dashboard-ops
Operate and troubleshoot AlphaGPT's Streamlit dashboard, local state inputs, read-only DB/RPC views, Plotly visualizers, and emergency STOP control.
- License 许可
- Apache-2.0
- 最近更新
- 2026-09-03
- Stars
- ★ 351
- Forks
- 27
01 前言 Front Matter
- name
- dashboard-ops
- license
- Apache 2.0
- metadata
- { "disco-role": "operating" }
- description
- Operate and troubleshoot AlphaGPT's Streamlit dashboard, local state inputs, read-only DB/RPC views, Plotly visualizers, and emergency STOP control.
- disable-model-invocation
- true
02 资源文件 3
- 参考文档references/dashboard-reference.md9.4 KB
- 参考文档references/troubleshooting.md7.5 KB
- 脚本scripts/alpha_gpt_dashboard_fixture.py5.0 KB
03 SKILL.md 正文
Dashboard Ops
Use this sub-skill when the task is to launch, inspect, fixture, or troubleshoot
AlphaGPT's Streamlit monitoring dashboard. It covers the dashboard entrypoint,
DashboardService, Plotly visualizers, local state/log files, the market
overview query, auto-refresh behavior, and the emergency STOP button.
Do not route here for
- Trading decisions, live runner lifecycle, portfolio risk policy, or execution safety; use live-strategy.
- Populating Postgres/Timescale tables or debugging Birdeye/DexScreener ingestion; use data-pipeline.
- Training, validating, or rewriting formulas in
best_meme_strategy.json; use factor-mining.
Read or run
- Read references/dashboard-reference.md for launch workflow, UI layout, local file schemas, market query, visualizer inputs, log tail, auto-refresh, empty states, and read-only service prerequisites.
- Read references/troubleshooting.md for Streamlit/import/version errors, empty panels, unknown wallet, DB failures, missing/malformed files, and STOP button caveats.
- Run scripts/alpha_gpt_dashboard_fixture.py to create sample
portfolio_state.json,best_meme_strategy.json, andstrategy.logfiles in a chosen directory without DB, RPC, or network access. - Use the root environment checker with
--scope dashboardfor import/env checks. - Read ../../references/install-and-operations.md when dashboard import problems appear to be dependency or Solana-pin related.
Safe default workflow
- Decide whether the user needs fixture/offline inspection or a live Streamlit server.
- For fixture work, write sample state files into an isolated output directory and do not contact DB/RPC.
- For live dashboard work, confirm the working directory contains or can read
portfolio_state.json,best_meme_strategy.json, andstrategy.logwhen expected. - Confirm DB env vars before expecting the Market Scanner tab to show rows.
- Confirm
QUICKNODE_RPC_URLand wallet/private-key readability only when wallet balance display matters. - Warn that the dashboard STOP button writes a local
STOP_SIGNAL; it only affects a runner watching the same path or configured stop path. - Treat an empty panel as a diagnosable state, not automatically a crash.
Core runtime facts
dashboard/app.pydefines a Streamlit page with wallet status, refresh, emergency stop, metrics, Portfolio/Market Scanner/Logs tabs, and optional auto-refresh.DashboardService.get_market_overview(self, limit=50)queries the latestohlcvtime joined withtokens, ordered by liquidity.DashboardService.load_portfolio()readsportfolio_state.jsonand derivespnl_pctwhen price fields exist.DashboardService.load_strategy_info()displaysbest_meme_strategy.jsonor a not-trained placeholder.DashboardService.get_recent_logs(n=50)tailsstrategy.logif present.plot_pnl_distribution(portfolio_df)expectssymbolandpnl_pct;plot_market_scatter(market_df)expectsliquidity,volume,fdv, andsymbol.
Quick route map
| User intent | Read/run | Notes |
|---|---|---|
| "Create dashboard test files." | scripts/alpha_gpt_dashboard_fixture.py | No DB/RPC/network; use --output-dir. |
| "Why is the portfolio empty?" | references/troubleshooting.md | Missing file, empty JSON, or no positions can all be valid. |
| "Why is Market Scanner blank?" | references/dashboard-reference.md and references/troubleshooting.md | Check DB join/time assumptions and ingestion. |
| "What does EMERGENCY STOP do?" | references/dashboard-reference.md | It writes a local file; it is not a chain transaction canceler. |
| "Can I run Streamlit now?" | references/dashboard-reference.md | Confirm env, working directory, and whether live DB/RPC is authorized. |
Verification stance
The native dashboard is a help-only or tiny-fixture candidate. Prefer static checks and the bundled fixture generator for verification. Do not require a live Streamlit server, wallet, DB, RPC endpoint, or network connection unless the user explicitly authorizes live operational checks and supplies the required services.
04 同仓库更多 Skills
passnet-triton-opt
Make a single PassNet Triton kernel fast AND numerically faithful: performance model (when a replacement can win at all), block/grid/warp tuning, autotune policy, launch overhead, and per-op numeric recipes to pass the dtype baseline tolerances. Use when a pass matches and is correct but speedup below expected, or when correctness fails by small numeric margins.
passnet-solve
END-TO-END playbook for solving one PassNet sample: analyze the computation graphs, decide the optimization strategy (what to fuse, what to replace, what to leave alone), drive the iteration loop, and maximize the sample score. This is the ENTRY skill — invoke it first for any PassNet optimization task; it tells you when to use passnet-pattern-fusion, passnet-triton-opt and passnet-feedback.
passnet-skill
PassNet GPU kernel optimization via compiler passes. Design and implement Triton-based optimization passes, create pass files under ./pass_dir/, self-evaluate with pass_evaluator, and iterate to maximize GPU speedup.
passnet-pattern-fusion
Author PassNet pass files: write patterns that actually MATCH the FX graph, pick fusion regions (which ops to absorb into one Triton kernel), and structure multi-pass submissions with the mandatory shared-dispatch architecture. Use when creating or fixing pattern()/replacement_args()/replacement_func() files, when a pass "failed to match", or when deciding how to fuse multiple kernels/ops.
passnet-orchestrate
Round/batch planning and decision support for PassNet graph optimization. Use for multi-sample triage, worker allocation, eval-budget policy, and post-eval keep/revert decisions. For solving one concrete sample end to end, use passnet-solve as the entry skill; consult this skill only when you need broader planning or a specific decision gate.
passnet-feedback
Fast, deterministic feedback for PassNet work: pre-flight pattern/match verification WITHOUT burning a GPU evaluation, per-node bottleneck analysis of a sample's graphs, and eval-log parsing into per-variant status + estimated score + failure classification. Use BEFORE every GPU evaluation (check_pattern), at the START of a sample (analyze_graph), and AFTER every evaluation (parse_eval_log).