Upload 6 files

CHANGELOG.md

@@ -201,3 +201,12 @@
 - Run Explorer filters and search now operate locally from a cached run list for a much snappier interaction.
 - Refresh now explicitly reloads the connected bucket; ordinary filter/search changes no longer re-query the bucket.
 - Selecting a run highlights it, switches the detail/progress context, and keeps the build form independent.
+
+## V42 — Product tabs and run recovery hardening
+
+- Renamed the custom UI top-level workflow tabs to the intended product model: **New Build**, **Live Progress**, and **Space Test**.
+- Kept the build form isolated from selected historical runs; selecting a run only updates Live Progress and details.
+- Hardened run discovery from the connected bucket by combining `ls` and targeted `glob` patterns so partial/in-progress runs are recovered more reliably.
+- Persisted `launch.json`, `summary.json`, and a minimal `state.json` immediately after a Job launch, so running Jobs appear in the Run Explorer before the worker writes final artifacts.
+- Progress polling now falls back to launch metadata when `state.json`/`events.jsonl` are not ready yet.
+- Job links are preserved or reconstructed from launch metadata for active and historical runs.

README.md

@@ -159,13 +159,17 @@ The custom dashboard separates two user intents more clearly:
 - Selecting a previous run in the Run Explorer only updates progress, details, links, and report preview. It does not modify the build form.
 - Run filtering/search is now local and cached after the latest bucket refresh, so status filters should feel instant. Use **Refresh runs** to reload the bucket.
 
-## V40 product UI
+## Current product UI
 
-The public custom UI now uses three top-level tabs instead of a left sidebar:
+The public custom UI uses three top-level workflow tabs:
 
-1. **Run new build** — prepare and launch a fresh private Space build from a model card.
-2. **Live progress** — inspect active or historical runs, follow the compact step timeline, open Job/Space/Settings/Artifacts, and browse the connected bucket's Run Explorer.
-3. **Test Space** — run a live smoke test after manual hardware selection or after a generated Space is ready.
+1. **New Build** — prepare and launch a fresh private Space build from a model card.
+2. **Live Progress** — inspect active or historical runs, follow the compact step timeline, open Job/Space/Settings/Artifacts, and browse the connected bucket's Run Explorer.
+3. **Space Test** — run a live smoke test after manual hardware selection or after a generated Space is ready.
+
+The build form is intentionally isolated from selected historical runs. Selecting a run only updates Live Progress, details, quick links, and report preview; it never mutates the New Build form.
+
+The Run Explorer reads launch metadata, summary metadata, state, gate files, and validation files from the connected user's bucket. New Jobs write launch metadata immediately so in-progress runs can appear before the worker has finished writing final artifacts.
 
 The live progress timeline is designed to avoid horizontal scrolling and uses explicit step states: completed, active, pending, and failed. Historical run Job links are reconstructed from `job_id` when possible, so the Job button should remain available even when older summaries did not store a full `job_url`.
 

app.py

@@ -149,7 +149,7 @@ def register_custom_routes(fastapi_app: FastAPI) -> None:
         return JSONResponse(
             {
                 "name": "Agentic Space Factory",
-                "version": "v40-tabbed-product-ui",
+                "version": "v42-tabbed-runs-recovery",
                 "bucket_default": settings.bucket_name,
                 "workflows": ["build_from_model_card", "validate_existing_space", "runs_explorer"],
                 "custom_ui_status": "root_custom_ui",
@@ -332,14 +332,17 @@ def register_custom_routes(fastapi_app: FastAPI) -> None:
         bucket_source = user_bucket_source(username=ctx["username"], bucket_name=bucket_name)
         bundle = read_run_bundle(run_id, bucket_source=bucket_source, token=ctx["token"])
         state = bundle.get("state") or {}
+        launch = bundle.get("launch") or {}
+        effective_state = {**launch, **state}
         events = bundle.get("events") or []
-        progress = progress_from_events(events, state=state)
+        progress = progress_from_events(events, state=effective_state)
+        summary = bundle.get("summary") or {}
         progress.update(
             {
                 "run_id": run_id,
                 "bucket_source": bucket_source,
-                "state": state,
-                "summary": bundle.get("summary") or {},
+                "state": effective_state,
+                "summary": summary,
                 "inference_gate": bundle.get("inference_gate") or {},
                 "generation_smoke": bundle.get("generation_smoke") or {},
                 "hardware_strategy": bundle.get("hardware_strategy") or {},
@@ -348,8 +351,8 @@ def register_custom_routes(fastapi_app: FastAPI) -> None:
                 "links": _api_links(
                     run_id=run_id,
                     bucket_source=bucket_source,
-                    target_space=state.get("target_space") or (bundle.get("summary") or {}).get("target_space"),
-                    job_url=state.get("job_url") or (bundle.get("summary") or {}).get("job_url"),
+                    target_space=effective_state.get("target_space") or summary.get("target_space"),
+                    job_url=effective_state.get("job_url") or summary.get("job_url"),
                 ),
             }
         )