| 1 | .\"------------------------------------------------------------
|
|---|
| 2 | .\" Id - set Rv,revision, and Dt, Date using rcs-Id tag.
|
|---|
| 3 | .de Id
|
|---|
| 4 | .ds Rv \\$3
|
|---|
| 5 | .ds Dt \\$4
|
|---|
| 6 | ..
|
|---|
| 7 | .Id $Id: mg_invf_rebuild.1 16583 2008-07-29 10:20:36Z davidb $
|
|---|
| 8 | .\"------------------------------------------------------------
|
|---|
| 9 | .TH mg_invf_rebuild 1 \*(Dt CITRI
|
|---|
| 10 | .SH NAME
|
|---|
| 11 | mg_invf_rebuild \- Rebuild an mg inverted file with or without skips.
|
|---|
| 12 | .SH SYNOPSIS
|
|---|
| 13 | .B mg_invf_rebuild
|
|---|
| 14 | .RB [ \-h ]
|
|---|
| 15 | .if n .ti +9n
|
|---|
| 16 | [
|
|---|
| 17 | .BR \-0 " |"
|
|---|
| 18 | .B \-1
|
|---|
| 19 | [
|
|---|
| 20 | .BI \-k " num"
|
|---|
| 21 | ] |
|
|---|
| 22 | .B \-2
|
|---|
| 23 | [
|
|---|
| 24 | .BI \-s " num"
|
|---|
| 25 | ]
|
|---|
| 26 | [
|
|---|
| 27 | .BI \-m " num"
|
|---|
| 28 | ]
|
|---|
| 29 | ]
|
|---|
| 30 | .if n .ti +9n
|
|---|
| 31 | [
|
|---|
| 32 | .BI \-d " directory"
|
|---|
| 33 | ]
|
|---|
| 34 | .BI \-f " name"
|
|---|
| 35 | .SH DESCRIPTION
|
|---|
| 36 | .B mg_invf_rebuild
|
|---|
| 37 | builds a new inverted file, with or without skipping, from an old
|
|---|
| 38 | inverted file. This file uses
|
|---|
| 39 | .I *.invf.ORG
|
|---|
| 40 | and
|
|---|
| 41 | .I *.invf.idx.ORG
|
|---|
| 42 | as the
|
|---|
| 43 | source from which it builds
|
|---|
| 44 | .I *.invf
|
|---|
| 45 | and
|
|---|
| 46 | .IR *.invf.idx .
|
|---|
| 47 | If
|
|---|
| 48 | .I *.invf.ORG
|
|---|
| 49 | or
|
|---|
| 50 | .I *.invf.idx.ORG
|
|---|
| 51 | do not exist, the program renames
|
|---|
| 52 | .I *.invf
|
|---|
| 53 | or
|
|---|
| 54 | .I *.invf.idx
|
|---|
| 55 | to
|
|---|
| 56 | .I *.invf.ORG
|
|---|
| 57 | or
|
|---|
| 58 | .I *.invf.idx.ORG
|
|---|
| 59 | as appropriate. The old inverted file may contain skipping. This
|
|---|
| 60 | means that it is possible to delete the
|
|---|
| 61 | .I *.ORG
|
|---|
| 62 | files
|
|---|
| 63 | after the new inverted file is built.
|
|---|
| 64 | .SH OPTIONS
|
|---|
| 65 | Options may appear in any order.
|
|---|
| 66 | .TP "\w'\fB\-m\fP \fInum\fP'u+2n"
|
|---|
| 67 | .B \-h
|
|---|
| 68 | This displays a usage line on
|
|---|
| 69 | .IR stderr .
|
|---|
| 70 | .TP
|
|---|
| 71 | .B \-0
|
|---|
| 72 | This generates a non-skipped inverted file. This option is normally
|
|---|
| 73 | only needed if the
|
|---|
| 74 | .I *.ORG
|
|---|
| 75 | files have been deleted.
|
|---|
| 76 | .TP
|
|---|
| 77 | .B \-1
|
|---|
| 78 | This generates a skipped inverted file. The
|
|---|
| 79 | .BI \-k " num"
|
|---|
| 80 | argument specifies the number of pointers hopped over with each skip.
|
|---|
| 81 | .TP
|
|---|
| 82 | .B \-2
|
|---|
| 83 | This option generates a skipped inverted file. The skipped inverted
|
|---|
| 84 | file is built so that it is `optimal' for ranking using a specific
|
|---|
| 85 | number of accumulators. Each term in the inverted file has a
|
|---|
| 86 | different skip length. The arguments
|
|---|
| 87 | .BR \-s " and " \-m
|
|---|
| 88 | control the sizes of the skips.
|
|---|
| 89 | .TP
|
|---|
| 90 | .BI \-k " num"
|
|---|
| 91 | This specifies the number of pointers that should be hopped over with
|
|---|
| 92 | each skip. This option is only valid if
|
|---|
| 93 | .B \-1
|
|---|
| 94 | is specified.
|
|---|
| 95 | .TP
|
|---|
| 96 | .BI \-m " num"
|
|---|
| 97 | This specifies the intended number of accumulators that will be used
|
|---|
| 98 | when ranking queries are done on the collection.
|
|---|
| 99 | .TP
|
|---|
| 100 | .BI \-s " num"
|
|---|
| 101 | This specifies the minimum size for skips. If the calculation of the
|
|---|
| 102 | optimal skip size results in a number smaller than
|
|---|
| 103 | .IR num ,
|
|---|
| 104 | the skip size is set to
|
|---|
| 105 | .IR num .
|
|---|
| 106 | .SH ENVIRONMENT
|
|---|
| 107 | .TP "\w'\fBMGDATA\fP'u+2n"
|
|---|
| 108 | .SB MGDATA
|
|---|
| 109 | If this environment variable exists, then its value is used as the
|
|---|
| 110 | default directory where the
|
|---|
| 111 | .BR mg (1)
|
|---|
| 112 | collection files are. If this variable does not exist, then the
|
|---|
| 113 | directory \*(lq\fB.\fP\*(rq is used by default. The command line
|
|---|
| 114 | option
|
|---|
| 115 | .BI \-d " directory"
|
|---|
| 116 | overrides the directory in
|
|---|
| 117 | .BR MGDATA .
|
|---|
| 118 | .SH FILES
|
|---|
| 119 | .TP 20
|
|---|
| 120 | .B *.invf
|
|---|
| 121 | Inverted file.
|
|---|
| 122 | .TP
|
|---|
| 123 | .B *.invf.ORG
|
|---|
| 124 | Original inverted file.
|
|---|
| 125 | .TP
|
|---|
| 126 | .B *.invf.idx
|
|---|
| 127 | The index into the inverted file.
|
|---|
| 128 | .TP
|
|---|
| 129 | .B *.invf.idx.ORG
|
|---|
| 130 | The original index into the inverted file.
|
|---|
| 131 | .TP
|
|---|
| 132 | .B *.invf.dict.build
|
|---|
| 133 | Compressed stemmed dictionary.
|
|---|
| 134 | .SH "SEE ALSO"
|
|---|
| 135 | .na
|
|---|
| 136 | .BR mg (1),
|
|---|
| 137 | .BR mg_compression_dict (1),
|
|---|
| 138 | .BR mg_fast_comp_dict (1),
|
|---|
| 139 | .BR mg_get (1),
|
|---|
| 140 | .BR mg_invf_dict (1),
|
|---|
| 141 | .BR mg_invf_dump (1),
|
|---|
| 142 | .BR mg_passes (1),
|
|---|
| 143 | .BR mg_perf_hash_build (1),
|
|---|
| 144 | .BR mg_text_estimate (1),
|
|---|
| 145 | .BR mg_weights_build (1),
|
|---|
| 146 | .BR mgbilevel (1),
|
|---|
| 147 | .BR mgbuild (1),
|
|---|
| 148 | .BR mgdictlist (1),
|
|---|
| 149 | .BR mgfelics (1),
|
|---|
| 150 | .BR mgquery (1),
|
|---|
| 151 | .BR mgstat (1),
|
|---|
| 152 | .BR mgtic (1),
|
|---|
| 153 | .BR mgticbuild (1),
|
|---|
| 154 | .BR mgticdump (1),
|
|---|
| 155 | .BR mgticprune (1),
|
|---|
| 156 | .BR mgticstat (1).
|
|---|
