nexus-agent-cli 0.5.0__tar.gz → 0.7.0__tar.gz

{nexus_agent_cli-0.5.0 → nexus_agent_cli-0.7.0}/PKG-INFO

@@ -1,7 +1,7 @@
-Metadata-Version: 2.4
+Metadata-Version: 2.5
 Name: nexus-agent-cli
-Version: 0.5.0
-Summary: Nexus internal agent ops CLI — pull run params, push analysis results, send notifications. Admin only.
+Version: 0.7.0
+Summary: Nexus internal agent ops CLI — pull run params, push analysis results, drive the Discord KB / blog-draft data plane, send notifications. Admin only.
 Requires-Python: >=3.10
 Requires-Dist: httpx>=0.27
 Requires-Dist: platformdirs>=4.2
@@ -85,6 +85,42 @@ nexus-agent push-config --ticker tqqq --time-range 2026-06-01_2026-06-14 \
 nexus-agent notify --to a@x.com --to b@y.com --subject "Weekly report" --body-file summary.txt
 ```
 
+### `discord` —— Discord KB / blog draft 数据平面
+
+```bash
+# 1) 起草前必调：取当前生效的风格 playbook
+#    ⚠️ 没有 active playbook 时返回 409 并退出 1 —— 必须中止本次运行，
+#       不允许降级为无风格约束照写
+nexus-agent discord playbook-active --out playbook.json --body-out playbook.md
+
+# 2) 取材（窗口由服务端水位线定义，**没有 --date**）
+nexus-agent discord clusters-pending --out clusters.json
+
+# 3) 回写富化结果（只改 title/summary/tickers/topics，不碰 blog_state）
+nexus-agent discord push-enrichment --file enrichment.json
+
+# 4) 回传草稿（markdown 正文 + 结构化 meta）
+#    meta.json: {title, draft_date, source_cluster_ids: [...非空], front_matter?, critic_report?}
+nexus-agent discord push-blog-draft --file draft.md --meta meta.json
+
+# 5) 反馈折叠：够 10 条可折叠反馈才跑 playbook_reviser
+nexus-agent discord feedback-pending --out feedback.json
+
+# 6) 提交 playbook。meta.json: {rules: [...], change_log, derived_from?, folded_feedback_ids?}
+#    ⚠️ 不要在 meta 里放 token_est —— 服务端用 count_tokens 实算（客户端会直接拒绝）
+nexus-agent discord push-playbook --file playbook.md --meta meta.json --activate   # v1 才用 --activate
+nexus-agent discord push-playbook --file playbook.md --meta meta.json             # 修订版：先落库待审批
+nexus-agent discord activate-playbook --version 2                                  # 相当于在 #blog-review 点 ✅
+```
+
+⚠️ **`clusters-pending` 刻意没有 `--date`。** 取材窗口是「上次成功运行覆盖到的时刻 →
+现在」，不是自然日 —— 按自然日切会出现某天的笔记谁也没收或被收两次。响应里的
+`window` 字段会告诉你这次实际覆盖了哪一段。
+
+⚠️ **`push-blog-draft` 目前只落库。** 把草稿发到 `#blog-review` 并挂审批 reaction
+尚未接线，响应里的 `discord_delivery` 会是 `not_wired`。所以 `activate-playbook`
+是目前唯一的"✅"入口。
+
 ## 配置（环境变量，均可选）
 
 | 变量 | 默认 | 说明 |
@@ -107,15 +143,43 @@ NEXUS_AGENT_API_BASE_URL=http://127.0.0.1:8000/api nexus-agent run-config
 `run-config/`、`upload/market-analysis/`、`upload/portfolio-recommendation/`、
 `upload/backtest/`、`notify/`。
 
+`discord` 子命令组调用 `apps/discord_kb` 的 7 个 admin 端点
+（`clusters/pending/`、`clusters/enrich/`、`blog-drafts/`、`style-playbook/active/`、
+`style-playbook/`、`style-playbook/<version>/activate/`、`feedback/pending/`）。
+设计与不变量见 nexus-service `docs/DISCORD_KB_AND_BOT_DESIGN.md`（§A.6 是这一组的来源）。
+
+## 测试
+
+```bash
+python -m unittest discover -s tests
+```
+
+只用标准库（`unittest` + `http.server`），不需要额外依赖、不出网、不读凭据目录 ——
+`tests/test_api_client_shapes.py` 把请求打到 127.0.0.1 的捕获服务器上，验路径与
+JSON body 的形状（这类错误在真实调用里只会表现为服务端 400/422，那时已分不清是
+本地拼错还是服务端变了）。
+
 ## 发布新版本
 
 ```bash
-# 1. 改 pyproject.toml 的 version
+# 1. 版本号有**两个**事实源，必须一起改：
+#      pyproject.toml 的 version   → 决定包版本
+#      nexus_agent_cli/__init__.py 的 __version__ → 决定 `nexus-agent version` 打印什么
+#    只改前者的话，装上新版后 `nexus-agent version` 仍打印旧号 —— 不报错、只是骗人。
+#    （2026-08-22 发 0.7.0 时就差点这样发出去。）
 # 2. 构建并发布到 PyPI（用 .zshrc 里的 UV_PUBLISH_PASSWORD all-projects token）
 rm -rf dist && python -m build
+python -m twine check dist/*          # 先查，再传
 TWINE_USERNAME=__token__ TWINE_PASSWORD="$UV_PUBLISH_PASSWORD" twine upload dist/*
 ```
 
+发布前建议核一遍（同一版本号 PyPI 不允许重传，错了只能跳版本）：
+
+```bash
+grep '^version' pyproject.toml; grep __version__ nexus_agent_cli/__init__.py   # 两处一致
+python -c "import zipfile,glob; print(*sorted(zipfile.ZipFile(glob.glob('dist/*.whl')[0]).namelist()),sep=chr(10))"
+```
+
 ## 迁移说明
 
 当前 AWS（EventBridge → ECS Fargate）的每周定时运行**保持不变**。本 CLI 提供并行的

{nexus_agent_cli-0.5.0 → nexus_agent_cli-0.7.0}/README.md

@@ -74,6 +74,42 @@ nexus-agent push-config --ticker tqqq --time-range 2026-06-01_2026-06-14 \
 nexus-agent notify --to a@x.com --to b@y.com --subject "Weekly report" --body-file summary.txt
 ```
 
+### `discord` —— Discord KB / blog draft 数据平面
+
+```bash
+# 1) 起草前必调：取当前生效的风格 playbook
+#    ⚠️ 没有 active playbook 时返回 409 并退出 1 —— 必须中止本次运行，
+#       不允许降级为无风格约束照写
+nexus-agent discord playbook-active --out playbook.json --body-out playbook.md
+
+# 2) 取材（窗口由服务端水位线定义，**没有 --date**）
+nexus-agent discord clusters-pending --out clusters.json
+
+# 3) 回写富化结果（只改 title/summary/tickers/topics，不碰 blog_state）
+nexus-agent discord push-enrichment --file enrichment.json
+
+# 4) 回传草稿（markdown 正文 + 结构化 meta）
+#    meta.json: {title, draft_date, source_cluster_ids: [...非空], front_matter?, critic_report?}
+nexus-agent discord push-blog-draft --file draft.md --meta meta.json
+
+# 5) 反馈折叠：够 10 条可折叠反馈才跑 playbook_reviser
+nexus-agent discord feedback-pending --out feedback.json
+
+# 6) 提交 playbook。meta.json: {rules: [...], change_log, derived_from?, folded_feedback_ids?}
+#    ⚠️ 不要在 meta 里放 token_est —— 服务端用 count_tokens 实算（客户端会直接拒绝）
+nexus-agent discord push-playbook --file playbook.md --meta meta.json --activate   # v1 才用 --activate
+nexus-agent discord push-playbook --file playbook.md --meta meta.json             # 修订版：先落库待审批
+nexus-agent discord activate-playbook --version 2                                  # 相当于在 #blog-review 点 ✅
+```
+
+⚠️ **`clusters-pending` 刻意没有 `--date`。** 取材窗口是「上次成功运行覆盖到的时刻 →
+现在」，不是自然日 —— 按自然日切会出现某天的笔记谁也没收或被收两次。响应里的
+`window` 字段会告诉你这次实际覆盖了哪一段。
+
+⚠️ **`push-blog-draft` 目前只落库。** 把草稿发到 `#blog-review` 并挂审批 reaction
+尚未接线，响应里的 `discord_delivery` 会是 `not_wired`。所以 `activate-playbook`
+是目前唯一的"✅"入口。
+
 ## 配置（环境变量，均可选）
 
 | 变量 | 默认 | 说明 |
@@ -96,15 +132,43 @@ NEXUS_AGENT_API_BASE_URL=http://127.0.0.1:8000/api nexus-agent run-config
 `run-config/`、`upload/market-analysis/`、`upload/portfolio-recommendation/`、
 `upload/backtest/`、`notify/`。
 
+`discord` 子命令组调用 `apps/discord_kb` 的 7 个 admin 端点
+（`clusters/pending/`、`clusters/enrich/`、`blog-drafts/`、`style-playbook/active/`、
+`style-playbook/`、`style-playbook/<version>/activate/`、`feedback/pending/`）。
+设计与不变量见 nexus-service `docs/DISCORD_KB_AND_BOT_DESIGN.md`（§A.6 是这一组的来源）。
+
+## 测试
+
+```bash
+python -m unittest discover -s tests
+```
+
+只用标准库（`unittest` + `http.server`），不需要额外依赖、不出网、不读凭据目录 ——
+`tests/test_api_client_shapes.py` 把请求打到 127.0.0.1 的捕获服务器上，验路径与
+JSON body 的形状（这类错误在真实调用里只会表现为服务端 400/422，那时已分不清是
+本地拼错还是服务端变了）。
+
 ## 发布新版本
 
 ```bash
-# 1. 改 pyproject.toml 的 version
+# 1. 版本号有**两个**事实源，必须一起改：
+#      pyproject.toml 的 version   → 决定包版本
+#      nexus_agent_cli/__init__.py 的 __version__ → 决定 `nexus-agent version` 打印什么
+#    只改前者的话，装上新版后 `nexus-agent version` 仍打印旧号 —— 不报错、只是骗人。
+#    （2026-08-22 发 0.7.0 时就差点这样发出去。）
 # 2. 构建并发布到 PyPI（用 .zshrc 里的 UV_PUBLISH_PASSWORD all-projects token）
 rm -rf dist && python -m build
+python -m twine check dist/*          # 先查，再传
 TWINE_USERNAME=__token__ TWINE_PASSWORD="$UV_PUBLISH_PASSWORD" twine upload dist/*
 ```
 
+发布前建议核一遍（同一版本号 PyPI 不允许重传，错了只能跳版本）：
+
+```bash
+grep '^version' pyproject.toml; grep __version__ nexus_agent_cli/__init__.py   # 两处一致
+python -c "import zipfile,glob; print(*sorted(zipfile.ZipFile(glob.glob('dist/*.whl')[0]).namelist()),sep=chr(10))"
+```
+
 ## 迁移说明
 
 当前 AWS（EventBridge → ECS Fargate）的每周定时运行**保持不变**。本 CLI 提供并行的

nexus_agent_cli-0.7.0/nexus_agent_cli/__init__.py

@@ -0,0 +1 @@
+__version__ = "0.7.0"

nexus_agent_cli-0.7.0/nexus_agent_cli/api_client.py

@@ -0,0 +1,323 @@
+from __future__ import annotations
+
+import json
+from typing import Any
+
+import httpx
+
+from nexus_agent_cli.auth_pkce import ensure_fresh_id_token
+from nexus_agent_cli.config import api_base_url
+
+
+def _headers() -> dict[str, str]:
+    token = ensure_fresh_id_token()
+    return {"Authorization": f"Bearer {token}", "Accept": "application/json"}
+
+
+def _raise_for_error(response: httpx.Response) -> None:
+    if response.status_code < 400:
+        return
+    try:
+        detail: Any = response.json()
+    except json.JSONDecodeError:
+        detail = response.text
+    raise RuntimeError(f"HTTP {response.status_code}: {detail}")
+
+
+def get_json(path: str, *, params: dict[str, Any] | None = None, timeout: float = 60.0) -> Any:
+    with httpx.Client(timeout=timeout) as client:
+        response = client.get(f"{api_base_url()}{path}", headers=_headers(), params=params)
+    _raise_for_error(response)
+    return response.json()
+
+
+def post_json(path: str, body: Any, *, timeout: float = 120.0) -> Any:
+    headers = {**_headers(), "Content-Type": "application/json"}
+    with httpx.Client(timeout=timeout) as client:
+        response = client.post(f"{api_base_url()}{path}", headers=headers, json=body)
+    _raise_for_error(response)
+    return response.json()
+
+
+# ── Nexus agent endpoints (admin only) ──────────────────────────────────────
+
+
+def get_run_config(tickers: str | None = None) -> Any:
+    """GET /nexus-agent/run-config/ — tickers + date range for this week's run."""
+    params = {"tickers": tickers} if tickers else None
+    return get_json("/nexus-agent/run-config/", params=params)
+
+
+def get_strategy_data(
+    ticker: str, start_date: str | None = None, end_date: str | None = None
+) -> Any:
+    """GET /nexus-agent/strategy-data/ — strategy trigger rows (Sheets creds stay server-side)."""
+    params: dict[str, Any] = {"ticker": ticker}
+    if start_date:
+        params["start_date"] = start_date
+    if end_date:
+        params["end_date"] = end_date
+    return get_json("/nexus-agent/strategy-data/", params=params)
+
+
+def get_backtest(ticker: str) -> Any:
+    """GET /nexus-agent/backtest/ — latest aggregated backtest data for one ticker."""
+    return get_json("/nexus-agent/backtest/", params={"ticker": ticker})
+
+
+def get_market_analysis(ticker: str) -> Any:
+    """GET /nexus-agent/market-analysis/ — latest market analysis report for one ticker."""
+    return get_json("/nexus-agent/market-analysis/", params={"ticker": ticker})
+
+
+def get_portfolio_recommendation() -> Any:
+    """GET /nexus-agent/portfolio-recommendation/ — latest portfolio recommendation."""
+    return get_json("/nexus-agent/portfolio-recommendation/")
+
+
+def upload_market_analysis(
+    ticker: str, content: str, language: str = "zh", metadata: Any = None
+) -> Any:
+    """POST /nexus-agent/upload/market-analysis/
+
+    metadata: 结构化 metadata(emit_metadata.py 产出)。服务端存成 sidecar,
+    读端点原样返回给前端右侧面板。省略时后端不写 sidecar,前端退回 markdown 渲染。
+    """
+    body: dict[str, Any] = {"ticker": ticker, "content": content, "language": language}
+    if metadata is not None:
+        body["metadata"] = metadata
+    return post_json("/nexus-agent/upload/market-analysis/", body)
+
+
+def upload_portfolio_recommendation(
+    content: str,
+    content_cn: str | None = None,
+    time_range: str | None = None,
+    metadata: Any = None,
+) -> Any:
+    """POST /nexus-agent/upload/portfolio-recommendation/"""
+    body: dict[str, Any] = {"content": content}
+    if content_cn is not None:
+        body["content_cn"] = content_cn
+    if time_range is not None:
+        body["time_range"] = time_range
+    if metadata is not None:
+        body["metadata"] = metadata
+    return post_json("/nexus-agent/upload/portfolio-recommendation/", body)
+
+
+def upload_backtest(ticker: str, levels: dict[str, Any]) -> Any:
+    """POST /nexus-agent/upload/backtest/ — levels: {aggressive/conservative/normal: {...}}"""
+    return post_json("/nexus-agent/upload/backtest/", {"ticker": ticker, **levels})
+
+
+def get_position_config(
+    ticker: str, kind: str = "normal", date_range: str | None = None
+) -> Any:
+    """GET /nexus-agent/position-config/ — signal_position config from S3.
+
+    Mirrors the worker's download_latest_normal_config_from_s3: the server reads
+    configs/signal_position/{ticker}/<date_range>/{ticker}-{kind}.json and returns it
+    as JSON. With date_range, fetch that specific historical version; otherwise the
+    latest. This is the Position-Research baseline the local skill cannot otherwise read.
+    """
+    params: dict[str, Any] = {"ticker": ticker, "kind": kind}
+    if date_range:
+        params["date_range"] = date_range
+    return get_json("/nexus-agent/position-config/", params=params)
+
+
+def list_history(
+    artifact_type: str, ticker: str | None = None, kind: str | None = None, since: str | None = None
+) -> Any:
+    """GET /nexus-agent/history/ — list historical versions of an artifact.
+
+    artifact_type: config / market-analysis / portfolio.
+    Returns {type, count, items: [{key, created_at, date_range?}, ...]} newest-first.
+    """
+    params: dict[str, Any] = {"type": artifact_type}
+    if ticker:
+        params["ticker"] = ticker
+    if kind:
+        params["kind"] = kind
+    if since:
+        params["since"] = since
+    return get_json("/nexus-agent/history/", params=params)
+
+
+def upload_position_config(
+    ticker: str, configs: dict[str, Any], time_range: str | None = None
+) -> Any:
+    """POST /nexus-agent/upload/position-config/ — write the 3 position configs to S3.
+
+    configs: {aggressive/conservative/normal: {<level>: {...}}}. Distinct from
+    upload_backtest (which uploads backtest aggregation, not position configs).
+    Server writes configs/signal_position/{ticker}/{time_range}/{ticker}-{kind}.json
+    and invalidates the dashboard cache.
+    """
+    body: dict[str, Any] = {"ticker": ticker, **configs}
+    if time_range is not None:
+        body["time_range"] = time_range
+    return post_json("/nexus-agent/upload/position-config/", body)
+
+
+def upload_position_report(ticker: str, content: str) -> Any:
+    """POST /nexus-agent/upload/position-report/ — archive a position-update narrative (md)."""
+    return post_json(
+        "/nexus-agent/upload/position-report/", {"ticker": ticker, "content": content}
+    )
+
+
+# ── discord_kb endpoints (admin only) ───────────────────────────────────────
+#
+# 这些端点是 Discord KB / blog draft 的**数据平面**：Discord bot token、OpenAI key、
+# DB 凭据全部只在服务端，本地运行看不到任何密钥。
+# 设计见 nexus-service/docs/DISCORD_KB_AND_BOT_DESIGN.md §A.6。
+
+
+def discord_clusters_pending() -> Any:
+    """GET /discord-kb/clusters/pending/ — 取材：窗口内 pending + 所有 deferred 的簇。
+
+    ⚠️ **刻意不接受 date 参数。** 取材窗口由服务端的水位线定义（上次成功运行覆盖到的
+    时刻 → 现在），不是任何自然日边界 —— 见设计文档 §B.6：cron 在 UTC 01:00 跑，而
+    北京时间晚上 11 点记的笔记属于 UTC 前一天下午，按自然日切窗口会出现某天的笔记
+    谁也没收、或者被收两次。加一个 --date 就等于把那个坑重新开出来。
+    响应里的 window 字段会告诉你这次实际覆盖了哪一段。
+    """
+    return get_json("/discord-kb/clusters/pending/")
+
+
+def discord_push_enrichment(clusters: list[dict[str, Any]]) -> Any:
+    """POST /discord-kb/clusters/enrich/ — 回写富化结果（title/summary/tickers/topics）。
+
+    只改描述性字段，**不碰 blog_state** —— 状态迁移只走服务端的 service 层。
+    """
+    return post_json("/discord-kb/clusters/enrich/", {"clusters": clusters})
+
+
+def discord_push_blog_draft(
+    title: str,
+    body_md: str,
+    draft_date: str,
+    source_cluster_ids: list[str],
+    front_matter: Any = None,
+    critic_report: Any = None,
+) -> Any:
+    """POST /discord-kb/blog-drafts/ — 提交草稿。
+
+    服务端落库并把来源簇标 `drafting`（同一事务）。`source_cluster_ids` 必须非空：
+    草稿要可追溯到来源，且标不上 drafting 的簇明天会被当 pending 重新起草。
+    ⚠️ 发 #blog-review 与挂审批 reaction **尚未接线**，响应里的 discord_delivery
+    会是 "not_wired"。
+    """
+    body: dict[str, Any] = {
+        "title": title,
+        "body_md": body_md,
+        "draft_date": draft_date,
+        "source_cluster_ids": source_cluster_ids,
+    }
+    if front_matter is not None:
+        body["front_matter"] = front_matter
+    if critic_report is not None:
+        body["critic_report"] = critic_report
+    return post_json("/discord-kb/blog-drafts/", body)
+
+
+def discord_playbook_active() -> Any:
+    """GET /discord-kb/style-playbook/active/ — 当前生效的风格 playbook。起草前必调。
+
+    ⚠️ 没有 active 时服务端返回 **409**（不是 200 + null）：调用方必须中止本次运行
+    并报警，绝不允许降级为无风格约束照写（§6.5）。所以这里的 RuntimeError
+    不要 catch 掉当成"没配置"。
+    """
+    return get_json("/discord-kb/style-playbook/active/")
+
+
+def discord_feedback_pending() -> Any:
+    """GET /discord-kb/feedback/pending/ — 可折叠的审批反馈。
+
+    过滤条件全在服务端（verdict ∈ {revised, rejected} 且 note 非空 且未折叠 且指向
+    当前 active 版本），本地不要重复实现 —— approved 只进指标，expired 绝不当负信号
+    （§C.1）。响应带 should_revise 供本地判断够不够触发一次 revise。
+    """
+    return get_json("/discord-kb/feedback/pending/")
+
+
+def discord_push_playbook(
+    body_md: str,
+    rules: list[Any],
+    change_log: str = "",
+    derived_from: Any = None,
+    folded_feedback_ids: list[str] | None = None,
+    activate: bool = False,
+) -> Any:
+    """POST /discord-kb/style-playbook/ — 提交一版 playbook。
+
+    ⚠️ **不要传 token_est** —— 服务端用 count_tokens 实算，不接受调用方上报（§C.5）。
+    超 2000 token 或 40 条规则会 422 拒收；体积变大而 change_log 的「删除」段为空
+    同样 422。
+    `activate=True` 只应用于 v1（首次从语料抽取）；修订版应先落库，你在 #blog-review
+    ✅ 之后再调 activate。
+    """
+    body: dict[str, Any] = {"body_md": body_md, "rules": rules, "change_log": change_log}
+    if derived_from is not None:
+        body["derived_from"] = derived_from
+    if folded_feedback_ids:
+        body["folded_feedback_ids"] = folded_feedback_ids
+    if activate:
+        body["activate"] = True
+    return post_json("/discord-kb/style-playbook/", body)
+
+
+def discord_activate_playbook(version: int) -> Any:
+    """POST /discord-kb/style-playbook/{version}/activate/ — 相当于在 #blog-review 点 ✅。
+
+    折叠标记（folded_into）在这一步才写：若提交时就标记而最终没 ✅，那批反馈会被
+    永久消费掉、信号白丢且看不出来（§6.6）。
+    """
+    return post_json(f"/discord-kb/style-playbook/{int(version)}/activate/", {})
+
+
+def notify(
+    to: list[str],
+    subject: str,
+    body_text: str | None = None,
+    body_html: str | None = None,
+    cc: list[str] | None = None,
+) -> Any:
+    """POST /nexus-agent/notify/ — send a notification email via server-side SES."""
+    body: dict[str, Any] = {"to": to, "subject": subject}
+    if cc:
+        body["cc"] = cc
+    if body_text is not None:
+        body["body_text"] = body_text
+    if body_html is not None:
+        body["body_html"] = body_html
+    return post_json("/nexus-agent/notify/", body)
+
+
+def notify_report(
+    kind: str,
+    time_range: str,
+    ticker: str | None = None,
+    to: list[str] | None = None,
+    cc: list[str] | None = None,
+    expires_in: int | None = None,
+    metadata: Any = None,
+) -> Any:
+    """POST /nexus-agent/notify/report/ — templated weekly email with 24h presigned links.
+
+    Presigning happens server-side because the local analysis run holds no S3 credentials.
+    """
+    body: dict[str, Any] = {"kind": kind, "time_range": time_range}
+    if ticker:
+        body["ticker"] = ticker
+    if to:
+        body["to"] = to
+    if cc:
+        body["cc"] = cc
+    if expires_in is not None:
+        body["expires_in"] = expires_in
+    if metadata is not None:
+        body["metadata"] = metadata
+    return post_json("/nexus-agent/notify/report/", body)

{nexus_agent_cli-0.5.0 → nexus_agent_cli-0.7.0}/nexus_agent_cli/commands/__init__.py

@@ -3,9 +3,11 @@ from __future__ import annotations
 import typer
 
 from nexus_agent_cli.commands.auth.cmd import register as register_auth
+from nexus_agent_cli.commands.discord.cmd import register as register_discord
 from nexus_agent_cli.commands.run.cmd import register as register_run
 
 
 def register_all(app: typer.Typer) -> None:
     register_auth(app)
     register_run(app)
+    register_discord(app)