| 1 | |
package org.paneris.bibliomania.metasearch.amazon; |
| 2 | |
|
| 3 | |
import java.io.ByteArrayOutputStream; |
| 4 | |
import java.io.File; |
| 5 | |
import java.io.FileNotFoundException; |
| 6 | |
import java.io.IOException; |
| 7 | |
import java.net.URL; |
| 8 | |
import java.net.URLConnection; |
| 9 | |
import java.util.Enumeration; |
| 10 | |
|
| 11 | |
import org.melati.util.IoUtils; |
| 12 | |
import org.melati.poem.transaction.ToTidyList; |
| 13 | |
import org.melati.util.UTF8URLEncoder; |
| 14 | |
import org.paneris.bibliomania.BookStocking; |
| 15 | |
import org.paneris.bibliomania.fti.StringDbHash; |
| 16 | |
import org.paneris.bibliomania.metasearch.BookStockingFactory; |
| 17 | |
import org.paneris.bibliomania.metasearch.BookshopBackend; |
| 18 | |
|
| 19 | |
import com.sleepycat.db.DatabaseException; |
| 20 | |
|
| 21 | |
public class Backend implements BookshopBackend { |
| 22 | |
|
| 23 | |
protected final StringDbHash publisherAndImageOfISBN; |
| 24 | |
|
| 25 | |
public Backend(File dbHome, int publisherAndImageOfISBNCacheSize) |
| 26 | 0 | throws FileNotFoundException, DatabaseException { |
| 27 | 0 | publisherAndImageOfISBN = |
| 28 | |
new StringDbHash( |
| 29 | |
new File(dbHome, "amazon_publisherAndImageOfISBN.db").getPath(), |
| 30 | |
publisherAndImageOfISBNCacheSize); |
| 31 | 0 | } |
| 32 | |
|
| 33 | |
public String retrievedPublisherAndImage(BookStocking stocking) |
| 34 | |
throws IOException { |
| 35 | 0 | BookPage bookPage = |
| 36 | |
new BookPage(productPage_nonaffil(stocking), stocking.getIsbn_unsafe()); |
| 37 | |
|
| 38 | 0 | if (bookPage.publisher == null) |
| 39 | 0 | return null; |
| 40 | |
|
| 41 | 0 | return bookPage.image == null |
| 42 | |
? bookPage.publisher |
| 43 | |
: bookPage.publisher + "\u00b7" + bookPage.image; |
| 44 | |
} |
| 45 | |
|
| 46 | |
public void resolve(BookStocking stocking) { |
| 47 | 0 | if (stocking.getPublisher_unsafe() == null) { |
| 48 | 0 | String isbn = stocking.getIsbn_unsafe(); |
| 49 | 0 | String publisherAndImage = publisherAndImageOfISBN.get(isbn); |
| 50 | 0 | if (publisherAndImage == null) { |
| 51 | |
try { |
| 52 | 0 | publisherAndImage = retrievedPublisherAndImage(stocking); |
| 53 | 0 | } catch (IOException e) { |
| 54 | 0 | } |
| 55 | |
|
| 56 | 0 | if (publisherAndImage == null) |
| 57 | 0 | publisherAndImage = ""; |
| 58 | |
|
| 59 | 0 | publisherAndImageOfISBN.put(isbn, publisherAndImage); |
| 60 | 0 | publisherAndImageOfISBN.flush(); |
| 61 | |
} |
| 62 | |
|
| 63 | 0 | int sep = publisherAndImage.indexOf("\u00b7"); |
| 64 | 0 | if (sep == -1) { |
| 65 | 0 | stocking.setPublisher_unsafe(publisherAndImage); |
| 66 | 0 | stocking.setThumbnailurl_unsafe(null); |
| 67 | |
} else { |
| 68 | 0 | stocking.setPublisher_unsafe(publisherAndImage.substring(0, sep)); |
| 69 | 0 | stocking.setThumbnailurl_unsafe(publisherAndImage.substring(sep + 1)); |
| 70 | |
} |
| 71 | |
} |
| 72 | 0 | } |
| 73 | |
|
| 74 | |
public byte[] productPage_nonaffil(BookStocking stocking) |
| 75 | |
throws IOException { |
| 76 | 0 | return IoUtils.slurp(new URL(productURL_nonaffil(stocking)), 32768, 100000); |
| 77 | |
} |
| 78 | |
|
| 79 | |
public String productURL_nonaffil(BookStocking stocking) { |
| 80 | 0 | return "http://www.amazon.com/exec/obidos/ASIN/" |
| 81 | |
+ stocking.getIsbn_unsafe(); |
| 82 | |
} |
| 83 | |
|
| 84 | |
public String productURL(BookStocking stocking) { |
| 85 | 0 | return productURL_nonaffil(stocking); |
| 86 | |
} |
| 87 | |
|
| 88 | |
public String notifiedOfferedLinkHTML(BookStocking stocking) { |
| 89 | 0 | return ""; |
| 90 | |
} |
| 91 | |
|
| 92 | 0 | byte[] searchQueryPrefixBytes = "index=books&field-power=".getBytes(); |
| 93 | 0 | byte[] authorBytes = "author:+".getBytes(); |
| 94 | 0 | byte[] titleBytes = "title:+".getBytes(); |
| 95 | 0 | byte[] andBytes = "+AND+".getBytes(); |
| 96 | |
|
| 97 | |
public Enumeration booksMatching( |
| 98 | |
BookStockingFactory stockings, |
| 99 | |
String title, |
| 100 | |
String author, |
| 101 | |
String publisher, |
| 102 | |
String keyword, |
| 103 | |
ToTidyList toTidy) |
| 104 | |
throws IOException { |
| 105 | |
|
| 106 | 0 | URLConnection c = |
| 107 | |
new URL( |
| 108 | |
"http://www.amazon.com/exec/obidos/search-handle-form/" |
| 109 | |
+ "ref=s_sf_b_ps/?") |
| 110 | |
.openConnection(); |
| 111 | |
|
| 112 | 0 | ByteArrayOutputStream buf = new ByteArrayOutputStream(100); |
| 113 | 0 | buf.write(searchQueryPrefixBytes); |
| 114 | |
|
| 115 | 0 | if (author != null && !author.trim().equals("")) { |
| 116 | 0 | buf.write(authorBytes); |
| 117 | 0 | buf.write(UTF8URLEncoder.encode(author).getBytes()); |
| 118 | |
} |
| 119 | 0 | if (title != null && !title.trim().equals("")) { |
| 120 | 0 | if (author != null && !author.trim().equals("")) |
| 121 | 0 | buf.write(andBytes); |
| 122 | 0 | buf.write(titleBytes); |
| 123 | 0 | buf.write(UTF8URLEncoder.encode(title).getBytes()); |
| 124 | |
} |
| 125 | |
|
| 126 | 0 | c.setDoOutput(true); |
| 127 | 0 | buf.writeTo(c.getOutputStream()); |
| 128 | 0 | buf.writeTo(System.err); |
| 129 | 0 | return new SearchResults( |
| 130 | |
IoUtils.slurp(c.getInputStream(), 65536), |
| 131 | |
stockings); |
| 132 | |
} |
| 133 | |
} |