Add --diff option teraslice-cli

[sotojn]

This PR makes the following changes:

• Adds --diff option to teraslice cli view that will compare the local json job file with what teraslice has on the state cluster
• Bumps teraslice-cli from v2.1.0 to v2.2.0

Ref to issue #3694

[busma13]

Looks good

[sotojn]

Here is an example on how this option looks. In the json file I change the amount of workers from 1 to 2 and run:

earl tjm view test-job.json --diff

Output:

  {
   "_context": "job",
   "_created": "2024-08-01T22:34:50.219Z",
-   "_updated": "2024-08-06T16:03:58.734Z",
+   "_updated": "2024-08-06T16:03:58.777Z",
    "assets": [
      "standard",
      "file"
    ],
    "job_id": "48e59745-ed75-4060-998e-002262abefc3",
    "lifecycle": "once",
    "name": "data-to-s3",
    "operations": [
      {
        "_op": "data_generator",
        "size": 100000
      },
      {
        "_op": "s3_exporter",
        "format": "ldjson",
        "path": "josephorigin"
      }
    ],
-   "workers": 1
+   "workers": 2
  }

--------
job: data-to-s3
id: 48e59745-ed75-4060-998e-002262abefc3
cluster: http://localhost:5678
--------

It does appear that updated seems slightly off all the time and maybe we shouldn't compare it?

[busma13]

It does appear that updated seems slightly off all the time and maybe we shouldn't compare it?

I think it could be helpful to know that _update will be different between the two records.

[sotojn]

It does appear that updated seems slightly off all the time and maybe we shouldn't compare it?

I think it could be helpful to know that _update will be different between the two records.

Should this be some sort of warning like, "_update field is subject to being slightly different from job file, to whats on the job record in the state cluster". Or should I just ignore this comparison entirely?

[busma13]

My feeling is that it's fine how it is now and a warning isn't necessary. But maybe it would be better to not compare _update if only the milliseconds are different.

[ciorg]

If there is no difference could it out put something like - no difference between the configs instead of printing the whole config, not a big deal it's a pain to do so

Also, what color applies to what config? Green is the local file and red is the cluster config? Maybe something to let people know what applies to what.

So far as the update difference, which for some reason is always different than the cluster value, at least on all the jobs I've looked at so far. ( I'm not sure how the update is calculated, but it seems like that value should be the same for both? Or maybe it shouldn't even have millisecond precision? But that is a different issue)

It is interesting to know when the job was last updated, but if the diff is 100 ms it doesn't really tell me anything useful. So maybe if there's a way to hide that diff unless it's at least a few min difference that would be better.

[sotojn]

I've updated the behavior to include what is being compared to what. I also added logic around the "_updated" field to only display a diff if the time difference is greater than a minute.

Results if "_updated field is less than 1 minute":
Command: earl tjm view data-to-s3.json --diff

  {
    "assets": [
      "standard",
      "file"
    ],
    "job_id": "48e59745-ed75-4060-998e-002262abefc3",
    "lifecycle": "once",
    "name": "data-to-s3",
    "operations": [
      {
        "_op": "data_generator",
        "size": 100000
      },
      {
        "_op": "s3_exporter",
        "format": "ldjson",
        "path": "josephorigin"
      }
    ],
-   "workers": 1   <--- state cluster value
+   "workers": 2   <--- local job file value
  }

--------
job: data-to-s3
id: 48e59745-ed75-4060-998e-002262abefc3
cluster: http://domain123
--------

Results if "_updated field is greater than 1 minute":
Command: earl tjm view data-to-s3.json --diff

  {
-   "_updated": "2024-08-06T16:03:58.734Z",   <--- state cluster value
+   "_updated": "2024-07-06T16:03:58.777Z",   <--- local job file value
    "assets": [
      "standard",
      "file"
    ],
    "job_id": "48e59745-ed75-4060-998e-002262abefc3",
    "lifecycle": "once",
    "name": "data-to-s3",
    "operations": [
      {
        "_op": "data_generator",
        "size": 100000
      },
      {
        "_op": "s3_exporter",
        "format": "ldjson",
        "path": "josephorigin"
      }
    ],
-   "workers": 1   <--- state cluster value
+   "workers": 2   <--- local job file value
  }

--------
job: data-to-s3
id: 48e59745-ed75-4060-998e-002262abefc3
cluster: http://domain123
--------

[ciorg]

LGTM

[busma13]

Looks good.