Sigcat 1994: Conference Disk

home *** CD-ROM | disk | FTP | other *** search

/ Sigcat 1994: Conference Disk / SIGCAT 94 Conference Disk - DataDisc.ISO / gsbuild / gsbuild.doc < prev next >

Wrap

Text File | 1994-04-11 | 29.5 KB | 792 lines

GSBUILD.DOC 04/04/94 This file contains information on programs in the GSBUILD/GSSEARCH system. PROCEDURES PROGRAMS GBMK_ADR.EXE Reads binary db_name.LV4 file and translates it into ASCII data offsets file (default name = SIZE.TXT). GBMK_CFG then reads db_name.TMP and SIZE.TXT to make db_name.CFG. NOTE: db_name is the filename path\prefix of the data base. Generates GBMK_ADR.ERR file if a processing error is encountered. GBMK_ADR Command Line Format: GBMK_ADR db_name.LV4 siz_file <ENTER> db_name.LV4: See GSSEARCH.DOC. siz_file: See SIZE.TXT information. Versions: 1.0 93/08/23 GBMK_CFG.EXE Reads temporary configuration file (db_name.TMP) and data offsets file (SIZE.TXT) and makes a configuration file for GSSEARCH. NOTE: db_name is the filename path\prefix of the data base. The configuration file name is taken from the first line of the db_name.TMP file. Copies the pointer and data file names, the numbers of fields, indexes, and browse formats, and the title line from db_name.TMP to the db_name.CFG file. Then for each index field copies the index names, types, and data field number and merges the pointer data for that field from SIZE.TXT, and outputs the combined line to db_name.CFG file. Then copies the data field information and browse format information from the db_name.TMP file to the db_name.CFG file. Generates GBMK_CFG.ERR file if a processing error is encountered. GBMK_CFG Command Line Format: GBMK_CFG tmp_file siz_file <ENTER> tmp_file: Temporary configuration file for the data base. Default name: db_name.TMP. siz_file: ASCII file produced by GBMK_ADR.EXE. Default name: SIZE.TXT. Versions: 1.5 94/04/04 Revised to accomodate inclusion of # of decimal places in numeric field length. 1.4 93/10/21 Added test for too many lines of data in SIZE.TXT file. Added removal of bad .CFG file if errors are encountered. 1.3 93/09/23 Fixed left justification of index and field names. 1.2 93/08/15 Revised command line message trigger and added output of GBMK_CFG.ERR file. 1.0 93/07/14 GBMK_GBG.EXE Reads TAGGED TEXT files and db name.SCR file and generates the db_name.GBG file. NOTE: db_name is the filename path\prefix of the data base. Generates GBMK_GBG.ERR file if a processing error is encountered. GBMK_GBG Command Line Format: GBMK_GBG inf_file db_name scr_file [/a] <ENTER> inf_file: ASCII file containing the path/names of input tagged data files. GBMK_GBG will process all of the listed files, to produce a single db_name.GBG file. Default name: db_name.INF db_name: This is the path/prefix to which .GBG will be appended to make the db_name.GBG file. scr_file: Name of the script file for the data base. Default name: db_name.SCR /a Switch to append to an existing .GBG file (optional) Versions: 2.0 94/02/18 Revised record counter. 1.9 94/01/18 Added tests for disk space. 1.8 93/12/21 Added skip of 0x00 stopchar if Format == 0. 1.7 93/10/13 Increased size of GBG buffer to match TAG buffer. 1.6 93/10/13 Modified to insert 0xFF between paragraphs of non-formatted field. 1.5 93/09/28 Modified parsing of non-formatted field to remove extra spaces after CR/LF. 1.4 93/08/30 Made ReadTagFile function and removed tests for reads after each increment of TagPos. Tag file is read when < 0x6000 bytes are left in buffer (1 max size rec + 512 field tags.) 1.3 93/08/19 Made ReadTagFile function and added tests for reads after each increment of TagPos. 1.2 93/08/15 Revised command line message trigger and added output of GBMK_GBG.ERR file. 1.1 93/07/14 GBMK_TAG.EXE Converts dBase file to a TAGGED.TXT file. Deleted records (marked by a non-space in the delete-rec-marker byte are not processed. Null (empty) fields are not converted. Numeric fields that have all zeros or zeros with decimal point will not be converted if the /0 flag is set in the command line. Memo file data is translated as is. There is no conversion of extended ASCII characters. Generates GBMK_TAG.ERR file if a processing error is encountered. GBCHKTAG.EXE may be used to check a .TAG file for errors, and to generate a list of field names. Command line format: GBMK_TAG dbf_file tag_file [/n=numrecs] [/0] <ENTER> dbf_file: The path\name of the input DBF file. tag_file: The path\name of the output tagged text file. numrecs: Number of records to convert, default = all records. option: /0 Ignore null (0) numeric fields. Versions: 2.3 94/03/23 Added command line input of number of records to convert. 2.2 94/02/18 Revised record counter. 2.1 94/01/18 Revised error messages. 2.0 93/08/15 Fixed bugs in text/binary reads. Added skip of deleted record. Added seek to end of header. 1.0 93/07/14 GB_INDEX.EXE Reads db_name.SCR file and spawns the programs that build the data base. NOTE: db_name is the filename path\prefix of the data base. GB_INDEX command line format: GB_INDEX inf_file db_name scr_file <ENTER> inf_file: ASCII file containing the path/names of input tagged data files. GBMK_GBG will process all of the listed files, to produce a single db_name.GBG file. Default name: db_name.INF db_name: This is the path/prefix to which .GBG will be appended to make the db_name.GBG file. scr_file: Name of the script file for the data base. Default name: db_name.SCR Versions: 2.1 94/03/08 Added del SORTED.TXT, del MERGED.TXT. 2.0 93/12/15 Modified to use GBSORTC, GBSORTN, etc. Added check for existence of necessary programs. 1.0 93/11/01 GBXTRACT.EXE Reads db_name.GBG file, db_name.SCR file and STOPWORD.LST file. Extracts field(s) and generates XTRACTED.TXT file. Extracted search terms are parsed according to the stop characters specified in the .SCR file, and are compared to the stop words in STOPWORD.LST if the stop word flag is set. For non-numeric fields, spaces at the beginning of each extracted term are skipped. Normally data fields are indexed one at a time. However, up to 25 fields may be extracted at one time, to provide global indexing capability. Globally indexed fields must all be of the same type, with no mixing of numeric and character-based fields. NOTE: db_name is the filename path\prefix of the data base. Generates GBXTRACT.ERR file if a processing error is encountered. GBXTRACT command line format: GBXTRACT gbg_name script_name indexnum [indexnum ...]<ENTER> gbg_name: Name of the data file for the data base. Default name: db_name.GBG scr_name: Name of the script file for the data base. Default name: db_name.SCR indexnum: number of data field to extract (integer) A max of 25 field numbers may be entered. Versions: 1.9 93/02/17 Revised to only update counter every 10 records. 1.8 93/01/31 Revised to skip spaces at beginning of extracted character term. 1.7 93/01/18 Added tests for available disk space. 1.6 93/11/03 Changed buffer sizes to allow records up to 32k bytes. 1.5 93/10/19 Changed CR/LF stopchar in non-formatted field to 255 (0xFF). 1.4 93/10/04 Added goto nextfield if non-numeric found in field type N. Added data field type n for multiple-entry numeric fields. 1.3 93/10/01 Changed name of Fields structure to DataFields to avoid type clash with Fields structure in GSSEARCH.H Forced insertion of CR/LF or 0 into stop character list. 1.2 93/08/15 Revised command line message trigger and added output of GBXTRACT.ERR file. Fixed bug in fread trigger. 1.1 93/07/14 GBSORTC.EXE 01/06/93 GBSORTN.EXE 01/06/93 Reads the XTRACTED.TXT file and sorts it, generating SORTED.TXT. GBSORTC sorts character field output, GBSORTN sorts numeric field output. NOTE: Renamed from SORT.EXE, NSORT.EXE. GBMERGEC.EXE 01/06/93 GBMERGEN.EXE 01/06/93 Reads SORTED.TXT file and merges duplicate indexes, generating MERGED.TXT. GBMERGEC merges character field output, GBMERGEN merges numeric field output. NOTE: Renamed from MERGE.EXE, NMERGE.EXE. GBCOUNTC.EXE 01/06/93 GBCOUNTN.EXE 02/16/94 Reads MERGED.TXT file and counts indexes terms, and concatenates index address information into db_name.LV1 -- db_name.LV4 files. GBCOUNTC works on character field output, GBCOUNTN works on numeric field output. NOTE: Renamed from COUNT.EXE, NCOUNT.EXE. COLOR.EXE Changes color of menu screen. Command line format: COLOR colorstring <ENTER> where colorstring consists of a 2 or 3 character color defined as: intensity (Optional) - = low intensity (default) + = high intensity colors: R = Red G = Green B = Blue Y = Yellow M = Magenta C = Cyan W = White K = blacK EXAMPLE: COLOR -WB <ENTER> GSMENU.EXE, GSMENU.DAT See GSMENU.DOC. GSBUILD.MNU Menu data file for GSMENU.EXE. This calls the programs to build the data base. PREPB.EXE, PREPC.EXE Programs which allow interactive generation of the data base control files db_name.TMP and db_name.SCR. START.BAT Batch file which starts the GSMENU.EXE program using the GSBUILD.MNU data file. BUILD.BAT Batch file which starts the build process. This calls INDEX.BAT, and provides file control information as part of the command line format: db_path\INDEX.BAT INF_file db_path\db_name SCR_file INDEX.BAT Batch file which controls the building of the data base. Generated by PREPB.EXE. At each step, if a .ERR file is found the processing will stop. GBCHKTAG.EXE Checks a .TAG file for errors. Tests for extra curly brackets, control characters, tag names or records which are too long. Prints a list of field names, and errors found with line and byte number of occurrence. Command line format: GBCHKTAG tag_file output_file <ENTER> tag_file: The path\name of the output tagged text file. dbf_file: The path\name of the output data file. Versions: 2.3 04/08/94 Added record and byte counters. Renamed from CHK_TAGS. 2.2 12/20/93 Added test for unmatched curly brackets. Removed test for tildes. Moved reread section to get rid of bug giving erroneous error messages for "non-consecutive" CR and LFs. 2.1 11/03/93 Revised record test length to 32768 bytes. 2.0 10/06/93 Added test for record length. 1.0 03/12/93 GSSEARCH.EXE, GSSEARCH.HLP See GSSEARCH.DOC GSMK_HLP.EXE Program which allows user to make a database specific help file, which is accessed by GSSEARCH.EXE. See GSSEARCH.DOC for more information. Versions: 4.0 94/01/11 Revised to allow general data base help input as prefix.TXT. 3.0 93/12/07 Revised to put help text offsets for GSSEARCH.HLP at actual offset instead of offset+1. Revised to add CTRL-Z to end of text if not already present. 2.0 93/08/15 Changed to start with index field 0, and to allow non-sequential numbers (i.e. skipped numbers. ) 1.0 Unknown FILE FORMATS & INFORMATION progname.ERR Error file generated by programs in the GSBUILD software system. Contains error message describing the processing fault encountered. db_name.CFG See GSSEARCH.DOC. db_name.GBG See GSSEARCH.DOC. db_name.HLP See GSSEARCH.DOC. db_name.LV1 ... db_name.LV4 See GSSEARCH.DOC. db_name.TMP Prototype Configuration file (ASCII file) generated by PREPB.EXE from db_name.DBF and interactive input. Line # Contents 1 Name of configuration file to make. 2..7 Path/names of pointer and data files. 8 # data fields. (Y) 9 # indexed fields (X) 10 # browse formats specified 11 Title line 12..X+12 Indexed field data lines Index name (12 chars max) tilde character (stop read marker) Index type (integer) 1 = numeric, 3 = character; Data field number (starting with 0) X+12..X+12+Y Data field data lines Data field name (10 chars max) tilde character (stop read marker) Index type (integer) C = character, c = character (long), M = memo, N = numeric; Data field length. For Numeric fields the number of decimal places is indicated by an integer following a slash (/), e.g. 5/2. Y+1 Browse format #1 header line Y+2 Browse format #1 data line Y+3 Browse format #2 header line Y+4 Browse format #2 header line ... Example: HCDN.CFG \HCDN\HCDN.LV1~ \HCDN\HCDN.LV2~ \HCDN\HCDN.LV3~ \HCDN\HCDN.GBG~ \HCDN\HCDN.HLP~ \HCDN\HELP.GBG~ 10 6 2 *** HCDN TEST 06/08/93 *** ~ Sta_Num ~ 3 0 Sta_Name ~ 3 1 Region ~ 3 2 Sub_Region ~ 3 3 Drain_Area ~ 1 4 State ~ 3 7 STA_NUM ~ 8 C STA_NAME ~ 48 C REGION ~ 51 C SUB_REGION~ 51 C DRAIN_AREA~ 8/2 N NUM_YEARS ~ 3/0 N COMMENTS ~ 500 C STATE ~ 2 C LATITUDE ~ 6/4 N LONGITUDE ~ 7/4 N -1StaNum Station Name State Drn Area Lat Lon ~ 1 1 8 2 10 25 8 36 2 5 40 10 9 52 6 10 60 7 0 0 0 0 0 0 0 0 0 0 0 0 -2Sta_Num Station Name Area ~ 1 0 8 2 9 40 5 52 8 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 db_name.TAG TAGGED TEXT file (ASCII file) Contains data field-tagged record data. The record data is listed sequentially by data field. Fields must appear in the same order in the record as the list of field names in db_name.SCR. Only those fields with data should appear in the record. Each record must end with the {EOR} (end of record) field. Records may be arranged with one field per line or may be concatenated on one line. Blank lines between records are ignored. Field names -- MUST be delimited by left and right curly brackets; -- may be no more than ten characters, excluding brackets; -- may not contain blanks; Data may contain hard carriage returns. This allows retention of format for tables. Data MAY NOT CONTAIN tilde (~) characters. These are used as delimiting characters within the WORDNDX.TXT file. Data MAY NOT CONTAIN any control characters except carriage return (CR) and line feed (LF). CHK_TAGS.EXE may be used to check a TAG file for the existence of tilde characters, control characters, as well as other format errors. Example: {OFFICE}WRD {ROOM}SUITE 5 {MAILCITY}Lansing {STATE}MI {WACODE}517 {FTSNO}374-1608 {DEPT}DOI {BUREAU}USGS {CTRLDATE}06/25/89 {ELEV} 127.3 {FIRST}Brian {MI}D. {LAST}Abbott {EOR} ... more records ... SIZE.TXT file (ASCII file) Contains pointer data for each indexed field in the data base. Default name: SIZE.TXT NOTE: SIZE.TXT is produced by GBMK_ADR.EXE. Example: 2 1000 100 0 0 0 2 1000 214 60 3200 0 2 1000 47 120 10048 0 2 1000 2 180 11552 0 2 1000 12 240 11616 0 2 1000 95 300 12000 0 db_name.SCR SCRIPT DATA FILE (ASCII file) File containing information on the indexing parameters for the data base. For EACH data field contains: the data field name, NOTE: Data field name must be ended with a tilde (~) character. Maximum of 10 characters. data field indexing flag, 0 = do not index field 1 = index field retain format flag, Allows retention of format of tables, etc. Setting this to 1 for numeric fields allows retention of leading blanks for formatting. 0 = wrap data; Leading spaces are eliminated. CR/LF characters are converted to spaces except when this would result in two consecutive spaces. Consecutive CR/LF characters are converted to a single space. 1 = retain format; CR/LF characters are converted to 00. stop character string, List of characters to use as delimiters between words for indexing. Ignored if field is not indexed. Stop characters should NOT be used if the field is numeric. NOTE: Stop character list strings must be ended with a tilde (~) character. S for the stop character string denotes the standard stop character set for character fields: space ; semicolon . period : colon , comma ' apostrophe ? question mark - dash (minus sign) () parentheses * asterisk _ underscore ^ caret CR carriage return LF line feed " double quote N for the stop character string denotes the standard stop character set for NUMERIC fields: space CR carriage return LF line feed D for the stop character string denotes the standard stop character set for DATE fields: space CR carriage return LF line feed NOTE: Processing may be speeded considerably if the most likely stop characters (i.e. space, comma, period, etc.) are placed early in the list. stop words use flag, 0 = index all words, 1 = DO NOT INDEX words found in STOPWORD.LST file; Ignored if field is not indexed. See documentation on STOPWORD.LST file. data field type C = Character, short c = Character, long D = Date (mm/dd/yy) M = Memo N = Numeric, first word of field is the number. Subsequent words in the field are ignored. n = Numeric, with multiple numbers in the same data field. Non-numeric words are ignored. This type is useful for indexing numeric tables. NOTE: Commas will be removed from N and n numbers. Example: field index format stop stopword field name flag flag characters flag type OFFICE~ 1 0 S~ 0 C ROOM~ 1 0 S~ 0 C MAILSTOP~ 1 0 !@# %-_~ 0 C MAILCITY~ 1 0 S~ 0 C STATE~ 1 0 S~ 0 C WACODE~ 1 0 S~ 0 C FTSNO~ 1 0 S~ 0 C WEXCH~ 1 0 S~ 0 C WEXT~ 1 0 S~ 0 C DEPT~ 1 0 S~ 0 C BUREAU~ 1 0 S~ 0 C REMARKS~ 1 0 S~ 1 C CTRLDATE~ 1 0 D~ 0 D ELEV~ 1 1 N~ 0 N FIRST~ 1 0 S~ 0 C MI~ 1 0 S~ 0 C LAST~ 1 0 S~ 0 C EOR~ 0 1 S~ 0 C NOTES: The STATE field was flagged to not use stop words because their use would eliminate OR, IN and ND. Eliminating their use in the other fields (except REMARKS) also speeds processing. The CTRLDATE field has a format of MM/DD/YY. It was flagged to use a stop character which would NOT appear in the field so if would not be broken on the / characters. The ELEV field is numeric. The stop character flag is therefore set to N, signifying the standard numeric stop characters (space,CR,LF). It is also formatted with leading spaces to get the decimal places to line up in the browse display. In order to retain them the field was flagged to retain the format. STOPWORD.LST (ASCII file) List of words to be excluded from indexing. Words must be arranged one word to a line. All words will be converted to all capitals by the software. Words do not need to be alphanumerically sorted. Words must be continuous, i.e. no spaces within words. There may be no blank lines between words. NOTE: The normal stop word list contains IN, OR, and ND. Therefore the stopwords should not be used when indexing a field which contains state abbreviations, since Indiana, Oregon, and North Dakota will be excluded. Maximum number of words = 200. Maximum length of word = 10 characters. Example: ABOUT AFTER ALL ALSO AMONG AN AND ANY ... more words ... XTRACTED.TXT (Binary file) This file is generated by GBXTRACT.EXE, and is normally deleted after each index extraction. Contains information for each "term" indexed for the specified field. The information for each index term is contained in a block of 40 bytes, consisting of a 10 byte header and a 30 byte word area, which is arranged as follows: Bytes Use Type 0 .. 1 Record length in db_name.GBG file; int 2 .. 5 Record offset in db_name.GBG file; long 6 .. 7 # of word in data field; int 8 .. 9 Data field number; int 10 .. 39 Term, padded with trailing spaces char[30] Character fields start at byte 10. Numeric fields are right justified with the decimal point at byte 30. If there is no decimal point the ones digit is at byte 29.