fix(cli): miner post JSON drops no_response into rejected bucket

[statxc]

Summary

Mirror of #724 for the sibling gitt miner post command. Three response states (accepted=True / False / None) are rendered as three distinct buckets in human mode, but JSON mode emits only accepted and rejected, silently bucketing no_response into rejected via rejected = total - accepted.

Fixes #806

Problem

gittensor/cli/miner_commands/post.py had two divergent classifications of the same per-validator response:

• Human table (3 buckets): [green]✓[/green] / [red]✗[/red] / [yellow]—[/yellow]
• JSON output (2 buckets): accepted and rejected = total - accepted — silently includes no-responses

Example with 5 validators (2 accept, 1 reject, 2 unreachable):

human: 2 ✓ / 1 ✗ / 2 —
JSON:  {"accepted": 2, "rejected": 3}     ← no_response folded into rejected

JSON consumers cannot distinguish "validator rejected my PAT" from "validator unreachable" — operationally distinct conditions that need different follow-up.

Fix

Mirror the structure #724 used for gitt miner check:

• Add _pat_post_row_category(row) -> {'accepted'|'rejected'|'no_response'} and _pat_post_aggregate_counts(results) -> {accepted, rejected, no_response} to gittensor/cli/miner_commands/helpers.py, sibling to the existing _pat_check_* pair.
• Replace the inline if/elif/else in post.py's human-mode loop with a _PAT_POST_STATUS_MARKUP[category] lookup driven by the same row-category helper.
• Spread **counts into the JSON envelope so all three buckets surface in the output.

After the fix:

JSON: {"accepted": 2, "rejected": 1, "no_response": 2}

Out of scope

This PR fixes only the JSON/human classification mismatch. Exit-code behavior on zero acceptances is intentionally not changed — the maintainer's position on that case is settled by #542 ("I wouldn't call that an operational failure, it's simply the output, no one accepted the PAT") and by the prior PR attempts #559, #550, #679.

[statxc]

Hi @anderdc how are you? could you review this?

[anderdc]

fix conflicts we just added some changes to those files

[statxc]

fix conflicts we just added some changes to those files

@anderdc I've fixed conflicts. Thanks.

[anderdc]

post.py:163-164 still has explicit 'accepted' and 'rejected' keys after the **counts spread, so rejected = len(results) - accepted_count overrides the correct value from counts and still bundles no_response. Drop those two keys (same shape as #724's change to check.py:107-114). Add an envelope-level test asserting accepted + rejected + no_response == total_validators.

[statxc]

post.py:163-164 still has explicit 'accepted' and 'rejected' keys after the **counts spread, so rejected = len(results) - accepted_count overrides the correct value from counts and still bundles no_response. Drop those two keys (same shape as #724's change to check.py:107-114). Add an envelope-level test asserting accepted + rejected + no_response == total_validators.

Thanks @anderdc I’ve updated this following you.

[statxc]

Thanks @anderdc