1/*
2 * Copyright (C) 2015 The Android Open Source Project
3 *
4 * Licensed under the Apache License, Version 2.0 (the "License");
5 * you may not use this file except in compliance with the License.
6 * You may obtain a copy of the License at
7 *
8 *      http://www.apache.org/licenses/LICENSE-2.0
9 *
10 * Unless required by applicable law or agreed to in writing, software
11 * distributed under the License is distributed on an "AS IS" BASIS,
12 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13 * See the License for the specific language governing permissions and
14 * limitations under the License.
15 */
16
17package com.android.statementservice.retriever;
18
19import android.util.Log;
20
21import com.android.volley.Cache;
22import com.android.volley.NetworkResponse;
23import com.android.volley.toolbox.HttpHeaderParser;
24
25import java.io.BufferedInputStream;
26import java.io.ByteArrayOutputStream;
27import java.io.IOException;
28import java.io.InputStream;
29import java.net.HttpURLConnection;
30import java.net.URL;
31import java.util.HashMap;
32import java.util.List;
33import java.util.Locale;
34import java.util.Map;
35
36/**
37 * Helper class for fetching HTTP or HTTPS URL.
38 *
39 * Visible for testing.
40 *
41 * @hide
42 */
43public class URLFetcher {
44    private static final String TAG = URLFetcher.class.getSimpleName();
45
46    private static final long DO_NOT_CACHE_RESULT = 0L;
47    private static final int INPUT_BUFFER_SIZE_IN_BYTES = 1024;
48
49    /**
50     * Fetches the specified url and returns the content and ttl.
51     *
52     * <p>
53     * Retry {@code retry} times if the connection failed or timed out for any reason.
54     * HTTP error code (e.g. 404/500) won't be retried.
55     *
56     * @throws IOException if it can't retrieve the content due to a network problem.
57     * @throws AssociationServiceException if the URL scheme is not http or https or the content
58     * length exceeds {code fileSizeLimit}.
59     */
60    public WebContent getWebContentFromUrlWithRetry(URL url, long fileSizeLimit,
61            int connectionTimeoutMillis, int backoffMillis, int retry)
62                    throws AssociationServiceException, IOException, InterruptedException {
63        if (retry <= 0) {
64            throw new IllegalArgumentException("retry should be a postive inetger.");
65        }
66        while (retry > 0) {
67            try {
68                return getWebContentFromUrl(url, fileSizeLimit, connectionTimeoutMillis);
69            } catch (IOException e) {
70                retry--;
71                if (retry == 0) {
72                    throw e;
73                }
74            }
75
76            Thread.sleep(backoffMillis);
77        }
78
79        // Should never reach here.
80        return null;
81    }
82
83    /**
84     * Fetches the specified url and returns the content and ttl.
85     *
86     * @throws IOException if it can't retrieve the content due to a network problem.
87     * @throws AssociationServiceException if the URL scheme is not http or https or the content
88     * length exceeds {code fileSizeLimit}.
89     */
90    public WebContent getWebContentFromUrl(URL url, long fileSizeLimit, int connectionTimeoutMillis)
91            throws AssociationServiceException, IOException {
92        final String scheme = url.getProtocol().toLowerCase(Locale.US);
93        if (!scheme.equals("http") && !scheme.equals("https")) {
94            throw new IllegalArgumentException("The url protocol should be on http or https.");
95        }
96
97        HttpURLConnection connection = null;
98        try {
99            connection = (HttpURLConnection) url.openConnection();
100            connection.setInstanceFollowRedirects(true);
101            connection.setConnectTimeout(connectionTimeoutMillis);
102            connection.setReadTimeout(connectionTimeoutMillis);
103            connection.setUseCaches(true);
104            connection.setInstanceFollowRedirects(false);
105            connection.addRequestProperty("Cache-Control", "max-stale=60");
106
107            if (connection.getResponseCode() != HttpURLConnection.HTTP_OK) {
108                Log.e(TAG, "The responses code is not 200 but "  + connection.getResponseCode());
109                return new WebContent("", DO_NOT_CACHE_RESULT);
110            }
111
112            if (connection.getContentLength() > fileSizeLimit) {
113                Log.e(TAG, "The content size of the url is larger than "  + fileSizeLimit);
114                return new WebContent("", DO_NOT_CACHE_RESULT);
115            }
116
117            Long expireTimeMillis = getExpirationTimeMillisFromHTTPHeader(
118                    connection.getHeaderFields());
119
120            return new WebContent(inputStreamToString(
121                    connection.getInputStream(), connection.getContentLength(), fileSizeLimit),
122                expireTimeMillis);
123        } finally {
124            if (connection != null) {
125                connection.disconnect();
126            }
127        }
128    }
129
130    /**
131     * Visible for testing.
132     * @hide
133     */
134    public static String inputStreamToString(InputStream inputStream, int length, long sizeLimit)
135            throws IOException, AssociationServiceException {
136        if (length < 0) {
137            length = 0;
138        }
139        ByteArrayOutputStream baos = new ByteArrayOutputStream(length);
140        BufferedInputStream bis = new BufferedInputStream(inputStream);
141        byte[] buffer = new byte[INPUT_BUFFER_SIZE_IN_BYTES];
142        int len = 0;
143        while ((len = bis.read(buffer)) != -1) {
144            baos.write(buffer, 0, len);
145            if (baos.size() > sizeLimit) {
146                throw new AssociationServiceException("The content size of the url is larger than "
147                        + sizeLimit);
148            }
149        }
150        return baos.toString("UTF-8");
151    }
152
153    /**
154     * Parses the HTTP headers to compute the ttl.
155     *
156     * @param headers a map that map the header key to the header values. Can be null.
157     * @return the ttl in millisecond or null if the ttl is not specified in the header.
158     */
159    private Long getExpirationTimeMillisFromHTTPHeader(Map<String, List<String>> headers) {
160        if (headers == null) {
161            return null;
162        }
163        Map<String, String> joinedHeaders = joinHttpHeaders(headers);
164
165        NetworkResponse response = new NetworkResponse(null, joinedHeaders);
166        Cache.Entry cachePolicy = HttpHeaderParser.parseCacheHeaders(response);
167
168        if (cachePolicy == null) {
169            // Cache is disabled, set the expire time to 0.
170            return DO_NOT_CACHE_RESULT;
171        } else if (cachePolicy.ttl == 0) {
172            // Cache policy is not specified, set the expire time to 0.
173            return DO_NOT_CACHE_RESULT;
174        } else {
175            // cachePolicy.ttl is actually the expire timestamp in millisecond.
176            return cachePolicy.ttl;
177        }
178    }
179
180    /**
181     * Converts an HTTP header map of the format provided by {@linkHttpUrlConnection} to a map of
182     * the format accepted by {@link HttpHeaderParser}. It does this by joining all the entries for
183     * a given header key with ", ".
184     */
185    private Map<String, String> joinHttpHeaders(Map<String, List<String>> headers) {
186        Map<String, String> joinedHeaders = new HashMap<String, String>();
187        for (Map.Entry<String, List<String>> entry : headers.entrySet()) {
188            List<String> values = entry.getValue();
189            if (values.size() == 1) {
190                joinedHeaders.put(entry.getKey(), values.get(0));
191            } else {
192                joinedHeaders.put(entry.getKey(), Utils.joinStrings(", ", values));
193            }
194        }
195        return joinedHeaders;
196    }
197}
198