@piwitests/reporter 0.49.0 → 0.50.0

package/dist/global-setup-module.js

@@ -648,11 +648,17 @@ function wildcard(pattern) {
   const escaped = pattern.split("*").map((part) => part.replace(/[.*+?^${}()|[\]\\]/g, "\\$&"));
   return new RegExp(`^${escaped.join(".*")}$`, "i");
 }
-function selectedProjects(projects, argv = process.argv) {
-  const names = cliProjectNames(argv);
-  if (!names || names.length === 0) return projects;
-  const patterns = names.map(wildcard);
-  const chosen = new Set(projects.filter((p) => patterns.some((re) => re.test(p.name ?? ""))));
+function selectedProjects(projects, argv = process.argv, filtered) {
+  let chosen;
+  if (Array.isArray(filtered)) {
+    const names = new Set(filtered.map((p) => p.name).filter(Boolean));
+    chosen = new Set(projects.filter((p) => filtered.includes(p) || !!p.name && names.has(p.name)));
+  } else {
+    const names = cliProjectNames(argv);
+    if (!names || names.length === 0) return projects;
+    const patterns = names.map(wildcard);
+    chosen = new Set(projects.filter((p) => patterns.some((re) => re.test(p.name ?? ""))));
+  }
   if (!argv.includes("--no-deps")) {
     const byName = new Map(projects.map((p) => [p.name ?? "", p]));
     const queue = [...chosen];
@@ -674,8 +680,9 @@ function asProxy(value) {
 }
 function baseUrlTargets(config, argv = process.argv) {
   const projects = config.projects?.length ? config.projects : [{ name: "", use: config.use }];
+  const filtered = config.projects?.length ? config.filteredProjects : void 0;
   const targets = /* @__PURE__ */ new Map();
-  for (const project of selectedProjects(projects, argv)) {
+  for (const project of selectedProjects(projects, argv, filtered)) {
     const url = project.use?.baseURL;
     if (typeof url !== "string" || !/^https?:\/\//i.test(url)) continue;
     const target = targets.get(url);

package/dist/index.d.ts

@@ -175,7 +175,9 @@ interface PiwiDashboardOptions {
      * runs no test and reaches no dashboard. The check runs in the reporter's
      * global setup, after `webServer` has started and before the config's own
      * `globalSetup`, so leave it off when that setup is what starts the app. Only
-     * the projects `--project` selects are checked. **Defaults to `false`**. Can
+     * the projects the run selects are checked, with their dependencies: the ones
+     * `--project` names, or every project but those declared `default: false`
+     * (Playwright 1.64+). **Defaults to `false`**. Can
      * also be set with `PIWI_CHECK_BASE_URL=true`.
      */
     checkBaseUrl?: boolean;

package/dist/index.js

@@ -3831,11 +3831,17 @@ function wildcard(pattern) {
   const escaped = pattern.split("*").map((part) => part.replace(/[.*+?^${}()|[\]\\]/g, "\\$&"));
   return new RegExp(`^${escaped.join(".*")}$`, "i");
 }
-function selectedProjects(projects, argv = process.argv) {
-  const names = cliProjectNames(argv);
-  if (!names || names.length === 0) return projects;
-  const patterns = names.map(wildcard);
-  const chosen = new Set(projects.filter((p) => patterns.some((re) => re.test(p.name ?? ""))));
+function selectedProjects(projects, argv = process.argv, filtered) {
+  let chosen;
+  if (Array.isArray(filtered)) {
+    const names = new Set(filtered.map((p) => p.name).filter(Boolean));
+    chosen = new Set(projects.filter((p) => filtered.includes(p) || !!p.name && names.has(p.name)));
+  } else {
+    const names = cliProjectNames(argv);
+    if (!names || names.length === 0) return projects;
+    const patterns = names.map(wildcard);
+    chosen = new Set(projects.filter((p) => patterns.some((re) => re.test(p.name ?? ""))));
+  }
   if (!argv.includes("--no-deps")) {
     const byName = new Map(projects.map((p) => [p.name ?? "", p]));
     const queue = [...chosen];
@@ -3857,8 +3863,9 @@ function asProxy(value) {
 }
 function baseUrlTargets(config, argv = process.argv) {
   const projects = config.projects?.length ? config.projects : [{ name: "", use: config.use }];
+  const filtered = config.projects?.length ? config.filteredProjects : void 0;
   const targets = /* @__PURE__ */ new Map();
-  for (const project of selectedProjects(projects, argv)) {
+  for (const project of selectedProjects(projects, argv, filtered)) {
     const url = project.use?.baseURL;
     if (typeof url !== "string" || !/^https?:\/\//i.test(url)) continue;
     const target = targets.get(url);
@@ -13693,6 +13700,97 @@ function compileFromCandidate(candidate) {
   return null;
 }
 
+// src/internal/ai/ground.ts
+var GROUNDED_METHODS = new Set(LOCATOR_METHODS.filter((method) => method !== "locator"));
+async function refLocator(page, ref) {
+  if (!/^[a-z0-9]+$/i.test(ref)) return null;
+  const getByRef = page.getByRef;
+  try {
+    const target = typeof getByRef === "function" ? getByRef.call(page, ref) : page.locator(`aria-ref=${ref}`);
+    return await target.count() === 1 ? target : null;
+  } catch {
+    return null;
+  }
+}
+async function matchesOnly(candidate, target) {
+  try {
+    return await candidate.count() === 1 && await candidate.and(target).count() === 1;
+  } catch {
+    return false;
+  }
+}
+async function includes(candidate, target) {
+  try {
+    return await candidate.and(target).count() === 1;
+  } catch {
+    return false;
+  }
+}
+function plainArg(arg, params, mask = true) {
+  switch (arg.type) {
+    case "string":
+      return mask ? maskValues(arg.value, params) : arg.value;
+    case "number":
+    case "boolean":
+      return arg.value;
+    case "object": {
+      const out = {};
+      for (const [key, value] of arg.entries) {
+        const plain = plainArg(value, params, mask);
+        if (plain === void 0) return void 0;
+        out[key] = plain;
+      }
+      return out;
+    }
+    default:
+      return void 0;
+  }
+}
+function chainToStructured(calls, params = {}) {
+  const links = [];
+  for (const call of calls) {
+    if (!GROUNDED_METHODS.has(call.method)) return null;
+    const args = [];
+    for (const arg of call.args) {
+      const plain = plainArg(arg, params, call.method !== "getByTestId");
+      if (plain === void 0) return null;
+      args.push(plain);
+    }
+    links.push({ method: call.method, args });
+  }
+  const [head, ...chain] = links;
+  if (!head) return null;
+  return chain.length > 0 ? { ...head, chain } : head;
+}
+async function normalizedLocator(target, params) {
+  if (typeof target.normalize !== "function") return null;
+  try {
+    const chain = tryParseLocatorChain(String(await target.normalize()));
+    return chain ? chainToStructured(chain.calls, params) : null;
+  } catch {
+    return null;
+  }
+}
+async function groundElement(page, element, params) {
+  const compiled = compileFromCandidate({
+    role: element.role,
+    name: element.name ?? null,
+    level: element.level ?? null
+  });
+  const target = element.ref ? await refLocator(page, element.ref) : null;
+  if (!target) return compiled;
+  if (compiled) {
+    const byName = buildLocator(page, compiled.locator, params);
+    if (await matchesOnly(byName, target)) return compiled;
+    if (!await includes(byName, target)) return compiled;
+  }
+  const normalized = await normalizedLocator(target, params);
+  if (!normalized || !await matchesOnly(buildLocator(page, normalized, params), target)) return compiled;
+  const fingerprint = compiled?.fingerprint ?? { role: element.role, name: element.name ?? null };
+  if (!compiled && element.level != null) fingerprint.level = element.level;
+  return { locator: normalized, fingerprint, score: compiled?.score ?? 0 };
+}
+
 // src/internal/ai/resolution.ts
 var MAX_SNAPSHOT_CHARS = 24e3;
 var DEFAULT_MAX_STEPS = 20;
@@ -13713,12 +13811,6 @@ var ServerStepResolver = class {
     return await res.json();
   }
 };
-function candidateOf(element) {
-  return { role: element.role, name: element.name ?? null, level: element.level ?? null };
-}
-function locatorFromElement(element) {
-  return compileFromCandidate(candidateOf(element));
-}
 function lazyLocator(resolveOnce) {
   let pending = null;
   const ensure = () => pending ??= resolveOnce();
@@ -13730,20 +13822,21 @@ function lazyLocator(resolveOnce) {
   };
   return new Proxy({}, handler);
 }
-function buildPostcondition(resolved, template) {
+async function buildPostcondition(resolved, template, ctx) {
   if (!resolved)
     throw new Error(`piwi AI: flow "${template}" resolved without a postcondition (the oracle is mandatory)`);
   if (resolved.assert === "url") {
     if (!resolved.url) throw new Error(`piwi AI: flow "${template}" postcondition is a url assert without a url`);
     return { assert: "url", url: resolved.url };
   }
-  const compiled = resolved.element ? locatorFromElement(resolved.element) : null;
+  const compiled = resolved.element ? await groundElement(ctx.page, resolved.element, ctx.params) : null;
   if (!compiled) throw new Error(`piwi AI: flow "${template}" postcondition element could not be compiled`);
   return { assert: resolved.assert, locator: compiled.locator };
 }
 var readMaskedSnapshot = async (page, params) => {
   const raw = await ariaSnapshotBestEffort(page.locator("body"));
-  return raw ? maskValues(raw, params) : "";
+  if (!raw) return "";
+  return raw.split(/(\[ref=\w+\])/).map((part, i) => i % 2 === 1 ? part : maskValues(part, params)).join("");
 };
 var observeXhrResponses = async (page, action, settleMs) => {
   const urls = [];
@@ -13795,7 +13888,7 @@ async function resolveLocator(template, ctx) {
     ...screenshot ? { screenshot } : {}
   });
   if (!decision.element) throw new Error(`piwi AI: resolver returned no element for "${template}"`);
-  const compiled = locatorFromElement(decision.element);
+  const compiled = await groundElement(ctx.page, decision.element, ctx.params);
   if (!compiled) throw new Error(`piwi AI: element for "${template}" could not be compiled to a stable locator`);
   const entry = {
     version: ARTIFACT_VERSION,
@@ -13828,14 +13921,14 @@ async function resolveRun(template, ctx) {
       ...screenshot ? { screenshot } : {}
     });
     if (decision.done) {
-      const postcondition = buildPostcondition(decision.postcondition, template);
+      const postcondition = await buildPostcondition(decision.postcondition, template, ctx);
       await assertPostcondition(postcondition, { page: ctx.page, params: ctx.params });
       return { version: ARTIFACT_VERSION, kind: "run", template, steps, postcondition };
     }
     if (!decision.element || !decision.action) {
       throw new Error(`piwi AI: resolver returned neither a step nor done for "${template}"`);
     }
-    const compiled = locatorFromElement(decision.element);
+    const compiled = await groundElement(ctx.page, decision.element, ctx.params);
     if (!compiled) throw new Error(`piwi AI: a step element for "${template}" could not be compiled`);
     const step = { locator: compiled.locator, action: decision.action, fingerprint: compiled.fingerprint };
     if (decision.value !== void 0) step.value = decision.value;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@piwitests/reporter",
-  "version": "0.49.0",
+  "version": "0.50.0",
   "description": "Playwright reporter that streams results, traces and HTML reports to a Piwi Dashboard instance",
   "url": "https://github.com/PiwiTests/platform",
   "homepage": "https://piwitests.dev",

package/templates/skills/investigate-failure/SKILL.md

@@ -28,7 +28,7 @@ Prefer the **Piwi MCP server** if it is connected to this agent (tools named `li
 7. **Verify.** Re-run the affected spec(s): `npx playwright test <file>`. Confirm they pass and the new run is green in the dashboard (`get_run_insights` compares against the last green run: regressions cleared, nothing new broken).
 
 8. **Report back.** Piwi learns from what you tell it, so end every investigation with these calls (they need reporter or admin access; each is logged with your API key):
-   - **Your diagnosis.** If the cluster had no diagnosis, or yours differs, record it with `record_diagnosis`: the model you run on and the diagnosis in the schema the tool describes (summary, confidenceScore, severity, ranked hypotheses with their evidence, suggestedFix with the unified-diff `patch`). Piwi validates the patch against the source that failed and shows the diagnosis as written by an agent.
+   - **Your diagnosis.** If the cluster had no diagnosis, or yours differs, record it with `record_diagnosis` (`clusterId`, or `executionId` for a failure in no cluster or with a cause of its own): the model you run on and the diagnosis in the schema the tool describes (summary, confidenceScore, severity, ranked hypotheses with their evidence, suggestedFix with the unified-diff `patch`). Piwi validates the patch against the source that failed and shows the diagnosis as written by an agent.
    - **Your rating of Piwi's diagnosis.** If you read one with `get_cluster_diagnosis`, rate it with `submit_diagnosis_feedback`: `up` when it pointed at the real cause, `down` with a one-line note on what it got wrong when it did not. A `down` keeps the next diagnosis from repeating it.
    - **Your fix.** Commit the change with the trailer `get_fix_plan` returns in `verify.commitTrailer` (`Piwi-Cluster: <clusterId>`) as the last line of the commit message, then call `report_fix_attempt` with the cluster, the kind (`patch`, `locator-edit` or `fix-plan`), the commit or the branch, the patch you applied and the `diagnosisId` you followed. When the tests pass on that commit, Piwi records the attempt verified and shows it on the cluster's activity.
    - Leave the cluster's status alone: the run that passes on your commit records the fix.