00001 /* Copyright 2005 The Apache Software Foundation or its licensors, as 00002 * applicable. 00003 * 00004 * Licensed under the Apache License, Version 2.0 (the "License"); 00005 * you may not use this file except in compliance with the License. 00006 * You may obtain a copy of the License at 00007 * 00008 * http://www.apache.org/licenses/LICENSE-2.0 00009 * 00010 * Unless required by applicable law or agreed to in writing, software 00011 * distributed under the License is distributed on an "AS IS" BASIS, 00012 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 00013 * See the License for the specific language governing permissions and 00014 * limitations under the License. 00015 */ 00016 00017 /** 00018 * @file lcn_index.h 00019 * @brief Routines for manipulating a lucene index 00020 */ 00021 00022 #ifndef _LCN_INDEX_H 00023 #define _LCN_INDEX_H 00024 00025 #include "lcn_types.h" 00026 #include "lcn_term.h" 00027 #include "lcn_document.h" 00028 00029 #include <apr_file_io.h> 00030 00031 #ifdef __cplusplus 00032 extern "C" { 00033 #endif /* __cplusplus */ 00034 00035 /** Opaque structure representing a lucene index. */ 00036 typedef struct lcn_index_t lcn_index_t; 00037 00038 /** Return the index contained in directory @a d in @a idx, allocated in 00039 * @a pool. 00040 */ 00041 lcn_error_t * 00042 lcn_index_open (lcn_index_t **idx, 00043 const char *d, 00044 apr_pool_t *pool); 00045 00046 /** Return the maximum document number contained in index @a idx. */ 00047 apr_uint32_t 00048 lcn_index_max_docs (lcn_index_t *idx); 00049 00050 /** Opaque iterator for iterating over a list of documents. */ 00051 typedef struct lcn_doc_iter_t lcn_doc_iter_t; 00052 00053 /** Return the current document number from @a itr. */ 00054 apr_uint32_t lcn_doc_iter_doc (lcn_doc_iter_t *itr); 00055 00056 /** Return the maximum document number that can be returned from @a itr. */ 00057 apr_uint32_t lcn_doc_iter_max_docs (lcn_doc_iter_t *itr); 00058 00059 /** Return the current document's frequency from @a itr. */ 00060 apr_uint32_t lcn_doc_iter_freq (lcn_doc_iter_t *itr); 00061 00062 /** Return the total number of hits for the term in the current segment. */ 00063 apr_uint32_t lcn_doc_iter_term_freq (lcn_doc_iter_t *itr); 00064 00065 /** Move @a itr to the next document. 00066 * 00067 * If the increment was successful, @a next is TRUE, if it moved us past 00068 * the end of the sequence @a next is FALSE. 00069 */ 00070 lcn_error_t * lcn_doc_iter_next (lcn_boolean_t *next, lcn_doc_iter_t *itr); 00071 00072 /* XXX need an interface to read multiple doc/freq pairs at once. */ 00073 00074 /* XXX need an interface to skip to a certain document. */ 00075 00076 /** Close @a itr. */ 00077 lcn_error_t * lcn_doc_iter_close (lcn_doc_iter_t *itr); 00078 00079 /** Return an iterator @a itr for iterating over all documents that contain 00080 * @a term within @a idx, allocated from within @a pool. 00081 */ 00082 lcn_error_t * 00083 lcn_index_term_docs (lcn_doc_iter_t **itr, 00084 lcn_index_t *idx, 00085 lcn_term_t *term, 00086 apr_pool_t *pool); 00087 00088 /** Return the document @a doc from @a idx that has the number @a docnum, 00089 * allocated from @a pool. 00090 */ 00091 lcn_error_t * 00092 lcn_index_get_document (lcn_document_t **doc, 00093 lcn_index_t *idx, 00094 apr_uint32_t docnum, 00095 apr_pool_t *pool); 00096 00097 #ifdef __cplusplus 00098 } 00099 #endif /* __cplusplus */ 00100 00101 #endif