open-source-search-engine/Msg3a.h

#ifndef _MSG3A_H_
#define _MSG3A_H_

#include "Msg39.h"
//#include "Stats.h"
//#include "Thesaurus.h"
//#include "PostQueryRerank.h" // rscore_t
#include "Msg0.h"
#include "Msg1.h"

// 90MB for 32 nodes we got now with about 1.3B docs
#define DEFAULT_POSDB_READSIZE 90000000

void setTermFreqWeights ( collnum_t collnum, // char *coll,
			  class Query *q );
			  //int64_t *termFreqs, 
			  //float *termFreqWeights ) ;

//#define MSG3A_TMP_BUFSIZE (MAX_RESULTS*18)

// ALWAYS get at least 20 docids so we can do better ranking
#define MIN_DOCS_TO_GET 20

#define RBUF_SIZE 2048

class Msg3a {
public:
	Msg3a();
	~Msg3a();
	void constructor();

	void reset ( );

	// . returns false if blocked, true otherwise
        // . sets errno on error
        // . "query/coll/docIds" should NOT be on the stack in case we block
	// . uses Query class to parse query
	// . uses Msg37 to retrieve term frequencies for each termId in query
	// . uses Indexdb class to intersect the lists to get results
	// . fills docIds buf with the resulting docIds
	// . sets *numDocIds to the # of resulting docIds
	// . if restrictindexdbForQuery is true we only read docIds from 
	//   indexdb root file
	// . this might ADJUST m_si->m_q.m_termFreqs[] to be more accurate
	// . NOTE: Msg39Request MUST NOT BE ON THE STACK! keep it persistent!
	bool getDocIds ( Msg39Request *r          ,
			 Query        *q          ,
			 void         *state      ,
			 void        (* callback) ( void *state ) ,
			 class Host *specialHost = NULL );


	bool gotTermFreqs();

	// Msg40 calls this to get Query m_q to pass to Summary class
	Query *getQuery ( ) { return m_q ; };

	// Msg40 calls these to get the data pointing into the reply
	int64_t *getDocIds        ( ) { return m_docIds;        };
	char      *getClusterLevels ( ) { return m_clusterLevels; };
	// we basically turn the scores we get from each msg39 split into
	// floats (rscore_t) and store them as floats so that PostQueryRerank
	// has an easier time
	double *getScores        ( ) { return m_scores;        };
	int32_t   getNumDocIds     ( ) { return m_numDocIds; };

	int32_t getSiteHash26 ( int32_t i ) { 
		// if it was in the seoResults cache this will be non-NULL
		if ( m_siteHashes26 ) return m_siteHashes26[i];
		// otherwise, this is the standard method:
		return g_clusterdb.getSiteHash26((char *)&m_clusterRecs[i]); 
	};

	bool sortFacetEntries();

	void printTerms ( ) ;

	// . estimates based on m_termFreqs, m_termSigns and m_numTerms
	// . received in reply
	int64_t  getNumTotalEstimatedHits ( ) { 
		return m_numTotalEstimatedHits; };

	// called when we got a reply of docIds
	bool gotAllShardReplies ( );

	bool gotCacheReply ( );

	bool mergeLists ( );

	bool gotReranked ( );

	bool gotClusterRecs ( bool calledFromMergeLists );

	bool gotClusterLevels ( );

	// . deserialize ourselves for the cache
	// . returns bytes written
	// . returns -1 and sets g_errno on error
	int32_t getStoredSize ( );
	int32_t serialize   ( char *buf , char *bufEnd );
	int32_t deserialize ( char *buf , char *bufEnd );


	// incoming parameters passed to Msg39::getDocIds() function
	Query     *m_q;
	int32_t       m_docsToGet;
	void      *m_state;
	void     (*m_callback ) ( void *state );

	// set by Msg3a initially
	//int32_t       m_indexdbSplit;
	int32_t m_numHosts;

	// used in XmlDoc.cpp for doing the seo keyword tool
	void      *m_hack;
	int32_t       m_hackQNum;
	char      *m_hackQPtr;
	char       m_inUse;

	bool m_moreDocIdsAvail;
	int32_t m_totalDocCount;

	// don't get more docIds than this from any one split
	int32_t       m_maxDocIdsToCompute;
	// this is set if IndexTable::addLists() had an error
	int32_t       m_errno;

	// this is now in here so Msg40 can send out one Msg3a per
	// collection if it wants to search an entire token
	Msg39Request m_rrr;

	// use msg37 to get TermFreqs
	//Msg37      m_msg37;
	//int64_t  m_termFreqs      [MAX_QUERY_TERMS];
	//float      m_termFreqWeights[MAX_QUERY_TERMS];

	// a multicast class to send the request, one for each split
	Multicast  m_mcast[MAX_SHARDS];

	// for timing how long things take
	int64_t  m_startTime;

	// this buffer should be big enough to hold all requests
	//char       m_request [MAX_MSG39_REQUEST_SIZE * MAX_SHARDS];
	int32_t       m_numReplies;

	int32_t m_skippedShards;

	// . # estimated total hits
	int64_t  m_numTotalEstimatedHits;

	// we have one request that we send to each split
	class Msg39Request *m_r;
	char               *m_rbufPtr;
	int32_t                m_rbufSize;
	char                m_rbuf [ RBUF_SIZE ];

	// now we send to the twin as well
	SafeBuf m_rbuf2;

	// each split gives us a reply
	class Msg39Reply   *m_reply       [MAX_SHARDS];
	int32_t                m_replyMaxSize[MAX_SHARDS];

	char m_debug;

	// final merged lists go here
	int64_t      *m_docIds        ;
	double         *m_scores        ;
	class DocIdScore **m_scoreInfos ;
	//key_t          *m_recs          ; // clusterdb recs
	key_t          *m_clusterRecs   ;
	char           *m_clusterLevels ;
	// this is new
	collnum_t      *m_collnums;
	int32_t            m_numDocIds     ;
	// the above ptrs point into this buffer
	char           *m_finalBuf;
	int32_t            m_finalBufSize;

	// when merging this list of docids into a final list keep
	// track of the cursor into m_docIds[]
	int32_t m_cursor;

	// what collection # are these docids from if m_collnums[] is NULL
	//collnum_t m_collnum;

	//
	// new things for seoresults cache
	//
	int32_t           *m_siteHashes26;
	key_t           m_ckey; // cachedb key
	Msg0            m_msg0;
	Msg1            m_msg1;
	RdbList         m_seoCacheList;

	// we don't have FacetStats because we have the actual 
	// Msg39Reply::ptr_facetHashList from each shard which contains
	// all the facet hash lists for each gbfacet: query term we had
	// and the query "Msg3a::m_q.m_qterms[].m_dt" is the hash table
	// where each key is a facethash for that gbfacet:xxxx term and
	// the value if the # of occurrences.
	//SectionStats    m_sectionStats;
};

#endif