Files
bananaSPLIT/bananaconfEN.json

65 lines
1.2 KiB
JSON

{
"INworkPath": "D:\\Test\\",
"OUTworkPath": "D:\\Test\\Separati\\",
"OUTnameFormat":"TEST_{docnum}_{year:04d}{month:02d}{day:02d}_{title}.txt",
"docStruct": {
"docSep": "\\s*Copyright [(0-9)]+",
"dateFormat":"{month} {day:d}, {year:d}{}",
"dateWords": [
"January",
"February",
"March",
"April",
"May",
"June",
"July",
"August",
"September",
"October",
"November",
"December"
],
"headWords": [
"BYLINE:",
"SECTION:",
"LENGTH:",
"DATELINE:",
"HIGHLIGHT:",
"Email:"
],
"tailWords": [
"Newstex ID" ,
"NOTES",
"LANGUAGE:",
"GRAPHIC:",
"TYPE:",
"URL:",
"LOAD-DATE:",
"PUBLICATION-TYPE:",
"DOCUMENT-TYPE:",
"CHARTS:",
"JOURNAL-CODE:"
]
},
"settings": {
"encoding": "utf-8",
"monthPosition": 0,
"getNewsPaperName": true,
"nameNotFoundStr" : "--ND--",
"includeTitle" : true,
"removeDuplicates": true,
"showSkipped": false,
"showRemovedDuplicates": true,
"maxTitleLen": 32,
"loadTXT": true,
"loadDOCX": false,
"removeOldFiles":true,
"saveSeparateFiles": true,
"saveBodyFile": true,
"saveBodyNumber":true,
"delLF": false,
"delWordBreak": true,
"delChars": "'|@|#"
}
}