feat(token-usage): attribute token usage by agent

[yuanxs21]

Description

Token usage was stored and shown only as (date, provider, model), so Settings → Token Usage could not tell which agent consumed the tokens. This PR stamps the request ContextVar agent id onto each usage event and adds a By Agent table on the Token Usage page.

Recording reads peek_current_agent_id(). Unset context is stored as "". New rows use unit-separator keys, so they do not merge into legacy provider:model rows. Old colon keys are left as-is (no migration).

GET /token-usage/details now returns one row per (date, agent, provider, model) with additive agent_id: null for pre-tracking rows, "" for recorded-but-unattributed. Console re-aggregates details as before; GET /token-usage summary shape is unchanged. The UI maps null and "" to Unattributed, and named ids through getAgentDisplayName.

This PR:

• Adds peek_current_agent_id() and records agent_id on _UsageEvent / disk entries / TokenUsageRecord
• Switches TokenRecordingModelWrapper usage reads to safe_attr(..., "usage") so ChatResponse (dict subclass) missing usage does not KeyError
• Aggregates details into a By Agent table (token-total descending) next to By Model / By Date; i18n for byAgent / agent / unattributed (and labels LLM calls as such)
• Leaves historical colon keys unattributed until they age out of the date window
• Adds unit coverage in tests/unit/token_usage/test_token_usage.py (no merge into legacy rows, model fallback from unit-separator keys, named agent_id round-trip, ContextVar set/empty and no active-agent fallback)

Related Issue: Fixes #(issue_number) or Relates to #(issue_number)

Security Considerations: [If applicable, e.g. channel auth, env/config handling]

Type of Change

• Bug fix
• New feature
• Breaking change
• Documentation
• Refactoring

Component(s) Affected

• Core / Backend (app, agents, config, providers, utils, local_models)
• Console (frontend web UI)
• Channels (DingTalk, Lark, QQ, Discord, iMessage, etc.)
• Skills
• CLI
• Documentation (website)
• Tests
• CI/CD
• Scripts / Deploy

Checklist

• I ran pre-commit run --all-files locally and it passes
• If pre-commit auto-fixed files, I committed those changes and reran checks
• I ran tests locally (pytest or as relevant) and they pass
• Documentation updated (if needed)
• Ready for review

For Channel Changes (DingTalk, Lark, QQ, Console, etc.)

• I ran ./scripts/check-channels.sh (or ./scripts/check-channels.sh --changed) and it passes
• Contract test exists in tests/contract/channels/test_<channel>_contract.py (REQUIRED)
• Contract test implements create_instance() with proper channel initialization
• All 19 contract verification points pass (see tests/contract/channels/__init__.py)
• Optional: Unit tests in tests/unit/channels/test_<channel>.py for complex internal logic

Testing

[How to test these changes]

Evidence

Examples of valid evidence:

• Terminal transcript of the test run (e.g. pytest tests/unit/app/chats/ -q output)
• Screenshot of the Console UI showing the fix
• CI artifact link
• pre-commit run --all-files summary

pre-commit run --all-files
# paste summary result

pytest
# paste summary result

Additional Notes

[Optional: any other context]

[github-actions]

Welcome to QwenPaw! 🐾

Hi @yuanxs21, this is your 34th Pull Request.

🙌 Join Developer Community

Thanks so much for your contribution! We'd love to invite you to join the official QwenPaw developer group! You can find the Discord and DingTalk group links under the "Developer Community" section on our docs page:
https://qwenpaw.agentscope.io/docs/community

We truly appreciate your enthusiasm—and look forward to your future contributions! 😊

We'll review your PR soon.