def load_corpus(directory):
texts = {}
docs = {}
for f in os.listdir(directory):
print 'Loading: ', directory + f
if f.endswith("txt8"):
with codecs.open(directory + f, 'r', 'ascii', 'ignore') as text:
texts[f[:-1]] = text.read()
elif f.endswith('docx'):
d = docx.clean(docx.opendocx(directory + f))
# converts to nltk text object
docs[f] = flatten(docx.getdocumenttext(d))
return texts, docs
'color': 'auto',
'space': 0,
'sz': 6,
'val': 'single',
},
},
'celstyle': [
{'align': 'center'},
{'align': 'left'},
{'align': 'right'},
],
'headstyle': { 'fill':'C6D9F1', 'themeFill':None, 'themeFillTint':None },
})
# Cleaning
docbody = docx.clean(docbody)
# ------------------------------
# Save output
# ------------------------------
# Prepare output file
outfile = zipfile.ZipFile('out.docx',mode='w',compression=zipfile.ZIP_DEFLATED)
# Copy unmodified sections
for f in template.namelist():
if not f in map(lambda i: i[0], actlist):
fo = template.open(f,'rU')
data = fo.read()
outfile.writestr(f,data)
fo.close()
