12Parsec/bananaSPLIT
3
0
Fork 0
Code Issues Releases Wiki Activity

file separati vengono isolati in cartelle differenti

Browse Source
This commit is contained in:
Emanuele
2019-12-02 20:12:00 +01:00
parent 8628c3dbfb
commit 01dd92e4da
2 changed files with 15 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
bananaSPLIT/libbananasplit/libsplit.py
View File

@@ -57,6 +57,9 @@ class bananaSPLITTER(threading.Thread):
except IOError as e: except IOError as e:
self.log.critical("Impossibile aprire il file: {}! [{}]".format(self.fileName,e)) self.log.critical("Impossibile aprire il file: {}! [{}]".format(self.fileName,e))
raise BaseException("OpenFile") raise BaseException("OpenFile")
except UnicodeDecodeError as ee:
self.log.critical("Il file [{}] contiene caratteri non compatibili con la codifica scelta! [{}]"
.format(self.fileParams['name'],ee))
pass pass
def remEmptyLines(self): def remEmptyLines(self):
@@ -205,6 +208,8 @@ class bananaSPLITTER(threading.Thread):
pass pass
def saveSeparate(self): def saveSeparate(self):
os.mkdir(self.paths['OUTworkPath']+slugify(self.fileName))
self.paths['OUTworkPath']=self.paths['OUTworkPath']+slugify(self.fileName)
self.log.info("Salvo gli articoli in file separati...") self.log.info("Salvo gli articoli in file separati...")
self.log.debug("Persorso: {0}".format(self.paths['OUTworkPath'].format('nomeFile'))) self.log.debug("Persorso: {0}".format(self.paths['OUTworkPath'].format('nomeFile')))
for ff in self.fileList: for ff in self.fileList:

18
bananaSPLIT/libtestmain.py
View File

@@ -4,6 +4,7 @@ Created on 1 dic 2019
@author: Emanuele Trabattoni @author: Emanuele Trabattoni
''' '''
import os import os
from glob import glob
from libsplit import bananaSPLITTER from libsplit import bananaSPLITTER
from libconfload import bananaCONF from libconfload import bananaCONF
from libfancylogger import fancyLogger from libfancylogger import fancyLogger
@@ -15,14 +16,15 @@ confl.open()
confl.use("testEN.json") confl.use("testEN.json")
splconf = confl.getParams("splitter") splconf = confl.getParams("splitter")
splconf["name"] = splconf["paths"]["INworkPath"]+"GUARDIAN 1989.txt" for f in glob(splconf["paths"]["INworkPath"]+"*.txt"):
splitter = bananaSPLITTER(fileParams=splconf, logger=logger) splconf["name"] = f
splitter.openFile() splitter = bananaSPLITTER(fileParams=splconf, logger=logger)
splitter.remEmptyLines() splitter.openFile()
splitter.splitFile() splitter.remEmptyLines()
splitter.removeDuplicates() splitter.splitFile()
splitter.saveBody() splitter.removeDuplicates()
splitter.saveSeparate() splitter.saveBody()
splitter.saveSeparate()