package org.greenstone.gsdl3.core;

import java.io.File;
import java.io.IOException;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.HashMap;
import java.util.Map;

import javax.servlet.Filter;
import javax.servlet.FilterChain;
import javax.servlet.FilterConfig;
import javax.servlet.ServletContext;
import javax.servlet.ServletException;
import javax.servlet.ServletOutputStream;
import javax.servlet.ServletRequest;
import javax.servlet.ServletResponse;
import javax.servlet.http.HttpSession;
import javax.servlet.http.HttpServletRequest;
import javax.servlet.http.HttpServletRequestWrapper;
import javax.servlet.http.HttpServletResponse;

import org.apache.commons.io.FileUtils;
import org.apache.commons.lang3.StringUtils;

import org.apache.log4j.Logger;
import org.greenstone.gsdl3.util.GSParams;
import org.greenstone.gsdl3.util.GSPath;
import org.greenstone.gsdl3.util.GSXML;
import org.greenstone.gsdl3.util.UserContext;
import org.greenstone.gsdl3.util.XMLConverter;
import org.greenstone.gsdl3.service.Authentication;
import org.w3c.dom.Document;
import org.w3c.dom.Element;
import org.w3c.dom.NodeList;

public class URLFilter implements Filter
{
  private FilterConfig _filterConfig = null;
  private static Logger _logger = Logger.getLogger(org.greenstone.gsdl3.core.URLFilter.class.getName());

  //Restricted URLs
  protected static final String SITECONFIG_URL = "sites/[^/]+/siteConfig.xml";
  protected static final String USERS_DB_URL = "etc/usersDB/.*";
  protected static final ArrayList<String> _restrictedURLs;
  static
  {
    ArrayList<String> restrictedURLs = new ArrayList<String>();
    restrictedURLs.add(SITECONFIG_URL);
    restrictedURLs.add(USERS_DB_URL);
    _restrictedURLs = restrictedURLs;
  }

  //Constants
  protected static final String DOCUMENT_PATH = "document";
  protected static final String COLLECTION_PATH = "collection";
  protected static final String GROUP_PATH = "group";
  protected static final String PAGE_PATH = "page";
  protected static final String SYSTEM_PATH = "system";
  protected static final String BROWSE_PATH = "browse";
  protected static final String SEARCH_PATH = "search";

  protected static final String METADATA_RETRIEVAL_SERVICE = "DocumentMetadataRetrieve";
  protected static final String ASSOCIATED_FILE_PATH = "/index/assoc/";
  protected static final String COLLECTION_FILE_PATH = "/collect/";
  protected static final String INTERFACE_PATH = "/interfaces/";
  protected static final String SITES_PATH = "/sites/";
  
  protected static final String SYSTEM_SUBACTION_CONFIGURE = "configure";
  protected static final String SYSTEM_SUBACTION_RECONFIGURE = "reconfigure";
  protected static final String SYSTEM_SUBACTION_ACTIVATE = "activate";
  protected static final String SYSTEM_SUBACTION_DEACTIVATE = "deactivate";

  public void init(FilterConfig filterConfig) throws ServletException
  {
    this._filterConfig = filterConfig;
  }

  public void destroy()
  {
    this._filterConfig = null;
  }

  @SuppressWarnings("deprecation")
  public void doFilter(ServletRequest request, ServletResponse response, FilterChain chain) throws IOException, ServletException
  {
    if (!(request instanceof HttpServletRequest)) {
      // Can this ever happen?
      _logger.error("The request was not an HttpServletRequest");
      return;
    }
	  
	
    HttpServletRequest hRequest = ((HttpServletRequest) request);
    HttpSession hSession = hRequest.getSession();
    ServletContext context = hSession.getServletContext();
	
    GSHttpServletRequestWrapper gRequest = new GSHttpServletRequestWrapper(hRequest);
	  
    // this is the part before the ?
    String url = hRequest.getRequestURI().toString();
    if (isURLRestricted(url)) {
      
      // TODO - should we make this a proper HTML page?
      response.getWriter().println("ERROR: Access to this page is forbidden.");
      return;
    }


    // Run security checks on files requested from a collection's index/assoc folder
    if (url.contains(ASSOCIATED_FILE_PATH)) {
      
	// now we need to get library name from the path, which is like
	// /greenstone3/library/sites/localsite/collect/collname/index/assoc/...
	String library_name = url.replaceFirst(context.getContextPath(), "");
	library_name = library_name.substring(0, library_name.indexOf(SITES_PATH));
	if (library_name.equals("")) {
	  response.getWriter().println("ERROR: Assoc file paths must now contain the library name");
	  return;
	} 
	// remove initial '/'
	library_name = library_name.substring(1);

	MessageRouter gsRouter = (MessageRouter) context.getAttribute(library_name+"Router");
				
	if (gsRouter == null) {  
	  _logger.error("Receptionist is null, stopping filter");
	  return;
	}
	// Sometimes we have a // before the filename - that mucks up the following code, so lets remove them
	url = url.replaceAll("//","/");
	String dir = null;
	int dirStart = url.indexOf(ASSOCIATED_FILE_PATH) + ASSOCIATED_FILE_PATH.length();
	int dirEnd = -1;
	if (dirStart < url.length() && url.indexOf("/", dirStart) != -1)
	  {
	    dirEnd = url.lastIndexOf("/");
	  }
	if (dirEnd != -1)
	  {
	    dir = url.substring(dirStart, dirEnd);
	  }
	if (dir == null)
	  {
	    return;
	  }
	
	String collection = null;
	int colStart = url.indexOf(COLLECTION_FILE_PATH) + COLLECTION_FILE_PATH.length();
	int colEnd = -1;
	if (colStart < url.length() && url.indexOf("/", colStart) != -1)
	  {
	    colEnd = url.indexOf("/", colStart);
	  }
	if (colEnd != -1)
	  {
	    collection = url.substring(colStart, colEnd);
	  }
	if (collection == null)
	  {
	    return;
	  }
	
	String file_name = url.substring(url.lastIndexOf("/")+1);

	

	// Query the MR with a request for the contains metadata for node "dir" - where dir is the assocfilepath
	// In the jdbm db, have entries like
	// [HASH1552e]
	// <contains>HASH1552e3sdlkjf7sdfsdfk
	// mapping assocfilepath to doc id
	String document = queryMRforDOCID(gsRouter, collection, dir);
	if (document == null) {
	  response.getWriter().println("ERROR: Couldn't find the document associated with assocfilepath: "+dir);
	  return;
	}
	
	//Query the MR for the security info for this document - can we show it? Or do we need to be logged in?
	// Or do we need to throw up the verify page?
	
	// While we are doing this, query the document for its srclinkFile metadata - then we can determine if the
	// file we are being asked for is the main doc (eg pdf) or just a supporting image on the page

	//Get the security info for this collection
	Document gsDoc = XMLConverter.newDOM();
	Element securityMessage = gsDoc.createElement(GSXML.MESSAGE_ELEM);
	Element securityRequest = GSXML.createBasicRequest(gsDoc, GSXML.REQUEST_TYPE_SECURITY, collection, new UserContext());
					
	securityMessage.appendChild(securityRequest);
	securityRequest.setAttribute(GSXML.NODE_OID, document);
					  
	// get the srclinkFile for the document
	Element metadata_request = GSXML.createBasicRequest(gsDoc, GSXML.REQUEST_TYPE_PROCESS, GSPath.appendLink(collection, "DocumentMetadataRetrieve"), new UserContext());
	Element param_list = gsDoc.createElement(GSXML.PARAM_ELEM + GSXML.LIST_MODIFIER);
	GSXML.addParameterToList(param_list, "metadata", "srclinkFile");
	metadata_request.appendChild(param_list);
	Element doc_list = gsDoc.createElement(GSXML.DOC_NODE_ELEM + GSXML.LIST_MODIFIER);
	metadata_request.appendChild(doc_list);
	Element d = gsDoc.createElement(GSXML.DOC_NODE_ELEM);
	d.setAttribute(GSXML.NODE_ID_ATT, document);
	doc_list.appendChild(d);
	securityMessage.appendChild(metadata_request);
	      
					
	Element mr_response = (Element)gsRouter.process(securityMessage);
	_logger.debug("security response = "+XMLConverter.getPrettyString(mr_response));

	boolean verifiable_file = true;
	// TODO check for errors

	Element meta_response = (Element) GSXML.getNamedElement(mr_response, GSXML.RESPONSE_ELEM, GSXML.TYPE_ATT, GSXML.REQUEST_TYPE_PROCESS);
	Element metadata_list = (Element)meta_response.getElementsByTagName(GSXML.METADATA_ELEM+GSXML.LIST_MODIFIER).item(0);
	String srcdoc = GSXML.getMetadataValue(metadata_list, "srclinkFile");
	if (!srcdoc.equals(file_name)) {
	  // the specified file is just a supporting file, not the main file.
	  // eg an image in an html doc.
	  verifiable_file = false;
	}

	Element securityResponse = (Element) GSXML.getNamedElement(mr_response, GSXML.RESPONSE_ELEM, GSXML.TYPE_ATT, GSXML.REQUEST_TYPE_SECURITY);
	ArrayList<String> groups = GSXML.getGroupsFromSecurityResponse(securityResponse);
				
	if (!groups.contains(""))
	  {
	    boolean found = false;
	    for (String group : groups)
	      {
		if (((HttpServletRequest) request).isUserInRole(group))
		  {
		    found = true;
		    break;
		  }
	      }

	    if (!found)
	      {
		// return an error page to the browser
		String new_url = context.getContextPath()+"/"+ library_name+"?a=p&sa=error&c="+collection+"&ec=wrong_group";			  
		((HttpServletResponse)response).sendRedirect(new_url);
		return;
	      }
	  }
	// if got here have no groups.
	// do we have human verify thing?
	if (verifiable_file) {
	  // we are asking for the main document - lets check human verify
					  
	  if (!securityResponse.getAttribute(GSXML.VERIFY_ATT).equals("")) {
	    // have we done the test previously?
	    HttpSession this_session =  ((HttpServletRequest) request).getSession();
	    if (this_session == null) {
	      _logger.error("KATH session is null");
	    } else {
	      _logger.error("KATH session id = "+this_session.getId());
	    }
	    if (this_session.getAttribute(GSParams.VERIFIED) != null ) {
	      _logger.error("KATH have verified in the session");
	      // we don't need to re-verify
	    } else {
	      _logger.error("KATH verfied not in session");
					      
	      // have we just  done the test?
	      String hmvf_response = gRequest.getParameter(GSParams.VERIFIED);
	      // hmvf param will be set by form if the verify page was submitted
	      if (hmvf_response != null && hmvf_response.equals("1")) {
		if (!securityResponse.getAttribute(GSXML.SITE_KEY_ATT).equals("")) {
		  String recaptcha_response = gRequest.getParameter(Authentication.RECAPTCHA_RESPONSE_PARAM);
		  String secret_key = securityResponse.getAttribute(GSXML.SECRET_KEY_ATT);
		  int result = Authentication.verifyRecaptcha(secret_key, recaptcha_response);
		  _logger.debug("recaptcha result code = "+result);
		  if (result == Authentication.NO_ERROR) {
		    _logger.debug("RECAPTCHA SUCCESS, hopefully going to the document");

		    this_session.setAttribute(GSParams.VERIFIED, "1");
		  } else {
		    _logger.error("something went wrong with recaptcha, error="+result);
		    _logger.error(Authentication.getErrorKey(result));
		    // display error page
		    //String new_url = context.getContextPath()+"/"+ context.getAttribute("LibraryName")+"?a=p&sa=error&c="+collection+"&ec=recap_fail";
		    String new_url = context.getContextPath()+"/"+ library_name+"?a=p&sa=error&c="+collection+"&ec=recap_fail";				
		    ((HttpServletResponse)response).sendRedirect(new_url);
						
		    return;
		  }
		}
					    
	      } else {
		// hmvf param is not set - we haven't shown them the form yet
		// we need to display the verify page
		//String new_url = context.getContextPath()+"/"+ context.getAttribute("LibraryName")+"?a=p&sa=verify&c="+collection+"&url="+url;
		String new_url = context.getContextPath()+"/"+ library_name+"?a=p&sa=verify&c="+collection+"&url="+url;			    
		((HttpServletResponse)response).sendRedirect(new_url);
		return;
	      }
	    }
	  }
	}
    
			
	// if we got here, we have passed all security checks and just want to view the file.
	// However, we need to remove the library_name from the URL. As can't change the
	// existing URL, we need to forward to the new one.
	// Remove the context and library name parts.
	// don't know what happens with the rest of the filter chain? Does this bypass that??
	url = url.replaceFirst(context.getContextPath(), "");
	url = url.replaceFirst("/"+library_name, "");
	request.getRequestDispatcher(url).forward(request, response);

	return;
    }
    else if (url.contains(INTERFACE_PATH))
      {			    
	String fileURL = url.replaceFirst(context.getContextPath(), "");
	File requestedFile = new File(context.getRealPath(fileURL));

	if (!requestedFile.exists())
	  {
	    int interfaceNameStart = fileURL.indexOf(INTERFACE_PATH) + INTERFACE_PATH.length();
	    int interfaceNameEnd = fileURL.indexOf("/", interfaceNameStart);
	    String interfaceName = fileURL.substring(interfaceNameStart, interfaceNameEnd);
	    String interfacesDir = fileURL.substring(0, interfaceNameStart);
	    File interfaceConfigFile = new File(context.getRealPath(interfacesDir + interfaceName + "/interfaceConfig.xml"));

	    if (interfaceConfigFile.exists())
	      {
		Document interfaceConfigDoc = XMLConverter.getDOM(interfaceConfigFile);

		String baseInterface = interfaceConfigDoc.getDocumentElement().getAttribute("baseInterface");
		if (baseInterface.length() > 0)
		  {
		    File baseInterfaceFile = new File(context.getRealPath(fileURL.replace("/" + interfaceName + "/", "/" + baseInterface + "/")));
		    if (baseInterfaceFile.exists())
		      {
			ServletOutputStream out = response.getOutputStream();
			out.write(FileUtils.readFileToByteArray(baseInterfaceFile));
			out.flush();
			out.close();
			return;
		      }
		  }
	      }
	  }
      }
    else
      {
	ArrayList<String> keywords = new ArrayList<String>();
	keywords.add(PAGE_PATH);
	keywords.add(BROWSE_PATH);
	keywords.add(SEARCH_PATH);
	keywords.add(DOCUMENT_PATH);
	//If we have a jsessionid on the end of our URL we want to ignore it
	int index;
	if ((index = url.indexOf(";jsessionid")) != -1)
	  {
	    url = url.substring(0, index);
	  }
	String[] segments = url.split("/");
	for (int i = 0; i < segments.length; i++)
	  {
	    String[] additionalParameters = null;
	    String[] defaultParamValues = null;
	    //COLLECTION
	    if (segments[i].equals(COLLECTION_PATH) && (i + 1) < segments.length) {
	      int j=i+1;
	      while(j+1 < segments.length && !keywords.contains(segments[j+1])) {
		j++;
	      }
					  
	      if (j>i+1) {
		// we had a group part
		String [] groups = Arrays.copyOfRange(segments, i+1, j);
		String group = StringUtils.join(groups, "/");
		gRequest.setParameter(GSParams.GROUP, group);
	      }
	      gRequest.setParameter(GSParams.COLLECTION, segments[j]);
	    }
	    // GROUP
	    else if(segments[i].equals(GROUP_PATH) && (i + 1) < segments.length)
	      {
		// assume for now, no other path parts for group links
		int j= segments.length - 1;
		String group;
		if (j==i+1) {
		  group = segments[j];
		} else {
		  String [] groups = Arrays.copyOfRange(segments, i+1, j+1);
		  group = StringUtils.join(groups, "/");
		}
		gRequest.setParameter(GSParams.GROUP, group);
		gRequest.setParameter(GSParams.ACTION, "p");
		gRequest.setParameter(GSParams.SUBACTION, "home");
					    
	      }
	    //DOCUMENT
	    else if (segments[i].equals(DOCUMENT_PATH) && (i + 1) < segments.length)
	      {
		gRequest.setParameter(GSParams.DOCUMENT, segments[i + 1]);
						
		additionalParameters = new String[] { GSParams.ACTION };
		defaultParamValues = new String[] { "d" };
		if ((i+2) < segments.length && segments[i+2].equals("print")) {
		  gRequest.setParameter(GSParams.SUBACTION, "print");
		  gRequest.setParameter("ed", "1");
						  
		}
					
	      }
	    //PAGE
	    else if (segments[i].equals(PAGE_PATH) && (i + 1) < segments.length)
	      {
		gRequest.setParameter(GSParams.SUBACTION, segments[i + 1]);

		additionalParameters = new String[] { GSParams.ACTION };
		defaultParamValues = new String[] { "p" };
	      }
	    //SYSTEM
	    else if (segments[i].equals(SYSTEM_PATH) && (i + 1) < segments.length)
	      {
		String sa = segments[i + 1];
		if (sa.equals(SYSTEM_SUBACTION_CONFIGURE) || sa.equals(SYSTEM_SUBACTION_RECONFIGURE))
		  {
		    sa = "c";
		  }
		else if (sa.equals(SYSTEM_SUBACTION_ACTIVATE))
		  {
		    sa = "a";
		  }
		else if (sa.equals(SYSTEM_SUBACTION_DEACTIVATE))
		  {
		    sa = "d";
		  }

		if (sa.equals("c") && (i + 2) < segments.length)
		  {
		    gRequest.setParameter(GSParams.SYSTEM_CLUSTER, segments[i + 2]);
		  }

		if (sa.equals("a") && (i + 2) < segments.length)
		  {
		    gRequest.setParameter(GSParams.SYSTEM_MODULE_TYPE, "collection");
		    gRequest.setParameter(GSParams.SYSTEM_MODULE_NAME, segments[i + 2]);
		  }

		if (sa.equals("d") && (i + 2) < segments.length)
		  {
		    gRequest.setParameter(GSParams.SYSTEM_CLUSTER, segments[i + 2]);
		  }

		gRequest.setParameter(GSParams.SUBACTION, sa);

		additionalParameters = new String[] { GSParams.ACTION };
		defaultParamValues = new String[] { "s" };
	      }
	    //ADMIN
	    else if (segments[i].equals("admin") && (i + 1) < segments.length)
	      {
		String pageName = segments[i + 1];

		gRequest.setParameter("s1.authpage", pageName);

		additionalParameters = new String[] { GSParams.ACTION, GSParams.REQUEST_TYPE, GSParams.SUBACTION, GSParams.SERVICE };
		defaultParamValues = new String[] { "g", "r", "authen", "Authentication" };
	      }
	    //BROWSE
	    else if (segments[i].equals(BROWSE_PATH) && (i + 1) < segments.length)
	      {
		String cl = "";
		for (int j = 1; (i + j) < segments.length; j++)
		  {
		    String currentSegment = segments[i + j].replace("CL", "").replace("cl", "");
		    if (currentSegment.contains("."))
		      {
			String[] subsegments = currentSegment.split("\\.");
			for (String subsegment : subsegments)
			  {
			    subsegment = subsegment.replace("CL", "").replace("cl", "");

			    if (cl.length() > 0)
			      {
				cl += ".";
			      }

			    if (subsegment.length() > 0)
			      {
				cl += subsegment;
			      }
			  }
			continue;
		      }
		    if (!currentSegment.matches("^(CL|cl)?\\d+$"))
		      {
			continue;
		      }

		    if (cl.length() > 0)
		      {
			cl += ".";
		      }

		    cl += currentSegment;
		  }

		gRequest.setParameter("cl", "CL" + cl);

		additionalParameters = new String[] { GSParams.ACTION, GSParams.REQUEST_TYPE, GSParams.SERVICE };
		defaultParamValues = new String[] { "b", "s", "ClassifierBrowse" };
	      }
	    //QUERY
	    else if (segments[i].equals(SEARCH_PATH))
	      {
		String serviceName = "";
		if ((i + 1) < segments.length)
		  {
		    serviceName = segments[i + 1];
		    gRequest.setParameter("s", serviceName);

		    additionalParameters = new String[] { GSParams.ACTION, GSParams.SUBACTION, GSParams.REQUEST_TYPE };
		    defaultParamValues = new String[] { "q", "", "d" };
		  }
		if ((i + 2) < segments.length)
		  {
		    if (serviceName.equals("TextQuery") || serviceName.equals("RawQuery"))
		      {

			gRequest.setParameter("s1.query", segments[i + 2]);
		      }
		    else if (serviceName.equals("FieldQuery"))
		      {
			gRequest.setParameter("s1.fqv", segments[i + 2]);
		      }
		    else if (serviceName.equals("AdvancedFieldQuery"))
		      {
			gRequest.setParameter("s1.fqv", segments[i + 2]);
		      }
		  }
	      }
	    if (additionalParameters != null)
	      {
		for (int j = 0; j < additionalParameters.length; j++)
		  {
		    if (gRequest.getParameter(additionalParameters[j]) == null)
		      {
			gRequest.setParameter(additionalParameters[j], defaultParamValues[j]);
		      }
		  }
	      }
	  }
      }

    chain.doFilter(gRequest, response);
  }

  private boolean isURLRestricted(String url)
  {
    for (String restrictedURL : _restrictedURLs)
      {
	if (url.matches(".*" + restrictedURL + ".*"))
	  {
	    return true;
	  }
      }

    return false;
  }

  private String queryMRforDOCID(MessageRouter gsRouter, String collection, String assocfiledir) {
    Document gsDoc = XMLConverter.newDOM();

    Element metaMessage = gsDoc.createElement(GSXML.MESSAGE_ELEM);
    Element metaRequest = GSXML.createBasicRequest(gsDoc, GSXML.REQUEST_TYPE_PROCESS, collection + "/" + METADATA_RETRIEVAL_SERVICE, new UserContext());
    metaMessage.appendChild(metaRequest);

    Element paramList = gsDoc.createElement(GSXML.PARAM_ELEM + GSXML.LIST_MODIFIER);
    metaRequest.appendChild(paramList);
    
    Element param = gsDoc.createElement(GSXML.PARAM_ELEM);    
    param.setAttribute(GSXML.NAME_ATT, "metadata");
    param.setAttribute(GSXML.VALUE_ATT, "contains");
    paramList.appendChild(param);

    Element docList = gsDoc.createElement(GSXML.DOC_NODE_ELEM + GSXML.LIST_MODIFIER);
    metaRequest.appendChild(docList);

    Element doc = gsDoc.createElement(GSXML.DOC_NODE_ELEM);
    doc.setAttribute(GSXML.NODE_ID_ATT, assocfiledir);
    docList.appendChild(doc);
    
    Element metaResponse = (Element) gsRouter.process(metaMessage);
    
    NodeList metadataList = metaResponse.getElementsByTagName(GSXML.METADATA_ELEM);
    if (metadataList.getLength() == 0) {
      
      _logger.error("Could not find the document related to this url");
      return null;
    }
      
    Element metadata = (Element) metadataList.item(0);
    String document = metadata.getTextContent();
    if (document != null && document.equals("")) {
      document = null;
    }
    return document;
      
 
  }
  
}