Partial implementation of consensus (see issue #3)

I can now build a consensus table that counts the characters in each
column of the alignment.

If we want to reduce these to a single consensus string, then we need to
decide how to resolve ties.

```
>seq_1
AAATATAT
>seq_2
AAAATTAT
>seq_3
AAATATAA
```

```
A       T
3       0
3       0
3       0
1       2
2       1
0       3
3       0
1       2
```

smof.py

@@ -50,6 +50,7 @@ def parse(argv=None):
     subcommands = [
         Cut,
         Clean,
+        Consensus,
         Filter,
         Grep,
         Md5sum,
@@ -2136,6 +2137,49 @@ def sortterm(x):
             yield s
 
 
+class Consensus(Subcommand):
+    def _parse(self):
+        cmd_name = 'consensus'
+        parser = self.subparsers.add_parser(
+            cmd_name,
+            usage=self.usage.format(cmd_name),
+            help="finds the consensus sequence for aligned sequence",
+            description="""Given input in aligned FASTA file format, where all
+            sequences are of equal length (possibly with gaps), `consensus`
+            will find the most common character in each column. Optionall, it
+            will instead provide the counts or proportions of each character at
+            each position."""
+        )
+        parser.add_argument(
+            'fh',
+            help='input fasta sequence (default = stdin)',
+            metavar='INPUT',
+            nargs="*"
+        )
+        parser.set_defaults(func=self.func)
+
+    def generator(self, args, gen):
+        raise NotImplemented
+
+    def write(self, args, gen, out=sys.stdout):
+        seqs = [s for s in gen.next()]
+        imax = max([len(s.seq) for s in seqs])
+        try:
+            transpose = [[s.seq[i] for s in seqs] for i in range(0, imax)]
+        except IndexError:
+            err("All sequences must be of equivalent length")
+
+        counts = Counter(('').join([s.seq for s in seqs]))
+        characters = list(counts.keys())
+
+        out.write("\t".join(characters))
+        out.write("\n")
+        for column in transpose:
+            c = Counter(column)
+            out.write("\t".join([str(c[x]) for x in characters]))
+            out.write("\n")
+
+
 # ==============
 # UNIX EMULATORS
 # ==============