Bucketing a collection of samples fails

[Chris Warburton]

From Hydra:

    <5>building path(s) â??/nix/store/1jpbm93rxi0rd30g5r6qzp0vw0kl8bz7-processed-hashed.jsonâ??
    b9f5a5jjla2s5ws0n6j0ic64pflcyzk6-astsOf: when expecting a Lazy Text, encountered Object instead
    ['1', '0']
    Traceback (most recent call last):
      File "/nix/store/vxbhq89jbpv9szm21piwbf3srpdndikr-process-samples.py", line 34, in <module>
        print(dumps(recurse([], data)))
      File "/nix/store/vxbhq89jbpv9szm21piwbf3srpdndikr-process-samples.py", line 28, in recurse
        return {k: recurse(path + [k], val[k]) for k in sort(val)}
      File "/nix/store/vxbhq89jbpv9szm21piwbf3srpdndikr-process-samples.py", line 28, in <dictcomp>
        return {k: recurse(path + [k], val[k]) for k in sort(val)}
      File "/nix/store/vxbhq89jbpv9szm21piwbf3srpdndikr-process-samples.py", line 28, in recurse
        return {k: recurse(path + [k], val[k]) for k in sort(val)}
      File "/nix/store/vxbhq89jbpv9szm21piwbf3srpdndikr-process-samples.py", line 28, in <dictcomp>
        return {k: recurse(path + [k], val[k]) for k in sort(val)}
      File "/nix/store/vxbhq89jbpv9szm21piwbf3srpdndikr-process-samples.py", line 27, in recurse
        return dict(val, sample=recurse(path, val['sample']))
      File "/nix/store/vxbhq89jbpv9szm21piwbf3srpdndikr-process-samples.py", line 30, in recurse
        return process(val)
      File "/nix/store/vxbhq89jbpv9szm21piwbf3srpdndikr-process-samples.py", line 20, in <lambda>
        input=dumps(astsFor(names)).encode('utf-8')).decode('utf-8'))
      File "/nix/store/klzk8v5mfipfcs2f5004riqmnycsfz8p-python3-3.4.4/lib/python3.4/subprocess.py", line 620, in check_output
        raise CalledProcessError(retcode, process.args, output=output)
    subprocess.CalledProcessError: Command '['/nix/store/9rj1jpgnf8nxj3bgpj06kwngp95013vq-hash']' returned non-zero exit status 1
    <3>builder for â??/nix/store/r1gm3hbrpfyna994pi0wzfcp3x1y1fbc-processed-hashed.json.drvâ?? failed with exit code 1

It looks to me like the Python runner is fetching ASTs of each sample,
and passing those into the bucketing command.

That doesn't fit with how our bucketing currently works: the bucketing
commands read in a list of names, and they look up those names using the
'astsOf' Haskell command from benchmarkingCommands.

I'll try removing this AST-lookup step from the Python, so the bucketers
can do it themselves.