feat(mesher/client): add live dashboard flows

Author: snowdamiz

mesher/api/helpers.mpl

@@ -2,7 +2,7 @@
 # Provides common utilities used across search, dashboard, and team handlers.
 # All functions are pub for cross-module import.
 
-from Storage.Queries import get_project_id_by_slug
+from Storage.Queries import get_project_id_by_slug, get_org_id_by_slug
 
 # Cluster-aware registry lookup.
 # In cluster mode (Node.self returns non-empty), uses Global.whereis for
@@ -35,6 +35,23 @@ pub fn resolve_project_id(pool :: PoolHandle, raw_id :: String) -> String do
   end
 end
 
+# Resolve an org identifier to a UUID.
+# If the identifier is 36 chars (UUID format), returns it directly.
+# Otherwise, treats it as a slug and looks up the org UUID from the database.
+# Returns the UUID string on success, or an empty string if slug not found.
+
+pub fn resolve_org_id(pool :: PoolHandle, raw_id :: String) -> String do
+  if String.length(raw_id) == 36 do
+    raw_id
+  else
+    let result = get_org_id_by_slug(pool, raw_id)
+    case result do
+      Ok( uuid) -> uuid
+      Err( _) -> ""
+    end
+  end
+end
+
 # Extract optional query parameter with a default value.
 # Request.query returns Option<String>; case match to Some/None.
 

mesher/api/team.mpl

@@ -15,7 +15,8 @@ from Storage.Queries import (
   revoke_api_key
 )
 from Types.User import User, OrgMembership
-from Api.Helpers import query_or_default, to_json_array, require_param, get_registry, resolve_project_id
+from Types.Project import ApiKey
+from Api.Helpers import query_or_default, to_json_array, require_param, get_registry, resolve_org_id, resolve_project_id
 
 # --- Shared helpers (leaf functions first, per define-before-use requirement) ---
 # Serialize a member row (with user info) to JSON.
@@ -54,16 +55,32 @@ end
 
 # Serialize an API key row to JSON.
 # Fields: id, project_id, key_value, label, created_at, revoked_at
-# If revoked_at is empty string, emit null instead of quoted empty string.
+# Uses Json.encode + Json.get so string fields remain truthful values.
+
+fn api_key_row_to_api_key(row) -> ApiKey do
+  ApiKey {
+    id : Map.get(row, "id"),
+    project_id : Map.get(row, "project_id"),
+    key_value : Map.get(row, "key_value"),
+    label : Map.get(row, "label"),
+    created_at : Map.get(row, "created_at"),
+    revoked_at : if String.length(Map.get(row, "revoked_at")) == 0 do
+      None
+    else
+      Some(Map.get(row, "revoked_at"))
+    end
+  }
+end
 
 fn api_key_to_json(row) -> String do
-  let id = Map.get(row, "id")
-  let project_id = Map.get(row, "project_id")
-  let key_value = Map.get(row, "key_value")
-  let label = Map.get(row, "label")
-  let created_at = Map.get(row, "created_at")
-  let revoked_at_raw = Map.get(row, "revoked_at")
-  let revoked_at = if String.length(revoked_at_raw) == 0 do
+  let api_key_json = Json.encode(api_key_row_to_api_key(row))
+  let id = Json.get(api_key_json, "id")
+  let project_id = Json.get(api_key_json, "project_id")
+  let key_value = Json.get(api_key_json, "key_value")
+  let label = Json.get(api_key_json, "label")
+  let created_at = Json.get(api_key_json, "created_at")
+  let revoked_at_raw = Json.get(api_key_json, "revoked_at")
+  let revoked_at = if revoked_at_raw == "" or revoked_at_raw == "null" do
     "null"
   else
     "\"#{revoked_at_raw}\""
@@ -78,6 +95,12 @@ fn extract_json_field(pool :: PoolHandle, body :: String, field :: String) -> St
   Ok(Json.get(body, field))
 end
 
+# Helper: resolve org path identifiers (UUID or slug) and fail explicitly when unknown.
+
+fn resolve_team_org_id(pool :: PoolHandle, raw_id :: String) -> String do
+  resolve_org_id(pool, raw_id)
+end
+
 # --- Team membership helper functions for case arm extraction (ORG-04) ---
 # Helper: handle successful add_member result.
 
@@ -164,7 +187,11 @@ end
 fn perform_role_update(pool :: PoolHandle, membership_id :: String, role :: String) do
   let result = update_member_role(pool, membership_id, role)
   case result do
-    Ok( _) -> update_role_success()
+    Ok( count) -> if count > 0 do
+      update_role_success()
+    else
+      HTTP.response(400, json { error : "invalid role or membership not found" })
+    end
     Err( e) -> HTTP.response(500, json { error : e })
   end
 end
@@ -213,14 +240,19 @@ end
 pub fn handle_list_members(request) do
   let reg_pid = get_registry()
   let pool = PipelineRegistry.get_pool(reg_pid)
-  let org_id = require_param(request, "org_id")
-  let result = get_members_with_users(pool, org_id)
-  case result do
-    Ok( rows) -> HTTP.response(200,
-    rows
-      |> List.map(fn (row) do member_to_json(row) end)
-      |> to_json_array())
-    Err( e) -> HTTP.response(500, json { error : e })
+  let raw_id = require_param(request, "org_id")
+  let org_id = resolve_team_org_id(pool, raw_id)
+  if String.length(org_id) == 0 do
+    HTTP.response(404, json { error : "organization not found" })
+  else
+    let result = get_members_with_users(pool, org_id)
+    case result do
+      Ok( rows) -> HTTP.response(200,
+      rows
+        |> List.map(fn (row) do member_to_json(row) end)
+        |> to_json_array())
+      Err( e) -> HTTP.response(500, json { error : e })
+    end
   end
 end
 
@@ -231,9 +263,14 @@ end
 pub fn handle_add_member(request) do
   let reg_pid = get_registry()
   let pool = PipelineRegistry.get_pool(reg_pid)
-  let org_id = require_param(request, "org_id")
+  let raw_id = require_param(request, "org_id")
+  let org_id = resolve_team_org_id(pool, raw_id)
   let body = Request.body(request)
-  validate_add_member(pool, org_id, body)
+  if String.length(org_id) == 0 do
+    HTTP.response(404, json { error : "organization not found" })
+  else
+    validate_add_member(pool, org_id, body)
+  end
 end
 
 # Handle POST /api/v1/orgs/:org_id/members/:membership_id/role
@@ -242,9 +279,15 @@ end
 pub fn handle_update_member_role(request) do
   let reg_pid = get_registry()
   let pool = PipelineRegistry.get_pool(reg_pid)
+  let raw_id = require_param(request, "org_id")
+  let org_id = resolve_team_org_id(pool, raw_id)
   let membership_id = require_param(request, "membership_id")
   let body = Request.body(request)
-  do_update_role(pool, membership_id, body)
+  if String.length(org_id) == 0 do
+    HTTP.response(404, json { error : "organization not found" })
+  else
+    do_update_role(pool, membership_id, body)
+  end
 end
 
 # Handle POST /api/v1/orgs/:org_id/members/:membership_id/remove
@@ -253,11 +296,21 @@ end
 pub fn handle_remove_member(request) do
   let reg_pid = get_registry()
   let pool = PipelineRegistry.get_pool(reg_pid)
+  let raw_id = require_param(request, "org_id")
+  let org_id = resolve_team_org_id(pool, raw_id)
   let membership_id = require_param(request, "membership_id")
-  let result = remove_member(pool, membership_id)
-  case result do
-    Ok( _) -> remove_success()
-    Err( e) -> HTTP.response(500, json { error : e })
+  if String.length(org_id) == 0 do
+    HTTP.response(404, json { error : "organization not found" })
+  else
+    let result = remove_member(pool, membership_id)
+    case result do
+      Ok( count) -> if count > 0 do
+        remove_success()
+      else
+        HTTP.response(404, json { error : "membership not found" })
+      end
+      Err( e) -> HTTP.response(500, json { error : e })
+    end
   end
 end
 

mesher/client/README.md

@@ -2,7 +2,7 @@
 
 This package is the canonical TanStack dashboard app for Mesher.
 
-It runs as a Vite-powered TanStack Start app from `hyperpush-mono/mesher/client/` and stays on the existing mock-data and client-state contract. Do not add backend calls, server functions, or widened URL/search-param semantics here as part of path-migration work.
+It runs as a Vite-powered TanStack Start app from `hyperpush-mono/mesher/client/` and now exercises the Issues route through a provider-owned same-origin `/api/v1` seam for both live reads and the supported maintainer actions. The shell stays visually intact by overlaying truthful Mesher issue/detail/timeline state onto the existing mock-shaped dashboard model, so unsupported fields remain visible and explicitly shell-only instead of disappearing or pretending to be live.
 
 ## Package root
 
@@ -21,8 +21,9 @@ npm --prefix mesher/client ci
 npm --prefix mesher/client run dev
 npm --prefix mesher/client run build
 PORT=3001 npm --prefix mesher/client run start
-npm --prefix mesher/client run test:e2e:dev
-npm --prefix mesher/client run test:e2e:prod
+bash mesher/scripts/seed-live-issue.sh
+npm --prefix mesher/client run test:e2e:dev -- --grep "issues live"
+npm --prefix mesher/client run test:e2e:prod -- --grep "issues live"
 ```
 
 From this package directory:
@@ -32,40 +33,69 @@ npm ci
 vite dev
 vite build
 node server.mjs
-npm run test:e2e:dev
-npm run test:e2e:prod
+bash ../scripts/seed-live-issue.sh
+npm run test:e2e:dev -- --grep "issues live"
+npm run test:e2e:prod -- --grep "issues live"
 ```
 
 ## Runtime contract
 
 - `vite dev` starts the local dashboard dev server on port `3000` by default.
 - `vite build` produces the production bundle in `dist/`.
 - `node server.mjs` serves the built app and static assets from `dist/client`.
-- `test:e2e:dev` verifies the direct-entry route and shell-state parity against the dev server.
-- `test:e2e:prod` verifies the same parity contract against the built production server.
+- `test:e2e:dev` verifies the live Issues seam against the dev server and same-origin Mesher proxy.
+- `test:e2e:prod` verifies the same live seam against the built production server.
+- `bash mesher/scripts/seed-live-issue.sh` deterministically seeds the read-proof issue and resets the action-proof issue back to an open/unresolved state before live verification.
+
+## Live Issues seam
+
+- Overview reads go through same-origin `/api/v1/projects/default/*` routes.
+- Selecting an issue reads `/api/v1/issues/:issue_id/events?limit=1`, `/api/v1/events/:event_id`, and `/api/v1/issues/:issue_id/timeline` through the provider-owned state path.
+- Supported live maintainer actions are `Resolve`, `Reopen`, and `Ignore`, which call same-origin `/api/v1/issues/:issue_id/{resolve,unresolve,archive}` from the existing detail action row.
+- The UI intentionally uses a mixed live/mock overlay: live Mesher truth replaces supported fields, while unsupported shell sections keep explicit fallback values.
+- `AI Analysis`, issue-link chrome, bounty chrome, and the retained proof-rail buttons remain visible but shell-only; they are not claimed as live backend actions.
+- Backend read and mutation failures are surfaced through the mounted Radix toaster and `issues-shell` / detail-panel `data-*` attributes instead of silently reverting.
 
 ## Important files
 
 - `src/routes/` — TanStack route tree for the dashboard shell and direct-entry pages.
 - `src/router.tsx` and `src/routeTree.gen.ts` — router assembly.
 - `server.mjs` — package-local production bridge for the built app.
 - `playwright.config.ts` — package-local dev/prod Playwright harness.
-- `tests/e2e/dashboard-route-parity.spec.ts` — route/UI parity proof on mock data.
+- `tests/e2e/issues-live-read.spec.ts` — live Issues read proof, including detail/timeline reads, sparse fallback retention, and failure toasts.
+- `tests/e2e/issues-live-actions.spec.ts` — live Issues action proof for resolve/reopen/ignore, same-origin routing, and destructive toast coverage.
+- `components/dashboard/dashboard-issues-state.tsx` — provider-owned overview + selected-issue Mesher read orchestration.
+- `components/ui/toaster.tsx` and `hooks/use-toast.ts` — mounted Radix toast surface reused for live read and mutation failures.
 - `app/globals.css` — shared global styles imported by the TanStack root route.
 
 ## Verification notes
 
-The canonical proof surface for this package is package-local:
+The canonical full-shell proof rail layers both seed helpers with the assembled walkthrough and the existing live Issues/admin+ops suites:
 
 ```bash
-npm --prefix mesher/client run build
-npm --prefix mesher/client run test:e2e:dev
-npm --prefix mesher/client run test:e2e:prod
+bash mesher/scripts/seed-live-issue.sh
+bash mesher/scripts/seed-live-admin-ops.sh
+npm --prefix mesher/client run test:e2e:dev -- --grep "issues live|admin and ops live|seeded walkthrough"
+npm --prefix mesher/client run test:e2e:prod -- --grep "issues live|admin and ops live|seeded walkthrough"
+```
+
+- `tests/e2e/seeded-walkthrough.spec.ts` is the canonical route-to-route shell proof. It covers direct-entry and in-app navigation parity across Issues, Performance, Solana Programs, Releases, Alerts, Bounties, Treasury, and Settings.
+- `tests/e2e/live-runtime-helpers.ts` owns the shared same-origin request tracking, direct-backend rejection, and filtered runtime diagnostics used by the walkthrough and route-level live suites.
+- `bash mesher/scripts/seed-live-issue.sh` resets the seeded live Issues read/action fixtures.
+- `bash mesher/scripts/seed-live-admin-ops.sh` resets the seeded Alerts and Settings admin/ops fixtures used by the assembled rail.
+
+The narrower route-level verification commands remain useful when you are only iterating on the Issues seam:
+
+```bash
+bash mesher/scripts/seed-live-issue.sh
+npm --prefix mesher/client run test:e2e:dev -- --grep "issues live"
+npm --prefix mesher/client run test:e2e:prod -- --grep "issues live"
 ```
 
-When a move or refactor breaks the package contract, expect the first signal to appear as one of:
+When the seam regresses, the first signal should appear as one of:
 
-- a missing file under `mesher/client/`
+- a failed same-origin `/api/v1` request in Playwright request tracking
+- a visible destructive toast for mutation or selected-issue read failures
+- a mismatched `issues-shell` or `issue-detail-panel` `data-*` attribute
 - a broken `dev`, `build`, `start`, or `test:e2e:*` script
-- a Playwright console/request failure in `tests/e2e/dashboard-route-parity.spec.ts`
 - a `node server.mjs` boot failure after build