def main(argv):
import getopt
def usage():
print 'usage: %s [-d] [-p pages] [-P password] [-c codec] [-o output] file ...' % argv[0]
return 100
try:
(opts, args) = getopt.getopt(argv[1:], 'dp:P:c:o:')
except getopt.GetoptError:
return usage()
if not args: return usage()
debug = 0
cmapdir = 'CMap'
cdbcmapdir = 'CDBCMap'
codec = 'ascii'
pages = set()
password = ''
outfp = stdout
for (k, v) in opts:
if k == '-d': debug += 1
elif k == '-p': pages.add(int(v))
elif k == '-P': password = v
elif k == '-c': codec = v
elif k == '-o': outfp = file(v, 'wb')
#
CMapDB.initialize(cmapdir, cdbcmapdir, debug=debug)
rsrc = PDFResourceManager(debug=debug)
for fname in args:
pdf2txt(outfp, rsrc, fname, pages, codec, password=password, debug=debug)
return
def convert_cmap(files, cmapdir, cdbcmapdir, force=False):
from cmap import CMapDB
CMapDB.initialize(cmapdir)
for fname in files:
if fname.endswith('.upr'): continue
cmapname = os.path.basename(fname)
cdbname = os.path.join(cdbcmapdir, cmapname+'.cmap.cdb')
if not force and os.path.exists(cdbname):
print >>stderr, 'Skipping: %r' % cdbname
continue
print >>stderr, 'Reading: %r...' % fname
cmap = CMapDB.get_cmap(cmapname)
dumpcdb(cmap, cdbname)
return
def __init__(self, spec):
try:
self.basefont = literal_name(spec["BaseFont"])
except KeyError:
if STRICT:
raise PDFFontError("BaseFont is missing")
self.basefont = "unknown"
self.cidsysteminfo = dict_value(spec.get("CIDSystemInfo", {}))
self.cidcoding = "%s-%s" % (
self.cidsysteminfo.get("Registry", "unknown"),
self.cidsysteminfo.get("Ordering", "unknown"),
)
try:
name = literal_name(spec["Encoding"])
except KeyError:
if STRICT:
raise PDFFontError("Encoding not specified")
name = "unknown"
try:
self.cmap = CMapDB.get_cmap(name, strict=STRICT)
except CMapDB.CMapNotFound, e:
raise PDFFontError(e)
def get_cmap(self, name):
return CMapDB.get_cmap(name, strict=STRICT)
def get_cmap(self, cmapname, strict=False):
return CMapDB.get_cmap(cmapname, strict=strict)
def get_cmap(self, cmapname, strict=False):
return CMapDB.get_cmap(cmapname, strict=strict)
