Atlas48/JargonFile
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Fix exclusions

Browse Source
This commit is contained in:
Bob Mottram 2014-04-26 15:27:41 +01:00
parent d7ce6ebeed
commit 1d0ec34604
1 changed files with 11 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
import/importjargon.py
View File

@ -129,7 +129,7 @@ def jargonCreateEntry(title, text, outputDir):
filename = outputDir filename = outputDir
if not outputDir.endswith('/'): if not outputDir.endswith('/'):
filename = filename + '/' filename = filename + '/'
filename = filename + title + '.txt' filename = filename + jargonSaneTitle(title) + '.txt'
# don't overwrite existing files # don't overwrite existing files
if os.path.isfile(filename): if os.path.isfile(filename):
@ -151,8 +151,8 @@ def jargonReadFile(filename, exclusions, outputDir):
parser.bodyText is not '' and \ parser.bodyText is not '' and \
len(parser.title) > 1: len(parser.title) > 1:
saneBodyText = jargonSaneText(parser.title, parser.bodyText) saneBodyText = jargonSaneText(parser.title, parser.bodyText)
parser.title = jargonSaneTitle(parser.title) if not ((jargonSaneTitle(parser.title) in exclusions) or \
if not parser.title in exclusions: (parser.title in exclusions)):
entryFilename = jargonCreateEntry(parser.title, saneBodyText, outputDir) entryFilename = jargonCreateEntry(parser.title, saneBodyText, outputDir)
if entryFilename is not '': if entryFilename is not '':
print entryFilename print entryFilename
@ -170,14 +170,21 @@ def jargonReadExclusions(filename):
exclusions = fp.readlines() exclusions = fp.readlines()
fp.close() fp.close()
tempExclusions = []
for i in range(len(exclusions)):
tempExclusions.append(exclusions[i].strip('\n'))
exclusions = tempExclusions
return exclusions return exclusions
def jargonImport(rootDir, excludeEntriesFilename, outputDir): def jargonImport(rootDir, excludeEntriesFilename, outputDir):
exclusions = jargonReadExclusions(excludeEntriesFilename) exclusions = jargonReadExclusions(excludeEntriesFilename)
print exclusions
for dirName, subdirList, fileList in os.walk(rootDir): for dirName, subdirList, fileList in os.walk(rootDir):
for filename in fileList: for filename in fileList:
jargonReadFile(dirName + '/' + filename, exclusions, outputDir) jargonReadFile(dirName + '/' + filename, exclusions, outputDir)
if __name__ == "__main__": if __name__ == "__main__":
jargonImport('../original','','../entries') jargonImport('../original','exclusions.txt','../entries')