add dump-segment tool mode to dump v10 segment metadata

[clintropolis]

Description

Since there is no longer a meta.smoosh in v10 segments, this PR adds an option to the dump segment tool to show v10 metadata. As a convenience to make this easy to run, I have added a bin/dump-segment to druid packaging, so its really easy to run against segments if you have a druid installation handy. Output is just the serialized json of the stored metadata, and can be piped into jq. For example:

$ ./bin/dump-segment --dump metadata_v10 -d ~/workspace/data/druid/segmentsCache/wikipedia-v10-no-rollup_2016-06-27T00\:00\:00.000Z_2016-06-28T00\:00\:00.000Z_2025-12-31T22\:19\:27.478Z/druid.segment | jq .
{
  "containers": [
    {
      "startOffset": 0,
      "size": 7069778
    }
  ],
  "files": {
    "__base/__time": {
      "container": 0,
      "startOffset": 0,
      "size": 120208
    },
    "__base/added": {
      "container": 0,
      "startOffset": 4241057,
      "size": 14
    },
...

or like do fancy jq stuff like show biggest internal files or whatever

$ ./bin/dump-segment --dump metadata_v10 -d ~/workspace/data/druid/segmentsCache/wikip10-no-rollup_2016-06-27T00\:00\:00.000Z_2016-06-28T00\:00\:00.000Z_2025-12-31T22\:19\:27.478Z/druid.segment | jq '.files | to_entries | map({file:.key, size:.value.size}) | sort_by(.size) | reverse'
[
  {
    "file": "__base/diffUrl.__stringDictionary",
    "size": 1908699
  },
  {
    "file": "__base/comment.__stringDictionary",
    "size": 1001152
  },
  {
    "file": "__base/page.__stringDictionary",
    "size": 732776
  },
  {
    "file": "__base/diffUrl.__valueIndexes",
    "size": 635276
  },
  {
    "file": "__base/page.__valueIndexes",
    "size": 586076
  },
  {
...