X-Git-Url: http://git.nikiroo.be/?a=blobdiff_plain;f=src%2Fbe%2Fnikiroo%2Ffanfix%2Fsupported%2FE621.java;h=aa3084f210cf99140d53d3878f326de46172d6f2;hb=1822d603af35b758310e182e50103f0a2bffbc85;hp=05d122e68b89efc71587fda628f9b17e720736d7;hpb=406447a4c017f2a4592429a67d546a6984460617;p=nikiroo-utils.git diff --git a/src/be/nikiroo/fanfix/supported/E621.java b/src/be/nikiroo/fanfix/supported/E621.java index 05d122e..aa3084f 100644 --- a/src/be/nikiroo/fanfix/supported/E621.java +++ b/src/be/nikiroo/fanfix/supported/E621.java @@ -1,11 +1,19 @@ package be.nikiroo.fanfix.supported; -import java.awt.image.BufferedImage; import java.io.IOException; import java.io.InputStream; +import java.io.UnsupportedEncodingException; +import java.net.MalformedURLException; import java.net.URL; +import java.net.URLDecoder; +import java.net.URLEncoder; +import java.util.AbstractMap; import java.util.ArrayList; +import java.util.Collection; +import java.util.Collections; +import java.util.LinkedList; import java.util.List; +import java.util.AbstractMap.SimpleEntry; import java.util.Map.Entry; import java.util.Scanner; @@ -13,6 +21,7 @@ import be.nikiroo.fanfix.Instance; import be.nikiroo.fanfix.data.Chapter; import be.nikiroo.fanfix.data.MetaData; import be.nikiroo.fanfix.data.Story; +import be.nikiroo.utils.Image; import be.nikiroo.utils.Progress; import be.nikiroo.utils.StringUtils; @@ -26,7 +35,7 @@ import be.nikiroo.utils.StringUtils; * * @author niki */ -class E621 extends BasicSupport { +class E621 extends BasicSupport_Deprecated { @Override public String getSourceName() { return "e621.net"; @@ -39,22 +48,58 @@ class E621 extends BasicSupport { meta.setTitle(getTitle(reset(in))); meta.setAuthor(getAuthor(source, reset(in))); meta.setDate(""); - meta.setTags(new ArrayList()); // TODDO ??? + meta.setTags(getTags(source, reset(in), false)); meta.setSource(getSourceName()); meta.setUrl(source.toString()); meta.setPublisher(getSourceName()); meta.setUuid(source.toString()); meta.setLuid(""); - meta.setLang("EN"); + meta.setLang("en"); meta.setSubject("Furry"); meta.setType(getType().toString()); meta.setImageDocument(true); - meta.setCover(getCover(source)); + meta.setCover(getCover(source, reset(in))); meta.setFakeCover(true); return meta; } + private List getTags(URL source, InputStream in, boolean authors) { + List tags = new ArrayList(); + + if (isSearch(source)) { + String tagLine = getLine(in, "id=\"tag-sidebar\"", 1); + if (tagLine != null) { + String key = "href=\""; + for (int pos = tagLine.indexOf(key); pos >= 0; pos = tagLine + .indexOf(key, pos + 1)) { + int end = tagLine.indexOf("\"", pos + key.length()); + if (end >= 0) { + String href = tagLine.substring(pos, end); + String subkey; + if (authors) + subkey = "?name="; + else + subkey = "?title="; + if (href.contains(subkey)) { + String tag = href.substring(href.indexOf(subkey) + + subkey.length()); + try { + tags.add(URLDecoder.decode(tag, "UTF-8")); + } catch (UnsupportedEncodingException e) { + // supported JVMs must have UTF-8 support + e.printStackTrace(); + } + } + } + } + + } + } + + return tags; + } + @Override public Story process(URL url, Progress pg) throws IOException { // There is no chapters on e621, just pagination... @@ -79,7 +124,7 @@ class E621 extends BasicSupport { } return ("e621.net".equals(host) || "e926.net".equals(host)) - && url.getPath().startsWith("/pool/"); + && (isPool(url) || isSearch(url)); } @Override @@ -87,10 +132,13 @@ class E621 extends BasicSupport { return true; } - private BufferedImage getCover(URL source) throws IOException { - InputStream in = Instance.getCache().open(source, this, true); - String images = getChapterContent(new URL(source.toString() + "?page=" - + 1), in, 1); + private Image getCover(URL source, InputStream in) throws IOException { + URL urlForCover = source; + if (isPool(source)) { + urlForCover = new URL(source.toString() + "?page=1"); + } + + String images = getChapterContent(urlForCover, in, 1, null); if (!images.isEmpty()) { int pos = images.indexOf("
"); if (pos >= 0) { @@ -102,7 +150,18 @@ class E621 extends BasicSupport { return null; } - private String getAuthor(URL source, InputStream in) throws IOException { + private String getAuthor(URL source, InputStream in) { + if (isSearch(source)) { + StringBuilder builder = new StringBuilder(); + for (String author : getTags(source, in, true)) { + if (builder.length() > 0) + builder.append(", "); + builder.append(author); + } + + return builder.toString(); + } + String author = getLine(in, "href=\"/post/show/", 0); if (author != null) { String key = "href=\""; @@ -144,7 +203,7 @@ class E621 extends BasicSupport { return null; } - private String getTitle(InputStream in) throws IOException { + private String getTitle(InputStream in) { String title = getLine(in, "", 0); if (title != null) { int pos = title.indexOf('>'); @@ -191,8 +250,60 @@ class E621 extends BasicSupport { } @Override - protected List<Entry<String, URL>> getChapters(URL source, InputStream in) - throws IOException { + protected List<Entry<String, URL>> getChapters(URL source, InputStream in, + Progress pg) throws IOException { + if (isPool(source)) { + return getChaptersPool(source, in, pg); + } else if (isSearch(source)) { + return getChaptersSearch(source, in, pg); + } + + return new LinkedList<Entry<String, URL>>(); + } + + private List<Entry<String, URL>> getChaptersSearch(URL source, + InputStream in, Progress pg) throws IOException { + List<Entry<String, URL>> urls = new ArrayList<Entry<String, URL>>(); + + String search = source.getPath(); + if (search.endsWith("/")) { + search = search.substring(0, search.length() - 1); + } + + int pos = search.lastIndexOf('/'); + if (pos >= 0) { + search = search.substring(pos + 1); + } + + String baseUrl = "https://e621.net/post/index/"; + if (source.getHost().contains("e926")) { + baseUrl = baseUrl.replace("e621", "e926"); + } + + for (int i = 1; true; i++) { + URL url = new URL(baseUrl + i + "/" + search + "/"); + try { + InputStream pageI = Instance.getCache().open(url, this, false); + try { + if (getLine(pageI, "No posts matched your search.", 0) != null) + break; + urls.add(new AbstractMap.SimpleEntry<String, URL>("Page " + + Integer.toString(i), url)); + } finally { + pageI.close(); + } + } catch (Exception e) { + break; + } + } + + // They are sorted in reverse order on the website + Collections.reverse(urls); + return urls; + } + + private List<Entry<String, URL>> getChaptersPool(URL source, + InputStream in, Progress pg) throws IOException { List<Entry<String, URL>> urls = new ArrayList<Entry<String, URL>>(); int last = 1; // no pool/show when only one page @@ -219,29 +330,16 @@ class E621 extends BasicSupport { } for (int i = 1; i <= last; i++) { - final String key = Integer.toString(i); - final URL value = new URL(source.toString() + "?page=" + i); - urls.add(new Entry<String, URL>() { - public URL setValue(URL value) { - return null; - } - - public URL getValue() { - return value; - } - - public String getKey() { - return key; - } - }); + urls.add(new AbstractMap.SimpleEntry<String, URL>(Integer + .toString(i), new URL(source.toString() + "?page=" + i))); } return urls; } @Override - protected String getChapterContent(URL source, InputStream in, int number) - throws IOException { + protected String getChapterContent(URL source, InputStream in, int number, + Progress pg) throws IOException { StringBuilder builder = new StringBuilder(); String staticSite = "https://static1.e621.net"; if (source.getHost().contains("e926")) { @@ -277,4 +375,34 @@ class E621 extends BasicSupport { return builder.toString(); } + + @Override + protected URL getCanonicalUrl(URL source) { + if (isSearch(source)) { + // /post?tags=tag1+tag2 -> ../post/index/1/tag1%32tag2 + String key = "post?tags="; + if (source.toString().contains(key)) { + int pos = source.toString().indexOf(key); + String tags = source.toString().substring(pos + key.length()); + tags = tags.replace("+", "%20"); + try { + return new URL(source.toString().substring(0, pos) + + "post/index/1/" + tags); + } catch (MalformedURLException e) { + Instance.getTraceHandler().error(e); + } + } + } + return super.getCanonicalUrl(source); + } + + private boolean isPool(URL url) { + return url.getPath().startsWith("/pool/"); + } + + private boolean isSearch(URL url) { + return url.getPath().startsWith("/post/index/") + || (url.getPath().equals("/post") && url.getQuery().startsWith( + "tags=")); + } }