* Made eucjp_to_utf8.py more versatile and user friendly.

diff --git a/eucjp_to_utf8.py b/eucjp_to_utf8.py
index bcfbbb139580da9d274c59bc7360c429a54da60e..b1d9c687c385cd1a78ab66b040c02ef65f8b9903 100755 (executable)
--- a/eucjp_to_utf8.py
+++ b/eucjp_to_utf8.py
@@ -3,28 +3,91 @@
 """
     eucjp_to_utf8.py
 
-    Convert EUC-JP encoded text to UTF-8.
-    Input files with .gz extension are automatically decompressed.
-
-    USAGE:  eucjp_to_utf8.py [infile [outfile]]
+    Convert optional gzip compressed EUC-JP encoded text to UTF-8.
 
     Copyright (c) 2021 Urban Wallasch <irrwahn35@freenet.de>
     Modified ("3-clause") BSD License
 """
 
 import sys
+import os
 import codecs
 
-if len(sys.argv) > 1:
-    iname = sys.argv[1]
-    ifile = open(iname, 'rb')
-    if iname[-3:] == '.gz':
-        import gzip
-        ifile = gzip.GzipFile(fileobj=ifile)
-else:
-    ifile = sys.stdin.detach()
+def usage(msg=''):
+    eprint(
+"""%s\n
+USAGE:
+  %s [-d] [-n] [-v] [infile [outfile]]
+    -d : decompress gzip compressed input file; default: infer from infile extension
+    -n : do not attempt to decode EUC-JP; default: auto-detect
+    -v : print informational messages
+  Writes to stdout, if no outfile specified.
+  Reads from stdin, if no infile specified.
+""" % (msg, os.path.basename(sys.argv[0])))
+    sys.exit(1)
+
+decomp = False
+recode = True
+verbose = False
+ifile = None
+ofile = None
+
+def eprint(*args, **kwargs):
+    print(*args, file=sys.stderr, **kwargs)
+
+def veprint(*args, **kwargs):
+    if verbose:
+        eprint(*args, **kwargs)
 
-ofile = open(sys.argv[2], 'w') if len(sys.argv) > 2 else sys.stdout
+for arg in sys.argv[1:]:
+    try:
+        if arg[0] == '-':
+            for opt in arg[1:]:
+                if opt == 'd':
+                    decomp = True
+                elif opt == 'n':
+                    recode = False
+                elif opt == 'v':
+                    verbose = True
+                else:
+                    usage("Unexpected option: '-%s'" % opt)
+        elif not ifile:
+            iname = arg
+            ifile = open(iname, 'rb')
+            if not decomp and iname[-3:] == '.gz':
+                veprint('Info: gzip decompression auto-enabled')
+                decomp = True
+        elif not ofile:
+            ofile = open(arg, 'w')
+        else:
+            usage("Unexpected argument: '%s'" % arg)
+    except Exception as e:
+        usage(str(e))
 
+if not ofile:
+    ofile = sys.stdout
+    if not ifile:
+        ifile = sys.stdin.detach()
+
+if decomp:
+    import gzip
+    ifile = gzip.GzipFile(fileobj=ifile)
+
+cnt = 0
 for line in ifile:
-    print(line.decode('euc_jp'), end='', file=ofile)
+    try:
+        if recode:
+            try:
+                oline = line.decode('euc_jp')
+            except UnicodeDecodeError as e:
+                veprint('Info: EUC-JP decoding auto-disabled.')
+                oline = line.decode('utf_8')
+                recode = False
+        else:
+            oline = line.decode('utf_8')
+        print(oline, end='', file=ofile)
+        cnt += 1
+    except Exception as e:
+        usage('%s%s' % (str(e), "" if decomp else "\n(Forgot '-d'?)"))
+
+veprint('Successfully processed %d lines.' % cnt)