update from master

[nikiroo-utils.git] / supported / FimfictionApi.java

package be.nikiroo.fanfix.supported;

import java.io.IOException;
import java.io.InputStream;
import java.net.URL;
import java.util.AbstractMap;
import java.util.ArrayList;
import java.util.HashMap;
import java.util.List;
import java.util.Map;
import java.util.Map.Entry;
import java.util.TreeMap;

import org.jsoup.nodes.Document;

import be.nikiroo.fanfix.Instance;
import be.nikiroo.fanfix.bundles.Config;
import be.nikiroo.fanfix.data.MetaData;
import be.nikiroo.fanfix.data.Story;
import be.nikiroo.utils.IOUtils;
import be.nikiroo.utils.Image;
import be.nikiroo.utils.Progress;

/**
 * Support class for <a href="http://www.fimfiction.net/">FimFiction.net</a>
 * stories, a website dedicated to My Little Pony.
 * <p>
 * This version uses the new, official API of FimFiction.
 * 
 * @author niki
 */
class FimfictionApi extends BasicSupport {
	private String oauth;
	private String json;

	private Map<Integer, String> chapterNames;
	private Map<Integer, String> chapterContents;

	public FimfictionApi() throws IOException {
		if (Instance.getInstance().getConfig().getBoolean(Config.LOGIN_FIMFICTION_APIKEY_FORCE_HTML, false)) {
			throw new IOException("Configuration is set to force HTML scrapping");
		}

		String oauth = Instance.getInstance().getConfig().getString(Config.LOGIN_FIMFICTION_APIKEY_TOKEN);

		if (oauth == null || oauth.isEmpty()) {
			String clientId = Instance.getInstance().getConfig().getString(Config.LOGIN_FIMFICTION_APIKEY_CLIENT_ID)
					+ "";
			String clientSecret = Instance.getInstance().getConfig()
					.getString(Config.LOGIN_FIMFICTION_APIKEY_CLIENT_SECRET) + "";

			if (clientId.trim().isEmpty() || clientSecret.trim().isEmpty()) {
				throw new IOException("API key required for the beta API v2");
			}

			oauth = generateOAuth(clientId, clientSecret);

			Instance.getInstance().getConfig().setString(Config.LOGIN_FIMFICTION_APIKEY_TOKEN, oauth);
			Instance.getInstance().getConfig().updateFile();
		}

		this.oauth = oauth;
	}

	@Override
	protected Document loadDocument(URL source) throws IOException {
		json = getJsonData();
		return null;
	}

	@Override
	public String getOAuth() {
		return oauth;
	}

	@Override
	protected boolean isHtml() {
		return true;
	}

	/**
	 * Extract the full JSON data we will later use to build the {@link Story}.
	 * 
	 * @return the data in a JSON format
	 * 
	 * @throws IOException
	 *             in case of I/O error
	 */
	private String getJsonData() throws IOException {
		// extract the ID from:
		// https://www.fimfiction.net/story/123456/name-of-story
		String storyId = getKeyText(getSource().toString(), "/story/", null,
				"/");

		// Selectors, so to download all I need and only what I need
		String storyContent = "fields[story]=title,description,date_published,cover_image";
		String authorContent = "fields[author]=name";
		String chapterContent = "fields[chapter]=chapter_number,title,content_html,authors_note_html";
		String includes = "author,chapters,tags";

		String urlString = String.format(
				"https://www.fimfiction.net/api/v2/stories/%s?" //
						+ "%s&%s&%s&" //
						+ "include=%s", //
				storyId, //
				storyContent, authorContent, chapterContent,//
				includes);

		// URL params must be URL-encoded: "[ ]" <-> "%5B %5D"
		urlString = urlString.replace("[", "%5B").replace("]", "%5D");

		URL url = new URL(urlString);
		InputStream jsonIn = Instance.getInstance().getCache().open(url, this, false);
		try {
			return IOUtils.readSmallStream(jsonIn);
		} finally {
			jsonIn.close();
		}
	}

	@Override
	protected MetaData getMeta() throws IOException {
		MetaData meta = new MetaData();

		meta.setTitle(getKeyJson(json, 0, "type", "story", "title"));
		meta.setAuthor(getKeyJson(json, 0, "type", "user", "name"));
		meta.setDate(bsHelper.formatDate(
				getKeyJson(json, 0, "type", "story", "date_published")));
		meta.setTags(getTags());
		meta.setSource(getType().getSourceName());
		meta.setUrl(getSource().toString());
		meta.setPublisher(getType().getSourceName());
		meta.setUuid(getSource().toString());
		meta.setLuid("");
		meta.setLang("en");
		meta.setSubject("MLP");
		meta.setType(getType().toString());
		meta.setImageDocument(false);

		String coverImageLink = getKeyJson(json, 0, "type", "story",
				"cover_image", "full");
		if (!coverImageLink.trim().isEmpty()) {
			URL coverImageUrl = new URL(coverImageLink.trim());

			// No need to use the oauth, cookies... for the cover
			// Plus: it crashes on Android because of the referer
			try {
				InputStream in = Instance.getInstance().getCache().open(coverImageUrl, null, true);
				try {
					Image img = new Image(in);
					if (img.getSize() == 0) {
						img.close();
						throw new IOException(
								"Empty image not accepted");
					}
					meta.setCover(img);
				} finally {
					in.close();
				}
			} catch (IOException e) {
				Instance.getInstance().getTraceHandler()
						.error(new IOException("Cannot get the story cover, ignoring...", e));
			}
		}

		return meta;
	}

	private List<String> getTags() {
		List<String> tags = new ArrayList<String>();
		tags.add("MLP");

		int pos = 0;
		while (pos >= 0) {
			pos = indexOfJsonAfter(json, pos, "type", "story_tag");
			if (pos >= 0) {
				tags.add(getKeyJson(json, pos, "name").trim());
			}
		}

		return tags;
	}

	@Override
	protected String getDesc() {
		String desc = getKeyJson(json, 0, "type", "story", "description");
		return unbbcode(desc);
	}

	@Override
	protected List<Entry<String, URL>> getChapters(Progress pg) {
		chapterNames = new TreeMap<Integer, String>();
		chapterContents = new TreeMap<Integer, String>();

		int pos = 0;
		while (pos >= 0) {
			pos = indexOfJsonAfter(json, pos, "type", "chapter");
			if (pos >= 0) {
				int posNumber = indexOfJsonAfter(json, pos, "chapter_number");
				int posComa = json.indexOf(",", posNumber);
				final int number = Integer.parseInt(json.substring(posNumber,
						posComa).trim());
				final String title = getKeyJson(json, pos, "title");
				String notes = getKeyJson(json, pos, "authors_note_html");
				String content = getKeyJson(json, pos, "content_html");

				if (!notes.trim().isEmpty()) {
					notes = "<br/>* * *<br/>" + notes;
				}

				chapterNames.put(number, title);
				chapterContents.put(number, content + notes);
			}
		}

		List<Entry<String, URL>> urls = new ArrayList<Entry<String, URL>>();
		for (String title : chapterNames.values()) {
			urls.add(new AbstractMap.SimpleEntry<String, URL>(title, null));
		}

		return urls;
	}

	@Override
	protected String getChapterContent(URL source, int number, Progress pg) {
		return chapterContents.get(number);
	}

	@Override
	protected boolean supports(URL url) {
		return "fimfiction.net".equals(url.getHost())
				|| "www.fimfiction.net".equals(url.getHost());
	}

	/**
	 * Generate a new token from the client ID and secret.
	 * <p>
	 * Note that those tokens are long-lived, and it would be badly seen to
	 * create a lot of them without due cause.
	 * <p>
	 * So, please cache and re-use them.
	 * 
	 * @param clientId
	 *            the client ID offered on FimFiction
	 * @param clientSecret
	 *            the client secret that goes with it
	 * 
	 * @return a new generated token linked to that client ID
	 * 
	 * @throws IOException
	 *             in case of I/O errors
	 */
	static private String generateOAuth(String clientId, String clientSecret)
			throws IOException {
		URL url = new URL("https://www.fimfiction.net/api/v2/token");
		Map<String, String> params = new HashMap<String, String>();
		params.put("client_id", clientId);
		params.put("client_secret", clientSecret);
		params.put("grant_type", "client_credentials");
		InputStream in = Instance.getInstance().getCache().openNoCache(url, null, params, null, null);

		String jsonToken = IOUtils.readSmallStream(in);
		in.close();

		// Extract token type and token from: {
		// token_type = "Bearer",
		// access_token = "xxxxxxxxxxxxxx"
		// }

		String tokenType = getKeyText(jsonToken, "\"token_type\"", "\"", "\"");
		String token = getKeyText(jsonToken, "\"access_token\"", "\"", "\"");

		return tokenType + " " + token;
	}

	// afters: [name, value] pairs (or "" for any of them), can end without
	// value
	static private int indexOfJsonAfter(String json, int startAt,
			String... afterKeys) {
		ArrayList<String> afters = new ArrayList<String>();
		boolean name = true;
		for (String key : afterKeys) {
			if (key != null && !key.isEmpty()) {
				afters.add("\"" + key + "\"");
			} else {
				afters.add("\"");
				afters.add("\"");
			}

			if (name) {
				afters.add(":");
			}

			name = !name;
		}

		return indexOfAfter(json, startAt, afters.toArray(new String[] {}));
	}

	// afters: [name, value] pairs (or "" for any of them), can end without
	// value but will then be empty, not NULL
	static private String getKeyJson(String json, int startAt,
			String... afterKeys) {
		int pos = indexOfJsonAfter(json, startAt, afterKeys);
		if (pos < 0) {
			return "";
		}

		String result = "";
		String wip = json.substring(pos);

		pos = nextUnescapedQuote(wip, 0);
		if (pos >= 0) {
			wip = wip.substring(pos + 1);
			pos = nextUnescapedQuote(wip, 0);
			if (pos >= 0) {
				result = wip.substring(0, pos);
			}
		}

		result = result.replace("\\t", "\t").replace("\\\"", "\"");

		return result;
	}

	// next " but don't take \" into account
	static private int nextUnescapedQuote(String result, int pos) {
		while (pos >= 0) {
			pos = result.indexOf("\"", pos);
			if (pos == 0 || (pos > 0 && result.charAt(pos - 1) != '\\')) {
				break;
			}

			if (pos < result.length()) {
				pos++;
			}
		}

		return pos;
	}

	// quick & dirty filter
	static private String unbbcode(String bbcode) {
		String text = bbcode.replace("\\r\\n", "<br/>") //
				.replace("[i]", "_").replace("[/i]", "_") //
				.replace("[b]", "*").replace("[/b]", "*") //
				.replaceAll("\\[[^\\]]*\\]", "");
		return text;
	}

	/**
	 * Return the text between the key and the endKey (and optional subKey can
	 * be passed, in this case we will look for the key first, then take the
	 * text between the subKey and the endKey).
	 * 
	 * @param in
	 *            the input
	 * @param key
	 *            the key to match (also supports "^" at start to say
	 *            "only if it starts with" the key)
	 * @param subKey
	 *            the sub key or NULL if none
	 * @param endKey
	 *            the end key or NULL for "up to the end"
	 * @return the text or NULL if not found
	 */
	static private String getKeyText(String in, String key, String subKey,
			String endKey) {
		String result = null;

		String line = in;
		if (line != null && line.contains(key)) {
			line = line.substring(line.indexOf(key) + key.length());
			if (subKey == null || subKey.isEmpty() || line.contains(subKey)) {
				if (subKey != null) {
					line = line.substring(line.indexOf(subKey)
							+ subKey.length());
				}
				if (endKey == null || line.contains(endKey)) {
					if (endKey != null) {
						line = line.substring(0, line.indexOf(endKey));
						result = line;
					}
				}
			}
		}

		return result;
	}

	/**
	 * Return the first index after all the given "afters" have been found in
	 * the {@link String}, or -1 if it was not possible.
	 * 
	 * @param in
	 *            the input
	 * @param startAt
	 *            start at this position in the string
	 * @param afters
	 *            the sub-keys to find before checking for key/endKey
	 * 
	 * @return the text or NULL if not found
	 */
	static private int indexOfAfter(String in, int startAt, String... afters) {
		int pos = -1;
		if (in != null && !in.isEmpty()) {
			pos = startAt;
			if (afters != null) {
				for (int i = 0; pos >= 0 && i < afters.length; i++) {
					String subKey = afters[i];
					if (!subKey.isEmpty()) {
						pos = in.indexOf(subKey, pos);
						if (pos >= 0) {
							pos += subKey.length();
						}
					}
				}
			}
		}

		return pos;
	}
}