docs: consolidate tasks and clean up legacy files

This commit consolidates all outstanding tasks from `TODO.md`, `BUILD_ISSUES.md`, and inline `TODO` comments into a single `TASKS.md` file. This new file serves as the single source of truth for all pending development work, categorized by priority.

The task list has been enhanced with additional context, linking tasks to the architectural vision in `refactor.md` and making them more actionable for developers.

The following legacy files have been removed to avoid confusion:
- `TODO.md`
- `BUILD_ISSUES.md`

The `AGENTS.md` file has also been updated to remove a temporary debug log and now directs agents to the new `TASKS.md` file for guidance.

TASKS.md

@@ -1,75 +1,72 @@
 # Consolidated Tasks for Tercul
 
-This document is the single source of truth for all outstanding development tasks. It is generated from `TODO.md`, `BUILD_ISSUES.md`, and inline `TODO` comments.
+This document is the single source of truth for all outstanding development tasks, aligned with the architectural vision in `refactor.md`.
 
 ---
 
 ##  Urgent: Build Failures
 
-These issues are currently breaking the build and must be resolved before any other work can proceed. All issues are located in `internal/adapters/graphql/schema.resolvers.go`.
+These issues are currently breaking the build and must be resolved before any other work can proceed. They are all related to the ongoing refactor, where repositories are being replaced by application services. The `r.App` object in the resolver does not have direct repository access anymore.
 
-- [ ] **Resolver Errors:**
-    - [ ] `internal/adapters/graphql/schema.resolvers.go:10:2: "log" imported and not used`
-    - [ ] `internal/adapters/graphql/schema.resolvers.go:1071:24: r.App.AuthorRepo undefined (type *app.Application has no field or method AuthorRepo)`
-    - [ ] `internal/adapters/graphql/schema.resolvers.go:1073:24: r.App.AuthorRepo undefined (type *app.Application has no field or method AuthorRepo)`
-    - [ ] `internal/adapters/graphql/schema.resolvers.go:1089:36: r.App.Localization.GetAuthorBiography undefined (type *"tercul/internal/app/localization".Service has no field or method GetAuthorBiography)`
-    - [ ] `internal/adapters/graphql/schema.resolvers.go:1141:22: r.App.UserRepo undefined (type *app.Application has no field or method UserRepo)`
-    - [ ] `internal/adapters/graphql/schema.resolvers.go:1143:24: r.App.UserRepo undefined (type *app.Application has no field or method UserRepo)`
-    - [ ] `internal/adapters/graphql/schema.resolvers.go:1212:20: r.App.TagRepo undefined (type *app.Application has no field or method TagRepo)`
-    - [ ] `internal/adapters/graphql/schema.resolvers.go:1225:32: r.App.TagRepo undefined (type *app.Application has no field or method TagRepo)`
-    - [ ] `internal/adapters/graphql/schema.resolvers.go:1249:25: r.App.CategoryRepo undefined (type *app.Application has no field or method CategoryRepo)`
-    - [ ] `internal/adapters/graphql/schema.resolvers.go:1262:32: r.App.CategoryRepo undefined (type *app.Application has no field or method CategoryRepo)`
+- [ ] **Fix Resolver Build Errors in `internal/adapters/graphql/schema.resolvers.go`:**
+    - [ ] `line 10`: Remove the unused "log" import.
+    - [ ] `lines 1071, 1073`: Replace `r.App.AuthorRepo` calls with the appropriate application service method (e.g., `r.App.Authors.FindByID`).
+    - [ ] `line 1089`: Correct the call to `r.App.Localization.GetAuthorBiography` as the method does not exist. The correct service and method must be identified.
+    - [ ] `lines 1141, 1143`: Replace `r.App.UserRepo` calls with the correct user application service method.
+    - [ ] `lines 1212, 1225`: Replace `r.App.TagRepo` calls with the correct tag application service method.
+    - [ ] `lines 1249, 1262`: Replace `r.App.CategoryRepo` calls with the correct category application service method.
 
 ---
 
 ## High Priority
 
-### Architecture & Refactoring
+### Architecture & Refactoring (see `refactor.md`)
 - [ ] **Implement Full Observability:**
-  - [ ] Centralize logging across the application.
-  - [ ] Add Prometheus metrics for key operations.
-  - [ ] Add OpenTelemetry tracing for distributed request tracing.
+  - [ ] **Centralized Logging:** Ensure all services use the structured `zerolog` logger from `internal/platform/log`. Add request/user/span IDs to the logging context in the HTTP middleware.
+  - [ ] **Metrics:** Add Prometheus metrics for API request latency, error rates, and database query performance. Expose them on the `/metrics` endpoint.
+  - [ ] **Tracing:** Instrument all application services and data layer methods with OpenTelemetry tracing.
 - [ ] **Establish CI/CD Pipeline:**
-  - [ ] Add `make lint test test-integration` to the CI pipeline.
-  - [ ] Set up automated deployments to a staging environment.
+  - [ ] **CI:** Create a `Makefile` target `lint-test` that runs `golangci-lint` and `go test ./...`. Configure the CI pipeline to run this on every push.
+  - [ ] **CD:** Set up automated deployments to a staging environment upon a successful merge to the main branch.
 - [ ] **Refactor Testing Utilities:**
-  - [ ] Remove database logic from `internal/testutil/testutil.go` to support mock-based integration tests.
+  - [ ] **Decouple from DB:** Remove all database connection logic from `internal/testutil/testutil.go`. Tests should use mock repositories and services, not a live database. This is a key step towards faster, more reliable unit tests.
 
 ### Features
 - [ ] **Implement Analytics Features:**
-  - [ ] Implement view counting for works and translations.
-  - [ ] Implement like counting for works and translations.
-  - [ ] Implement comment counting for works.
-  - [ ] Implement bookmark counting for works.
-  - [ ] Implement translation counting for works.
-  - [ ] Implement translation analytics to show popular translations.
-  - *Note: This is referenced in both `TODO.md` and `internal/jobs/linguistics/work_analysis_service.go`.*
+  - **Context:** This is required for user engagement insights. The following counts need to be implemented and stored, likely on the `Work` and `Translation` models.
+  - [ ] Implement view counting.
+  - [ ] Implement like counting.
+  - [ ] Implement comment counting.
+  - [ ] Implement bookmark counting.
+  - [ ] Implement translation counting.
+  - [ ] Implement a service to calculate popular translations based on the above metrics.
+  - *Note: This is referenced in the old `TODO.md` and a TODO comment in `internal/jobs/linguistics/work_analysis_service.go`.*
 
 - [ ] **Complete Unfinished Implementations:**
-    - [ ] `internal/app/work/commands.go`: Implement the `MergeWork` function.
-    - [ ] `internal/app/search/service.go`: Get content from the translation service.
-    - [ ] `cmd/tools/enrich/main.go`: The tool is broken and needs to be fixed.
+    - [ ] `internal/app/work/commands.go`: Implement the `MergeWork` command. This should handle merging duplicate work entries.
+    - [ ] `internal/app/search/service.go`: The search service needs to fetch content from the translation service to enrich its search index.
+    - [ ] `cmd/tools/enrich/main.go`: This command-line tool is broken. It needs to be investigated, fixed, and documented.
 
 ---
 
 ## Medium Priority
 
 ### Performance
-- [ ] **Batch Weaviate Operations:** Implement batching for Weaviate client operations to improve throughput.
-- [ ] **Add Performance Benchmarks:** Create benchmarks for critical API paths and business logic to monitor performance over time.
+- [ ] **Batch Weaviate Operations:** Refactor the Weaviate client in `internal/platform/search` to support batching of create/update operations to improve indexing performance.
+- [ ] **Add Performance Benchmarks:** Using Go's built-in benchmarking tools, add benchmarks for critical API queries and commands to detect performance regressions.
 
 ### Code Quality & Architecture
-- [ ] **Expand Weaviate Client:** Add support for all models to the Weaviate client.
-- [ ] **Add Documentation:** Improve code documentation and generate API documentation.
-- [ ] **Refactor Caching:** Replace bespoke cached repositories with the decorator pattern in `internal/data/cache`.
-- [ ] **Improve Configuration Handling:** Replace ad-hoc config with environment variable parsing and validation (e.g., using `koanf` or `envconfig`).
+- [ ] **Expand Weaviate Client:** The client in `internal/platform/search` should be extended to support indexing all relevant domain models (e.g., Authors, Tags).
+- [ ] **Add Documentation:** Add GoDoc comments to all public functions and types in the `internal/app` and `internal/domain` packages.
+- [ ] **Refactor Caching:** As per `refactor.md`, replace the current bespoke cached repositories with a decorator pattern in `internal/data/cache`. This will simplify cache invalidation logic.
+- [ ] **Improve Configuration Handling:** Replace the global config object with struct-based configuration loaded from environment variables (e.g., using `koanf` or `envconfig`), as outlined in `refactor.md`.
 
 ### Monitoring & Logging
-- [ ] **Add Monitoring:** Add monitoring for background jobs and API endpoints.
-- [ ] **Add Linguistics Metrics:** Add metrics for linguistics pipeline (analysis duration, cache hit/miss, provider usage).
+- [ ] **Add Job Monitoring:** Add specific monitoring and alerting for background jobs in `internal/jobs` to track success, failure, and duration.
+- [ ] **Add Linguistics Metrics:** Add Prometheus metrics for the linguistics pipeline, including analysis duration, cache hit/miss rates, and third-party API usage.
 
 ---
 
 ## Low Priority
 
-- [ ] **Refactor Transactional Runner:** Refactor `RunTransactional` in `internal/testutil/testutil.go` to be more mock-friendly.
+- [ ] **Refactor Transactional Runner:** Refactor the `RunTransactional` helper in `internal/testutil/testutil.go` to be more friendly to mock-based testing, likely by allowing a mock DB transaction to be passed in.