chore(agentic): remove unused sweep summary collection

Author: cquil11

.github/workflows/e2e-tests.yml

@@ -312,42 +312,8 @@ jobs:
         uses: ./.github/workflows/collect-evals.yml
         secrets: inherit
 
-    collect-agentic-results:
-        needs: [test-sweep-agentic, test-sweep-multi-node-agentic]
-        if: ${{ always() && (needs.test-sweep-agentic.result != 'skipped' || needs.test-sweep-multi-node-agentic.result != 'skipped') }}
-        runs-on: ubuntu-latest
-        steps:
-            - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
-              with:
-                  submodules: true
-
-            - uses: actions/setup-python@v6
-              with:
-                  python-version: '3.11'
-
-            - name: Install dependencies
-              run: pip install pandas matplotlib numpy
-
-            - name: Download agentic artifacts
-              uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8.0.1
-              with:
-                  pattern: 'agentic_*'
-                  path: results/
-
-            - name: Run aggregation
-              env:
-                  PYTHONPATH: utils/agentic-benchmark/scripts:utils/agentic-benchmark/analysis
-              run: |
-                  python utils/agentic-benchmark/scripts/collect_sweep_results.py results/ aggregated/
-
-            - name: Upload aggregated results
-              uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
-              with:
-                  name: agentic_aggregated
-                  path: aggregated/
-
     calc-success-rate:
-        needs: [collect-results, collect-evals, collect-agentic-results]
+        needs: [collect-results, collect-evals]
         if: ${{ always() }}
         runs-on: ubuntu-latest
 

utils/test_find_reusable_sweep_run.py

@@ -207,13 +207,11 @@ def test_artifact_names_excludes_expired_artifacts(monkeypatch) -> None:
         "paginated_github_api",
         lambda *args, **kwargs: [
             {"name": "results_bmk", "expired": True},
-            {"name": "agentic_aggregated", "expired": False},
+            {"name": "run-stats", "expired": False},
         ],
     )
 
-    assert reuse.artifact_names("repo", 123, "token") == {
-        "agentic_aggregated"
-    }
+    assert reuse.artifact_names("repo", 123, "token") == {"run-stats"}
 
 
 def test_main_skips_pr_synchronize_with_reuse_authorization(

utils/test_validate_reusable_sweep_artifacts.py

@@ -197,8 +197,6 @@ def test_multinode_agentic_identity_fields_match() -> None:
 def write_agentic_artifacts(
     root: Path,
     conc: int = 16,
-    *,
-    aggregate: bool = True,
 ) -> None:
     result_name = f"dsv4_tp8_conc{conc}_offloadcpu_result"
     point_dir = root / f"bmk_agentic_{result_name}"
@@ -207,12 +205,6 @@ def write_agentic_artifacts(
         json.dumps(agentic_result(conc))
     )
     (root / f"agentic_{result_name}").mkdir()
-    if aggregate:
-        aggregate_dir = root / "agentic_aggregated"
-        aggregate_dir.mkdir()
-        (aggregate_dir / "summary.csv").write_text(
-            f"exp_name,status\nagentic_{result_name},SUCCESS\n"
-        )
 
 
 def test_eval_validation_requires_raw_result_dirs_not_eval_debug_dirs(
@@ -398,7 +390,7 @@ def test_fixed_sequence_validation_rejects_duplicate_identity(
     assert "fixed-sequence artifacts contain 1 duplicate row(s)" in errors
 
 
-def test_agentic_validation_checks_points_raw_and_aggregate(tmp_path: Path) -> None:
+def test_agentic_validation_checks_points_and_raw_artifacts(tmp_path: Path) -> None:
     write_agentic_artifacts(tmp_path)
 
     assert validate_agentic_artifacts(tmp_path) == []
@@ -407,7 +399,7 @@ def test_agentic_validation_checks_points_raw_and_aggregate(tmp_path: Path) -> N
 def test_agentic_validation_accepts_run_sweep_point_artifacts(
     tmp_path: Path,
 ) -> None:
-    write_agentic_artifacts(tmp_path, aggregate=False)
+    write_agentic_artifacts(tmp_path)
 
     assert validate_agentic_artifacts(tmp_path) == []
 
@@ -420,8 +412,6 @@ def test_agentic_validation_accepts_additional_source_identity(
     extra_dir.mkdir()
     (extra_dir / "extra.json").write_text(json.dumps(agentic_result(32)))
     (tmp_path / "agentic_extra").mkdir()
-    summary = tmp_path / "agentic_aggregated" / "summary.csv"
-    summary.write_text(summary.read_text() + "agentic_extra,SUCCESS\n")
 
     assert validate_agentic_artifacts(tmp_path) == []
 
@@ -443,7 +433,7 @@ def test_agentic_validation_requires_point_and_raw_artifacts(
 def test_agentic_validation_rejects_duplicate_point_identity(
     tmp_path: Path,
 ) -> None:
-    write_agentic_artifacts(tmp_path, aggregate=False)
+    write_agentic_artifacts(tmp_path)
     point_dir = (
         tmp_path / "bmk_agentic_dsv4_tp8_conc16_offloadcpu_result"
     )

utils/validate_reusable_sweep_artifacts.py

@@ -4,7 +4,6 @@
 from __future__ import annotations
 
 import argparse
-import csv
 import json
 import re
 import shutil
@@ -250,9 +249,7 @@ def validate_agentic_artifacts(
     raw_names = {
         path.name
         for path in artifacts_dir.iterdir()
-        if path.is_dir()
-        and path.name.startswith("agentic_")
-        and path.name != "agentic_aggregated"
+        if path.is_dir() and path.name.startswith("agentic_")
     }
     if point_names != raw_names:
         missing_raw = point_names - raw_names
@@ -262,36 +259,6 @@ def validate_agentic_artifacts(
         for name in sorted(extra_raw):
             errors.append(f"unexpected raw agentic artifact dir: {name}")
 
-    aggregate_dir = artifacts_dir / "agentic_aggregated"
-    summary_path = aggregate_dir / "summary.csv"
-    if aggregate_dir.exists():
-        if not summary_path.is_file():
-            errors.append("missing agentic_aggregated/summary.csv")
-        else:
-            with open(summary_path, newline="") as handle:
-                summary_rows = [
-                    str(row.get("exp_name") or "")
-                    for row in csv.DictReader(handle)
-                    if row.get("exp_name")
-                ]
-            duplicate_names = [
-                name
-                for name, count in Counter(summary_rows).items()
-                if count > 1
-            ]
-            for name in sorted(duplicate_names):
-                errors.append(
-                    f"agentic aggregate has duplicate experiment: {name}"
-                )
-            summary_names = set(summary_rows)
-            if summary_names != raw_names:
-                for name in sorted(raw_names - summary_names):
-                    errors.append(f"agentic aggregate is missing experiment: {name}")
-                for name in sorted(summary_names - raw_names):
-                    errors.append(
-                        f"agentic aggregate has unexpected experiment: {name}"
-                    )
-
     return errors