Keep DTC Operations docs app running from merged repo

[alexeygrigorev]

Keep DTC Operations docs app running from merged repo

Status: pending
Tags: migration, portal, process-docs, frontend, backend, testing, P0
Depends on: #2
Blocks: Phase 1 DTC Operations import acceptance and later portal integration work

Scope

Validate and repair the imported DTC Operations docs app after the source is copied into dataops by #2. The merged repo must run the existing process-doc portal from repo-local paths and configuration, preserving docs navigation, search, lint, local editing, and the existing backend/frontend contract.

This issue should make the app runnable and testable from /home/alexey/git/dataops without requiring runtime reads from ../dtc-operations. Update configuration, import paths, scripts, and focused regression tests as needed, but keep the product behavior equivalent to the imported DTC Operations docs app.

Acceptance Criteria

• The docs app starts locally from dataops using repo-local backend, frontend, and content paths, with no runtime dependency on ../dtc-operations.
• Existing process-doc navigation and document detail views load content from merged-repo content/.
• Search index generation or loading works from merged content and returns relevant process-doc results in the local app.
• SOP/process-doc linting runs from dataops and reports results using merged-repo paths.
• Local-mode document editing can open, modify, save, and reload a process doc without calling production GitHub write APIs.
• Existing image or asset handling used by the docs app still resolves files from the merged repo.
• Regression coverage is added or updated for backend docs/search behavior and at least one Playwright portal flow covering load, search, and document edit/save.
• Any required local run/test commands are documented in the repo location expected by the existing project conventions.

Test Scenarios

Scenario: Operator starts the merged docs app locally

Given: #2 has imported DTC Operations into dataops
When: an operator runs the documented local start command from /home/alexey/git/dataops
Then: the docs app starts successfully and serves the process-doc portal without reading from ../dtc-operations

Scenario: Operator searches process documentation

Given: the local docs app is running with merged-repo content
When: an operator searches for a known SOP or process term
Then: relevant process documents appear and open from the merged content/ tree

Scenario: Operator edits a process document locally

Given: the local docs app is running in local edit mode
When: an operator edits a process doc, saves it, and reloads the document
Then: the saved change is visible locally and no production GitHub write is attempted

Scenario: Engineer runs regression checks

Given: dependencies are installed for the imported docs app
When: an engineer runs the backend docs/search tests, SOP lint, and focused Playwright flow
Then: all checks pass from the merged repo path

Out of Scope

• Importing DTC Operations source files into dataops; that belongs to Import DTC Operations into dataops #2.
• Adding DataTasks, podcast assistant, or cross-system workflow features.
• Rewriting the frontend framework or replacing the existing vanilla JavaScript portal shell.
• Production deployment cutover, DNS changes, or real GitHub content writes.
• Broad cleanup or refactoring beyond what is needed to keep the imported docs app running.

[alexeygrigorev]

Grooming complete: scoped as a P0 migration/portal regression issue for keeping the imported DTC Operations docs app runnable from dataops, blocked on #2.

[alexeygrigorev]

Software Engineer Report

Status: READY FOR TEST

Changed for #3:

• Added tests/docs_app/test_playwright_local_docs_flow.py, a browser-backed local portal flow that uses temporary repo-shaped content/, builds a temp search index, starts the local Lambda API, serves the real frontend, and verifies search, document open, rendered inline edit/save, reload persistence, and /content/ asset serving without GitHub or AWS credentials.
• Updated tests/docs_app/README.md with the optional Playwright command.
• Updated README.md to point to canonical assistants/podcast/.
• Updated lambda-functions/README.md local search/deploy references for DataOps.
• Updated lambda-functions/template.full.yaml and lambda-functions/template.runtime-secrets.yaml secret-name defaults from inherited dtc-operations/* to dataops-v1/*.

Verification run:

• uv run --project lambda-functions --extra search --with pytest python -m pytest tests/docs_app -q -rs -> 69 passed, 1 skipped. The skip is the optional Playwright test when playwright is not installed.
• uv run --project lambda-functions --extra search --with pytest --with playwright python -m pytest tests/docs_app -q -rs -> 70 passed.
• uv run --project lambda-functions --extra search python -m lambda_functions.build_search_index --docs-dir content --output /tmp/dataops-issue3-search.index -> indexed 357 documents.
• python3 scripts/sop_lint.py content/internal-admin/documentation/sops/creating-process-document.md -> OK: 1 file clean.
• git diff --check -> passed.
• Local Lambda smoke with CONTENT_ROOT=/home/alexey/git/dataops/content and SEARCH_INDEX_PATH=/tmp/dataops-issue3-search.index on port 8897:
  • GET /health -> 200
  • GET /docs -> 200, merged content/ documents returned
  • GET /search?q=invoice&limit=3 -> 200, relevant finance SOP results returned
  • GET /docs?path=content/internal-admin/documentation/sops/creating-process-document.md -> 200 with parsed SOP content
  • GET /lint -> 200 with merged-path lint report

Notes:

• I did not edit docs/local-development.md or docs command-plan files; Add top-level developer command plan #13 owns those. There are unrelated uncommitted edits in docs/README.md and docs/local-development.md in the shared worktree, left untouched.
• No source repos under ../dtc-operations, ../datatasks, or ../podcast-assistant were modified.
• No commit or push performed.

READY FOR TEST