feat: institutional + support + training admin sections (backend + frontend)

Ship three fully-wired admin areas end-to-end with APIs, seeds, tests and docs.

Backend (new `encoach_lms_api` addon + existing addons):
- Institutional: academic years/terms, departments, admission registers & admissions,
  courses/batches, lessons, fees (terms + student fees + invoicing with income-account
  auto-wiring), gradebook (assignments/grades), library, facilities (encoach.asset),
  student leave, result templates + marksheets (incl. delete-with-cascade).
- Support: `encoach.ticket` model + CRUD/assignee routes; payment records derived
  from `op.student.fees.details` and `account.move`; platform settings backed by
  `encoach.code` and `ir.config_parameter` (packages + grading config).
- Training: `encoach.vocab.item` + `encoach.grammar.rule` (plus progress models)
  with CRUD, pagination, search/level filters, and upsert-style progress endpoints.
  Odoo 19 compatibility: `_sql_constraints` replaced with `@api.constrains`;
  `ValidationError`/`UserError` mapped to HTTP 400.

Frontend:
- Rewire institutional admin pages (Academic Year Manager, Admissions, Courses,
  Lessons, Fees, Gradebook, Library, Facilities, Student Leave, Marksheets,
  Taxonomy, Resources) to real APIs with React Query invalidation and dialogs.
- New typed services: `payments.service.ts`, `platformSettings.service.ts`,
  `training.service.ts`. Updated `fees/gradebook/lms/courseware/taxonomy/
  resources/student-progress/generation` services + related types.
- Rewrite `VocabularyPage`, `GrammarPage`, `PaymentRecordPage`, `SettingsPage`,
  `TicketsPage` to consume live data with search/filter/progress/CRUD flows.
- New shared components: `TaxonomyCascade`, `MaterialViewer`, `teacher/TeacherLibrary`.
- Favicons/branding assets and misc. UX polish across teacher/student pages.

Tooling & QA:
- Seeders: `seed_demo.py`, `seed_demo_data.py`, `seed_institutional.py` (idempotent,
  covers institutional + support + training fixtures incl. income-account wiring).
- API write-flow test suites: `test_write_flows.py` (institutional),
  `test_support_flows.py` (support), `test_training_flows.py` (training),
  `test_ai_full.py`. All suites pass end-to-end.
- Docs: add `docs/PROJECT_SUMMARY.md` with per-section scope, artifacts and QA.
- `.gitignore`: ignore `pgdata_bak_*/`, `frontend/.vite/`, `frontend/dist/`,
  `frontend/node_modules/`.

Made-with: Cursor

custom_addons/encoach_core/models/permission.py

@@ -6,7 +6,9 @@ class EncoachPermission(models.Model):
     _description = "EnCoach Permission"
 
     code = fields.Char(required=True, index=True)
-    topic = fields.Char()
+    name = fields.Char()
+    description = fields.Text()
+    topic = fields.Char(string="Category")
     legacy_id = fields.Char(index=True)
 
     _code_unique = models.Constraint(

custom_addons/encoach_core/models/role.py

@@ -1,4 +1,4 @@
-from odoo import models, fields
+from odoo import models, fields, api
 
 
 class EncoachRole(models.Model):
@@ -6,10 +6,22 @@ class EncoachRole(models.Model):
     _description = "EnCoach Role"
 
     name = fields.Char(required=True)
+    code = fields.Char(index=True)
+    description = fields.Text()
     entity_id = fields.Many2one("encoach.entity", required=True, ondelete="cascade")
     permission_ids = fields.Many2many("encoach.permission", string="Permissions")
+    user_ids = fields.Many2many(
+        "res.users", "encoach_role_user_rel", "role_id", "user_id",
+        string="Users",
+    )
     is_default = fields.Boolean(default=False)
     legacy_id = fields.Char(index=True)
+    user_count = fields.Integer(compute="_compute_user_count", store=True)
+
+    @api.depends("user_ids")
+    def _compute_user_count(self):
+        for rec in self:
+            rec.user_count = len(rec.user_ids)
 
     def to_encoach_dict(self):
         self.ensure_one()