CIRCL
/
PyCIRCLean
mirror of https://github.com/CIRCL/PyCIRCLean
2
0
Fork 0
Code Issues Releases Wiki Activity

Use default python-magic, escape filenames

pull/2/head
Raphaël Vinot 2015-11-05 16:27:48 +01:00
parent 03f1d90f33
commit 1a2637b252
3 changed files with 6 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
.travis.yml
View File

@ -16,7 +16,6 @@ addons:
- p7zip-full - p7zip-full
install: install:
- pip install git+https://github.com/Rafiot/python-magic.git@travis
- sudo add-apt-repository ppa:fontforge/fontforge --yes - sudo add-apt-repository ppa:fontforge/fontforge --yes
- sudo add-apt-repository ppa:coolwanglu/pdf2htmlex --yes - sudo add-apt-repository ppa:coolwanglu/pdf2htmlex --yes
- sudo apt-get update -qq - sudo apt-get update -qq
@ -31,6 +30,7 @@ install:
- popd - popd
- pip install lxml - pip install lxml
- pip install officedissector - pip install officedissector
- pip install python-magic
- if [ $(python -c 'import sys; print(sys.version_info.major)') == '2' ]; then pip install oletools olefile ; fi - if [ $(python -c 'import sys; print(sys.version_info.major)') == '2' ]; then pip install oletools olefile ; fi
- pip install coveralls - pip install coveralls
- pip install codecov - pip install codecov

2
bin/filecheck.py
View File

@ -374,7 +374,7 @@ class KittenGroomerFileCheck(KittenGroomerBase):
self.cur_file.log_string += 'Archive extracted, processing content.' self.cur_file.log_string += 'Archive extracted, processing content.'
tmpdir = self.cur_file.dst_path + '_temp' tmpdir = self.cur_file.dst_path + '_temp'
self._safe_mkdir(tmpdir) self._safe_mkdir(tmpdir)
extract_command = '{} -p1 x {} -o{} -bd -aoa'.format(SEVENZ, self.cur_file.src_path, tmpdir) extract_command = '{} -p1 x "{}" -o"{}" -bd -aoa'.format(SEVENZ, self.cur_file.src_path, tmpdir)
self._run_process(extract_command) self._run_process(extract_command)
self.recursive += 1 self.recursive += 1
self.processdir(tmpdir, self.cur_file.dst_path) self.processdir(tmpdir, self.cur_file.dst_path)

8
bin/generic.py
View File

@ -244,7 +244,7 @@ class KittenGroomer(KittenGroomerBase):
name, ext = os.path.splitext(filename) name, ext = os.path.splitext(filename)
tmppath = os.path.join(tmpdir, name + '.pdf') tmppath = os.path.join(tmpdir, name + '.pdf')
self._safe_mkdir(tmpdir) self._safe_mkdir(tmpdir)
lo_command = '{} --format pdf -eSelectPdfVersion=1 --output {} {}'.format( lo_command = '{} --format pdf -eSelectPdfVersion=1 --output "{}" "{}"'.format(
UNOCONV, tmppath, self.cur_file.src_path) UNOCONV, tmppath, self.cur_file.src_path)
self._run_process(lo_command) self._run_process(lo_command)
self._pdfa(tmppath) self._pdfa(tmppath)
@ -252,7 +252,7 @@ class KittenGroomer(KittenGroomerBase):
def _pdfa(self, tmpsrcpath): def _pdfa(self, tmpsrcpath):
'''Way to process PDF/A file''' '''Way to process PDF/A file'''
pdf_command = '{} --dest-dir / {} {}'.format(PDF2HTMLEX, tmpsrcpath, pdf_command = '{} --dest-dir / "{}" "{}"'.format(PDF2HTMLEX, tmpsrcpath,
self.cur_file.dst_path + '.html') self.cur_file.dst_path + '.html')
self._run_process(pdf_command) self._run_process(pdf_command)
@ -266,7 +266,7 @@ class KittenGroomer(KittenGroomerBase):
# The magic comes from here: http://svn.ghostscript.com/ghostscript/trunk/gs/doc/Ps2pdf.htm#PDFA # The magic comes from here: http://svn.ghostscript.com/ghostscript/trunk/gs/doc/Ps2pdf.htm#PDFA
curdir = os.getcwd() curdir = os.getcwd()
os.chdir(self.ressources_path) os.chdir(self.ressources_path)
gs_command = '{} -dPDFA -dQUIET -dSAFER -dBATCH -dNOPAUSE -dNOOUTERSAVE -sProcessColorModel=DeviceCMYK -sDEVICE=pdfwrite -sPDFACompatibilityPolicy=1 -sOutputFile={} ./PDFA_def.ps {}'.format( gs_command = '{} -dPDFA -dQUIET -dSAFER -dBATCH -dNOPAUSE -dNOOUTERSAVE -sProcessColorModel=DeviceCMYK -sDEVICE=pdfwrite -sPDFACompatibilityPolicy=1 -sOutputFile="{}" ./PDFA_def.ps "{}"'.format(
GS, os.path.join(curdir, tmppath), os.path.join(curdir, self.cur_file.src_path)) GS, os.path.join(curdir, tmppath), os.path.join(curdir, self.cur_file.src_path))
self._run_process(gs_command) self._run_process(gs_command)
os.chdir(curdir) os.chdir(curdir)
@ -280,7 +280,7 @@ class KittenGroomer(KittenGroomerBase):
self.cur_file.log_string += 'Archive extracted, processing content.' self.cur_file.log_string += 'Archive extracted, processing content.'
tmpdir = self.cur_file.dst_path + '_temp' tmpdir = self.cur_file.dst_path + '_temp'
self._safe_mkdir(tmpdir) self._safe_mkdir(tmpdir)
extract_command = '{} -p1 x {} -o{} -bd -aoa'.format(SEVENZ, self.cur_file.src_path, tmpdir) extract_command = '{} -p1 x "{}" -o"{}" -bd -aoa'.format(SEVENZ, self.cur_file.src_path, tmpdir)
self._run_process(extract_command) self._run_process(extract_command)
self.recursive += 1 self.recursive += 1
self.processdir(tmpdir, self.cur_file.dst_path) self.processdir(tmpdir, self.cur_file.dst_path)