Add visualization recommender, UI controls and chart-job fallback

[rad1092]

Motivation

• Provide question-driven visualization guidance so the UI can suggest chart types and rationale for common intents (trend, comparison, relationship, quality, etc.).
• Allow users to opt into recommended visualizations and make chart generation robust by returning useful tabular fallbacks when plotting fails.

Description

• Add bitnet_tools/viz_recommender.py with recommend_chart_types(question) that returns an intent, recommended chart types, and a textual reason (includes default overview).
• Extend bitnet_tools/visualize.py to accept an optional selected_chart_types parameter, support line charts, and propagate selection through create_multi_charts().
• Add POST /api/viz/recommend and enhance /api/charts/jobs in bitnet_tools/web.py to accept selected_chart_types or auto-select via the recommender; implement _build_chart_fallback() and make _run_chart_job() return a preview/summary fallback instead of hard failing on chart errors.
• Update UI (bitnet_tools/ui/index.html, bitnet_tools/ui/app.js) with a “시각화 옵션 보기” button, a recommendation panel showing intent/recommended charts/reason, and send the selected/recommended chart types when starting chart jobs.
• Add tests: tests/test_viz_recommender.py for recommendation consistency and an API contract test in tests/test_web.py for /api/viz/recommend.

Testing

• Compiled modules successfully with python -m py_compile bitnet_tools/visualize.py and python -m py_compile bitnet_tools/web.py (succeeded).
• Static check of the UI script with node --check bitnet_tools/ui/app.js (succeeded).
• Ran unit tests with pytest -q tests/test_viz_recommender.py tests/test_web.py and all tests passed (19 passed).

---

Codex Task

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 65d6cefc38

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Recompute chart recommendation for the current question

This reuses appState.vizRecommendation if it exists, so after a user fetches recommendations once and then edits the question, startChartsJob() still sends the old selected_chart_types. Because /api/charts/jobs only auto-recommends when selected_chart_types is empty, the stale list overrides the new question and can generate charts for the wrong intent.

Useful? React with 👍 / 👎.

[chatgpt-codex-connector]

Validate each selected chart type as a string

The endpoint only checks that selected_chart_types is a list, but downstream chart generation normalizes entries with c.strip().lower() in create_file_charts(). If a client sends non-string items (for example [1]), the job is accepted and later fails inside chart generation instead of returning a clear 400 validation error, which makes client bugs look like chart-engine failures.

Useful? React with 👍 / 👎.