--retry-failed reaches the UI

The CLI could retry failed upload jobs; the web app couldn't, so a run
that hit a bug (twice today) left work only a terminal could reclaim.
/api/pipeline now reports the failed count, and the upload card grows a
"retry N failed" checkbox — shown only when there are failures — that
rides along with both Dry run and the real Upload. Help explains why
failures are skipped by default.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_016jXZFSTZQKzAC8fqpWSz9g

src/bggpipe/templates/pages/help.html

@@ -29,7 +29,7 @@
   <p><b><a href="/photos">Photos</a></b> — drag photos in, drop them in the <code>photos/</code> folder, or (on a paired phone) tap the drop zone and shoot straight from the camera; camera captures get unique <code>shelf-…</code> names so they can never overwrite each other. Each photo has its own page listing every title read from it and any reshoot tickets — boxes seen but not identified. Photograph those up close, drop the new shot in, and extract again. Re-uploading a photo under the same <i>file name</i> deliberately replaces it, and the next extract run re-reads it.</p>
   <p><b><a href="/titles">Titles</a></b> — every read off your shelves, alphabetized, with its status and photos. This is the proofread checkpoint: <a href="#curation">edit, split, remove</a>. Its badge counts <span class="chip shaky">shaky read</span> lines — the model wasn't sure and nothing has verified them; filter to them, then press <b>✓ looks right</b> or edit each one.</p>
   <p><b><a href="/review">Review</a></b> — the decisions only you can make: which game a title is, which edition a copy is, whether two same-game reads are really one box (merges show a veto), and whether an unmatched title is a real game BGG simply doesn't have (<b>keep locally</b>: it joins the Library, never uploads). Keyboard-first; see <a href="#keys">shortcuts</a>.</p>
-  <p><b><a href="/queue">Queue</a></b> — exactly what upload will do (new entries and version upgrades) and the log of everything it has done. Nothing reaches BGG that isn't visible here first.</p>
+  <p><b><a href="/queue">Queue</a></b> — exactly what upload will do (new entries and version upgrades) and the log of everything it has done. Nothing reaches BGG that isn't visible here first. A job that fails is skipped by later runs (so one broken game can't loop forever); when any exist, the Pipeline's upload card offers a <b>retry N failed</b> checkbox.</p>
   <p><b><a href="/library">Library</a></b> — your enriched collection: filter by board games or RPGs. RPG matches are identified and enriched but never uploaded — BGG collections can't hold them, so they stay local citizens.</p>
 </div>
 

src/bggpipe/templates/pages/pipeline.html

@@ -58,18 +58,27 @@ function render() {
     stageCard(5, "upload", uploadFacts,
       `${runBtn("upload", "Dry run")}
        <button class="danger" data-upload-real ${RUNNING || P.stub_data ? "disabled" : ""}>Upload</button>
-       <label>limit <input type="number" id="uplimit" min="1" placeholder="all"></label>`),
+       <label>limit <input type="number" id="uplimit" min="1" placeholder="all"></label>
+       ${P.upload_failed
+         ? `<label title="failed jobs are skipped on normal runs so a broken one can't loop">
+              <input type="checkbox" id="upretry"> retry ${P.upload_failed} failed</label>`
+         : ""}`),
     stageCard(6, "enrich", `<b>${P.games}</b> game(s) in the <a href="/library">library</a>`, runBtn("enrich")),
   ].join("");
 
   document.querySelectorAll("[data-run]").forEach(b =>
     b.addEventListener("click", () => runStage(b.dataset.run,
-      b.dataset.run === "upload" ? {dry_run: true} : {})));
+      b.dataset.run === "upload" ? {dry_run: true, ...uploadOpts()} : {})));
+  const uploadOpts = () => ({
+    limit: Number(document.getElementById("uplimit").value) || null,
+    retry_failed: !!document.getElementById("upretry")?.checked,
+  });
   const real = document.querySelector("[data-upload-real]");
   if (real) real.addEventListener("click", () => {
-    const limit = Number(document.getElementById("uplimit").value) || null;
+    const {limit, retry_failed} = uploadOpts();
-    if (!confirm(`Really add games to your live BGG collection${limit ? ` (limit ${limit})` : ""}? A browser window will open.`)) return;
+    if (!confirm(`Really add games to your live BGG collection${limit ? ` (limit ${limit})` : ""}` +
-    runStage("upload", {dry_run: false, limit});
+                 `${retry_failed ? ", retrying previously failed jobs" : ""}? A browser window will open.`)) return;
+    runStage("upload", {dry_run: false, limit, retry_failed});
   });
 
   const job = P.job;

src/bggpipe/webreview.py

@@ -194,6 +194,7 @@ class RemoveBody(BaseModel):
 class RunBody(BaseModel):
     dry_run: bool = True  # upload only; the safe direction is the default
     limit: int | None = None
+    retry_failed: bool = False
 
 
 class DismissBody(BaseModel):
@@ -222,10 +223,14 @@ def _default_stages(cfg: Config) -> dict[str, Callable[..., object]]:
 
         return run_diff(cfg)
 
-    def upload(dry_run: bool = True, limit: int | None = None) -> object:
+    def upload(
+        dry_run: bool = True,
+        limit: int | None = None,
+        retry_failed: bool = False,
+    ) -> object:
         from bggpipe.upload import run_upload
 
-        return run_upload(cfg, dry_run=dry_run, limit=limit)
+        return run_upload(cfg, dry_run=dry_run, limit=limit, retry_failed=retry_failed)
 
     def enrich() -> object:
         from bggpipe.enrich import run_enrich
@@ -824,6 +829,7 @@ def create_app(
                 "to_add": _csv_count(cfg.to_add_path),
                 "to_update": _csv_count(cfg.to_update_path),
                 "upload_log": dict(log_counts),
+                "upload_failed": log_counts.get("failed", 0),
                 "games": games,
                 "job": jobs.snapshot(),
             }
@@ -834,7 +840,12 @@ def create_app(
             raise HTTPException(404, f"unknown stage {stage!r}")
         body = body or RunBody()
         if stage == "upload":
-            fn = partial(stages["upload"], dry_run=body.dry_run, limit=body.limit)
+            fn = partial(
+                stages["upload"],
+                dry_run=body.dry_run,
+                limit=body.limit,
+                retry_failed=body.retry_failed,
+            )
         else:
             fn = stages[stage]
         if not jobs.start(stage, fn):

tests/test_web_dashboard.py

@@ -113,17 +113,21 @@ def test_upload_defaults_to_dry_run(tmp_path):
     calls = []
     jobs = JobRunner()
 
-    def upload(dry_run=True, limit=None):
+    def upload(dry_run=True, limit=None, retry_failed=False):
-        calls.append({"dry_run": dry_run, "limit": limit})
+        calls.append({"dry_run": dry_run, "limit": limit, "retry_failed": retry_failed})
 
     web = _app(cfg, stages={"upload": upload}, jobs=jobs)
     web.post("/api/run/upload")  # no body: the safe direction
     jobs.wait()
     web.post("/api/run/upload", json={"dry_run": False, "limit": 2})
     jobs.wait()
+    # a failed job is skipped on normal runs; the UI opts back in
+    web.post("/api/run/upload", json={"dry_run": False, "retry_failed": True})
+    jobs.wait()
     assert calls == [
-        {"dry_run": True, "limit": None},
+        {"dry_run": True, "limit": None, "retry_failed": False},
-        {"dry_run": False, "limit": 2},
+        {"dry_run": False, "limit": 2, "retry_failed": False},
+        {"dry_run": False, "limit": None, "retry_failed": True},
     ]