ghttp.h

一个由Mike Gashler完成的机器学习方面的includes neural net, naive bayesian classifier, decision tree, KNN, a genet

/*
	Copyright (C) 2006, Mike Gashler

	This library is free software; you can redistribute it and/or
	modify it under the terms of the GNU Lesser General Public
	License as published by the Free Software Foundation; either
	version 2.1 of the License, or (at your option) any later version.

	see http://www.gnu.org/copyleft/lesser.html
*/

#ifndef __GHTTP_H__
#define __GHTTP_H__

#include <time.h>

class GHttpClientSocket;
class GSocketServer;
class GPointerArray;
class GHttpServerBuffer;
class GQueue;
class GStringHeap;
class GConstStringHashTable;


// This class allows you to get files using the HTTP protocol
class GHttpClient
{
public:
	enum Status
	{
		Downloading,
		Error,
		NotFound,
		Done,
		Aborted,
	};

protected:
	char m_szHeaderBuf[258];
	char m_szServer[256];
	char m_szClientName[32];
	int m_nHeaderPos;
	int m_nContentSize;
	bool m_bChunked;
	unsigned char* m_pData;
	int m_nDataPos;
	GHttpClientSocket* m_pSocket;
	Status m_status;
	GQueue* m_pChunkQueue;
	bool m_bPastHeader;
	char* m_szRedirect;
	double m_dLastReceiveTime;
	char caLastModifiedString[60];
	bool m_bAmCurrentlyDoingJustHeaders;

public:
	GHttpClient();
	~GHttpClient();

	// Send a request to get a file.  Returns immediately (before the file
	// is downloaded).
	bool Get(const char* szUrl, bool actuallyDownloadTheData = true);

	// See what the status of the download is.  If everything is going okay,
	// it will return "Downloading" while downloading and "Done" when the file
	// is available.  pfProgress is an optional parameter.  If it is non-NULL,
	// it will return a number between 0 and 1 that indicates the ratio of
	// content (not including header data) already downloaded.
	Status CheckStatus(float* pfProgress);

	// Don't call this until the status is "Done".  It returns a pointer to the
	// file that was downloaded.  The buffer will be deleted when this object is
	// deleted, so if you want to retain the buffer, call DropData instead.
	unsigned char* GetData(int* pnSize);
	//get the date string of this 
	char* GetUrlModified() { /* todor check first */ return caLastModifiedString; }

	// Just like GetData except it forgets about the buffer so you'll have to
	// delete it yourself.
	unsigned char* DropData(int* pnSize);
	


	// This is called when the connection is lost
	void OnLoseConnection();

	void SetClientName(const char* szClientName);
	
	void Abort();	// called by the consumer, when an abort is desired. 

protected:

	void ProcessHeader(const unsigned char* szData, int nSize);
	void ProcessBody(const unsigned char* szData, int nSize);
	void ProcessChunkBody(const unsigned char* szData, int nSize);
	void GimmeWhatYouGot();

};





#define MAX_SERVER_LINE_SIZE 300
#define MAX_COOKIE_SIZE 300

// This class allows you to implement a simple HTTP daemon
class GHttpServer
{
protected:
	GSocketServer* m_pSocket;
	GPointerArray* m_pBuffers;
	GQueue* m_pQ;
	char m_szContentType[64];
	char m_szCookie[MAX_COOKIE_SIZE];
	bool m_bPersistCookie;
	time_t m_time;

public:
	GHttpServer(int nPort);
	virtual ~GHttpServer();

	// You should call this method constantly inside the main loop.
	// It returns true if it did anything, and false if it didn't, so
	// if it returns false you may want to sleep for a little while.
	bool Process();

	// Unescapes a URL. (i.e. replace "%20" with " ", etc.)
	static void UnescapeUrl(char* szOut, const char* szIn);

	// Parses the parameters in a URL and puts them in a table
	static void ParseParams(GStringHeap* pStringHeap, GConstStringHashTable* pTable, const char* szParams);

	// Specifies the content-type of the response
	void SetContentType(const char* szContentType);

	// Specifies the set-cookie header to be sent with the response
	void SetCookie(const char* szPayload, bool bPersist);

	// Sets the date (modified time) to be sent with the file so the client can cache it
	void SetTimeStamp(time_t t) { m_time = t; }

protected:
	virtual void OnProcessLine(int nConnection, const char* szLine) {}
	void ProcessPostData(int nConnection, GHttpServerBuffer* pClient, const unsigned char* pData, int nDataSize);
	void ProcessHeaderLine(int nConnection, GHttpServerBuffer* pClient, const char* szLine);
	void BeginRequest(GHttpServerBuffer* pClient, int eType, const char* szIn);
	void SendResponse(GHttpServerBuffer* pClient, int nConnection);
	void SendNotModifiedResponse(GHttpServerBuffer* pClient, int nConnection);

	// This method should set the content type and the date headers, and any other
	// headers deemed necessary
	virtual void SetHeaders(const char* szUrl, const char* szParams) = 0;

	// The primary purpose of this method is to push a response into pResponse.
	// Typically this method will call SetHeaders.
	virtual void DoGet(const char* szUrl, const char* szParams, int nParamsLen, const char* szCookie, GQueue* pResponse) = 0;

	// This method takes ownership of pData. Don't forget to delete it. When the POST is
	// caused by an HTML form, it's common for this method to just call DoGet (passing
	// pData for szParams) and then delete pData. (For convenience, a '\0' is already appended
	// at the end of pData.)
	virtual void DoPost(const char* szUrl, unsigned char* pData, int nDataSize, const char* szCookie, GQueue* pResponse) = 0;

	// This is called when the client does a conditional GET. It should return true
	// if you wish to re-send the file, and DoGet will be called.
	virtual bool HasBeenModifiedSince(const char* szUrl, const char* szDate) = 0;
};


#endif // __GHTTP_H__