submit: default validation + hf_username fields on the pending row

Bundle 1+2 C4. `_build_pending_row` adds three new keys matching the
schema bump from C1, in the schema.md column order:

- validation_status = "unvalidated" (every new row starts in the
unvalidated tier; maintainers promote via the dataset repo after
methodology review).
- validation_method = None (set by the maintainer at promotion time
to one of code | traces | api | manual).
- hf_username = None (placeholder; C10 plumbs through the value
from gradio.OAuthProfile.username once the LoginButton lands).

Pre-existing fields, submitted_at stamping, the cadgenbench version
+ data revision tags, the score-shaped null fields, the blob URL,
and the sha256 are untouched. Dedup gate still pivots on
submission_sha256 exactly as before.

tests/test_submit.py covers the C4 acceptance:

- new-row defaults: validation_status is "unvalidated",
validation_method + hf_username are None;
- sha256 path passthrough: the hash argument lands on the row's
submission_sha256 key intact (existing dedup gate's row-level
half is unchanged);
- existing-metadata regression guard: submission_id, status,
submitter_name, submission_name, agent_url, notes,
submission_blob_url, cadgenbench_data_revision, and every
score-shaped null field all keep their pre-C4 values.

`_resolve_data_revision` is monkeypatched so the suite has zero
Hub I/O; cadgenbench.__version__ + datetime.now both stay live
(pure / process-local). 7/7 unit tests green locally (3 new + 4
existing leaderboard).

submit.py

@@ -401,6 +401,12 @@ def _build_pending_row(
     metadata + ``status: pending`` + ``submission_blob_url`` +
     ``submission_sha256``; every score-shaped field is ``null`` until
     the worker flips the row.
+
+    Validation-tier fields default per the validation-policy decision
+    doc: ``validation_status: "unvalidated"`` (maintainers promote
+    post-eval), ``validation_method: None``. ``hf_username`` is also
+    ``None`` here; C10 wires it through from ``gr.OAuthProfile`` once
+    the LoginButton lands.
     """
     return {
         "submission_id": submission_id,
@@ -421,6 +427,9 @@ def _build_pending_row(
         "per_fixture_breakdown": None,
         "submission_blob_url": blob_url,
         "submission_sha256": submission_sha256,
+        "validation_status": "unvalidated",
+        "validation_method": None,
+        "hf_username": None,
     }
 
 

tests/test_submit.py

@@ -0,0 +1,88 @@
+"""Unit tests for the submit-tab pending-row builder.
+
+C4 contract: ``_build_pending_row`` defaults the three Bundle 1+2
+schema fields (``validation_status="unvalidated"``,
+``validation_method=None``, ``hf_username=None``) and keeps the
+existing metadata + sha256 fields intact.
+
+Hub I/O (``_resolve_data_revision`` reads from
+``HfApi.dataset_info``) is monkeypatched out, so the suite has zero
+network traffic.
+"""
+from __future__ import annotations
+
+import submit
+
+
+def _stub_meta() -> dict:
+    """Minimum meta.json shape that survives ``_load_and_validate_meta``."""
+    return {
+        "submitter_name": "team-test",
+        "submission_name": "Stub Agent v1",
+        "agent_url": "https://github.com/example/stub-agent",
+        "notes": "test row, not a real submission",
+        "agree_to_publish": True,
+    }
+
+
+def test_pending_row_defaults_new_fields(monkeypatch):
+    """Three Bundle 1+2 fields land with their schema defaults."""
+    monkeypatch.setattr(submit, "_resolve_data_revision", lambda: "test-rev")
+    row = submit._build_pending_row(
+        submission_id="sub-test-x",
+        meta=_stub_meta(),
+        blob_url="https://huggingface.co/datasets/example/sub-test-x.zip",
+        submission_sha256="a" * 64,
+    )
+    assert row["validation_status"] == "unvalidated"
+    assert row["validation_method"] is None
+    assert row["hf_username"] is None
+
+
+def test_pending_row_preserves_sha256(monkeypatch):
+    """Existing dedup path's row-level half: sha256 still gets stamped on the row."""
+    monkeypatch.setattr(submit, "_resolve_data_revision", lambda: "test-rev")
+    expected_hash = "f" * 64
+    row = submit._build_pending_row(
+        submission_id="sub-test-x",
+        meta=_stub_meta(),
+        blob_url="https://huggingface.co/datasets/example/sub-test-x.zip",
+        submission_sha256=expected_hash,
+    )
+    assert row["submission_sha256"] == expected_hash
+
+
+def test_pending_row_preserves_existing_metadata(monkeypatch):
+    """Pre-Bundle-1+2 fields keep their values from meta + args.
+
+    Regression guard: a future refactor of ``_build_pending_row`` that
+    accidentally drops one of these keys would silently change the
+    schema of every row the Space writes.
+    """
+    monkeypatch.setattr(submit, "_resolve_data_revision", lambda: "test-rev")
+    meta = _stub_meta()
+    row = submit._build_pending_row(
+        submission_id="sub-test-x",
+        meta=meta,
+        blob_url="https://example.test/sub-test-x.zip",
+        submission_sha256="0" * 64,
+    )
+    assert row["submission_id"] == "sub-test-x"
+    assert row["status"] == "pending"
+    assert row["failure_reason"] is None
+    assert row["submitter_name"] == meta["submitter_name"]
+    assert row["submission_name"] == meta["submission_name"]
+    assert row["agent_url"] == meta["agent_url"]
+    assert row["notes"] == meta["notes"]
+    assert row["submission_blob_url"] == "https://example.test/sub-test-x.zip"
+    assert row["cadgenbench_data_revision"] == "test-rev"
+    # Score-shaped fields are null on a fresh pending row.
+    for k in (
+        "aggregate_score",
+        "validity_rate",
+        "score_by_task_type",
+        "per_task_scores",
+        "per_fixture_scores",
+        "per_fixture_breakdown",
+    ):
+        assert row[k] is None