final String version = options.get(VERSION_KEY);
            if (null == version) {
                // No version in the options : the format is unexpected
                return false;
            }
            // Version 18 is the first one to include the whitelist
            // Obviously this is a big ## HACK ##
            return Integer.parseInt(version) >= 18;
        } catch (java.io.FileNotFoundException e) {
            return false;
        } catch (java.io.IOException e) {
            return false;
        } catch (NumberFormatException e) {
            return false;
        } catch (BufferUnderflowException e) {
            return false;
        } finally {
            if (inStream != null) {
                try {
                    inStream.close();
                } catch (IOException e) {
                    // do nothing
                }
            }
        }
    }

    /**
     * Returns a list of file addresses for a given locale, trying relevant methods in order.
     *
     * Tries to get binary dictionaries from various sources, in order:
     * - Uses a content provider to get a public dictionary set, as per the protocol described
     *   in BinaryDictionaryFileDumper.
     * If that fails:
     * - Gets a file name from the built-in dictionary for this locale, if any.
     * If that fails:
     * - Returns null.
     * @return The list of addresses of valid dictionary files, or null.
     */
    public static ArrayList<AssetFileAddress> getDictionaryFiles(final Locale locale,
            final Context context) {

        final boolean hasDefaultWordList = DictionaryFactory.isDictionaryAvailable(context, locale);
        // cacheWordListsFromContentProvider returns the list of files it copied to local
        // storage, but we don't really care about what was copied NOW: what we want is the
        // list of everything we ever cached, so we ignore the return value.
        BinaryDictionaryFileDumper.cacheWordListsFromContentProvider(locale, context,
                hasDefaultWordList);
        final File[] cachedWordLists = getCachedWordLists(locale.toString(), context);
        final String mainDictId = getMainDictId(locale);
        final DictPackSettings dictPackSettings = new DictPackSettings(context);

        boolean foundMainDict = false;
        final ArrayList<AssetFileAddress> fileList = CollectionUtils.newArrayList();
        // cachedWordLists may not be null, see doc for getCachedDictionaryList
        for (final File f : cachedWordLists) {
            final String wordListId = getWordListIdFromFileName(f.getName());
            final boolean canUse = f.canRead() && hackCanUseDictionaryFile(locale, f);
            if (canUse && isMainWordListId(wordListId)) {
                foundMainDict = true;
            }
            if (!dictPackSettings.isWordListActive(wordListId)) continue;
            if (canUse) {
                fileList.add(AssetFileAddress.makeFromFileName(f.getPath()));
            } else {
                Log.e(TAG, "Found a cached dictionary file but cannot read or use it");
            }
        }

        if (!foundMainDict && dictPackSettings.isWordListActive(mainDictId)) {
            final int fallbackResId =
                    DictionaryFactory.getMainDictionaryResourceId(context.getResources(), locale);
            final AssetFileAddress fallbackAsset = loadFallbackResource(context, fallbackResId);
            if (null != fallbackAsset) {
                fileList.add(fallbackAsset);
            }
        }

        return fileList;
    }
}