強化 PPT 產線與線上預覽

services/ppt_auto_generation_service.py

@@ -93,6 +93,51 @@ SCHEDULE_PROFILES = {
     "annual": ("annual",),
 }
 
+SCHEDULE_CADENCES = {
+    "daily": {
+        "label": "每日",
+        "schedule_text": "每日 20:30",
+        "gate": "每日固定產出",
+        "time": "20:30",
+        "description": "每日補齊營運日報，供 22:00 視覺 QA 接續審核。",
+    },
+    "weekly": {
+        "label": "每週",
+        "schedule_text": "每週一 20:40",
+        "gate": "週一固定產出",
+        "time": "20:40",
+        "description": "每週產出週報與市場情報，整理近 7 日變化。",
+    },
+    "monthly": {
+        "label": "每月",
+        "schedule_text": "每月 1 日 20:50",
+        "gate": "每月 1 日產出",
+        "time": "20:50",
+        "description": "每月產出月報、策略、競品、促銷、品類與管理型簡報。",
+    },
+    "quarterly": {
+        "label": "每季",
+        "schedule_text": "每季首日 21:00",
+        "gate": "1/4/7/10 月 1 日產出",
+        "time": "21:00",
+        "description": "每季首日產出季報，承接季度營運檢討。",
+    },
+    "half_yearly": {
+        "label": "每半年",
+        "schedule_text": "每半年首日 21:10",
+        "gate": "1/7 月 1 日產出",
+        "time": "21:10",
+        "description": "每半年首日產出半年報，彙整 H1 / H2 成果。",
+    },
+    "annual": {
+        "label": "每年",
+        "schedule_text": "每年 1/1 21:20",
+        "gate": "每年 1 月 1 日產出",
+        "time": "21:20",
+        "description": "每年首日產出年度總結，保留完整年度資料快照。",
+    },
+}
+
 _RUN_LOCK = threading.Lock()
 _LAST_RUN: dict | None = None
 
@@ -224,6 +269,49 @@ def get_report_type_options() -> list[dict]:
     ] + [{"key": "all", "label": "全部", "prefix": "all"}]
 
 
+def get_schedule_cadence_status(coverage_items: Sequence[dict] | None = None) -> list[dict]:
+    """Return the scheduler contract with optional month coverage counts."""
+    item_by_type = {
+        str(item.get("key")): item
+        for item in (coverage_items or [])
+        if item.get("key")
+    }
+    cadences: list[dict] = []
+    for key, meta in SCHEDULE_CADENCES.items():
+        report_types = SCHEDULE_PROFILES.get(key, ())
+        related_items = [item_by_type[report_type] for report_type in report_types if report_type in item_by_type]
+        ready_count = sum(1 for item in related_items if item.get("ready"))
+        missing_types = [
+            report_type
+            for report_type in report_types
+            if not item_by_type.get(report_type, {}).get("ready")
+        ]
+        total = len(report_types)
+        if total and not missing_types:
+            status = "ready"
+        elif ready_count > 0:
+            status = "partial"
+        else:
+            status = "missing"
+        cadences.append({
+            "key": key,
+            "label": meta["label"],
+            "schedule_text": meta["schedule_text"],
+            "gate": meta["gate"],
+            "time": meta["time"],
+            "description": meta["description"],
+            "report_types": list(report_types),
+            "report_labels": [REPORT_TYPE_LABELS.get(report_type, report_type) for report_type in report_types],
+            "ready_count": ready_count,
+            "missing_count": len(missing_types),
+            "missing_report_types": missing_types,
+            "total": total,
+            "progress_pct": round((ready_count / total * 100), 1) if total else 0,
+            "status": status,
+        })
+    return cadences
+
+
 def build_defined_ppt_jobs(
     *,
     latest_date: str | None = None,
@@ -513,9 +601,12 @@ def get_defined_report_coverage(
         for job in jobs
     ]
     missing = [item for item in items if not item["ready"]]
+    cadences = get_schedule_cadence_status(items)
     return {
         "enabled": is_ppt_auto_generation_enabled(),
         "items": items,
+        "cadences": cadences,
+        "cadence_summary": "、".join(cadence["schedule_text"] for cadence in cadences),
         "missing_report_types": [item["key"] for item in missing],
         "missing_count": len(missing),
         "ready_count": len(items) - len(missing),
@@ -524,6 +615,78 @@ def get_defined_report_coverage(
     }
 
 
+def get_generation_run_history(
+    *,
+    month_start: datetime | None = None,
+    month_end: datetime | None = None,
+    limit: int = 24,
+) -> list[dict]:
+    """Read persisted PPT generation runs without blocking the observability page."""
+    safe_limit = max(1, min(int(limit or 24), 100))
+    where_clauses = []
+    params: dict = {"limit": safe_limit}
+    if month_start is not None:
+        where_clauses.append("started_at >= :month_start")
+        params["month_start"] = month_start
+    if month_end is not None:
+        where_clauses.append("started_at < :month_end")
+        params["month_end"] = month_end
+    where_sql = f"WHERE {' AND '.join(where_clauses)}" if where_clauses else ""
+
+    try:
+        session = get_session()
+        try:
+            rows = session.execute(
+                sa_text(
+                    f"""
+                    SELECT schedule_kind, report_type, target_label, status,
+                           file_path, file_size, error_msg, started_at, finished_at
+                    FROM ppt_generation_runs
+                    {where_sql}
+                    ORDER BY started_at DESC NULLS LAST
+                    LIMIT :limit
+                    """
+                ),
+                params,
+            ).fetchall()
+        finally:
+            session.close()
+    except Exception:
+        return []
+
+    def _format_dt(value) -> str:
+        if hasattr(value, "strftime"):
+            return value.strftime("%Y-%m-%d %H:%M")
+        return str(value or "")
+
+    items = []
+    for row in rows:
+        schedule_kind = row[0] or "manual"
+        report_type = row[1] or ""
+        file_path = row[4] or ""
+        status = row[3] or ""
+        items.append({
+            "schedule_kind": schedule_kind,
+            "schedule_label": SCHEDULE_CADENCES.get(schedule_kind, {}).get("label", "手動"),
+            "report_type": report_type,
+            "report_label": REPORT_TYPE_LABELS.get(report_type, report_type or "未知"),
+            "target_label": row[2] or "",
+            "status": status,
+            "status_label": {
+                "ready": "已產生",
+                "missing_file": "未落盤",
+                "error": "失敗",
+                "planned": "已規劃",
+            }.get(status, status or "未知"),
+            "file_name": os.path.basename(file_path) if file_path else "",
+            "file_size_kb": round(float(row[5] or 0) / 1024, 1) if row[5] else None,
+            "error_msg": row[6] or "",
+            "started_at": _format_dt(row[7]),
+            "finished_at": _format_dt(row[8]),
+        })
+    return items
+
+
 def _generate_job(job: PPTAutoJob) -> str | None:
     from routes import openclaw_bot_routes as bot_routes
 

services/ppt_preview_service.py

@@ -0,0 +1,103 @@
+"""PPT online preview helpers.
+
+Browsers cannot reliably render .pptx files inline, so the app converts a
+deck to a cached PDF and embeds that PDF in the observability preview page.
+"""
+
+from __future__ import annotations
+
+import hashlib
+import os
+import shutil
+import subprocess
+import tempfile
+from dataclasses import dataclass
+from pathlib import Path
+
+
+@dataclass(frozen=True)
+class PPTPreviewResult:
+    ok: bool
+    pdf_path: str | None = None
+    cache_hit: bool = False
+    converter: str | None = None
+    error: str | None = None
+
+
+def find_libreoffice_binary() -> str | None:
+    return shutil.which("libreoffice") or shutil.which("soffice")
+
+
+def _preview_cache_path(pptx_path: Path, cache_dir: Path) -> Path:
+    stat = pptx_path.stat()
+    cache_key = hashlib.sha256(
+        f"{pptx_path.resolve()}:{stat.st_size}:{int(stat.st_mtime)}".encode("utf-8")
+    ).hexdigest()[:16]
+    safe_stem = "".join(ch if ch.isalnum() or ch in ("-", "_") else "_" for ch in pptx_path.stem)[:96]
+    return cache_dir / f"{safe_stem}_{cache_key}.pdf"
+
+
+def build_ppt_preview(
+    pptx_path: str | os.PathLike[str],
+    *,
+    cache_dir: str | os.PathLike[str] | None = None,
+    timeout_sec: int = 90,
+) -> PPTPreviewResult:
+    source = Path(pptx_path)
+    if not source.is_file():
+        return PPTPreviewResult(ok=False, error="pptx not found")
+    if source.suffix.lower() != ".pptx":
+        return PPTPreviewResult(ok=False, error="unsupported file type")
+
+    converter = find_libreoffice_binary()
+    if not converter:
+        return PPTPreviewResult(
+            ok=False,
+            error="LibreOffice is not installed in the app container.",
+        )
+
+    target_dir = Path(cache_dir or os.getenv("PPT_PREVIEW_CACHE_DIR", "/app/data/ppt_previews"))
+    target_dir.mkdir(parents=True, exist_ok=True)
+    target_pdf = _preview_cache_path(source, target_dir)
+    if target_pdf.is_file() and target_pdf.stat().st_size > 0:
+        return PPTPreviewResult(ok=True, pdf_path=str(target_pdf), cache_hit=True, converter=converter)
+
+    with tempfile.TemporaryDirectory(prefix="ppt_preview_") as tmp:
+        tmpdir = Path(tmp)
+        try:
+            proc = subprocess.run(
+                [
+                    converter,
+                    "--headless",
+                    "--convert-to",
+                    "pdf",
+                    "--outdir",
+                    str(tmpdir),
+                    str(source),
+                ],
+                capture_output=True,
+                timeout=timeout_sec,
+                check=False,
+            )
+        except subprocess.TimeoutExpired:
+            return PPTPreviewResult(ok=False, converter=converter, error="LibreOffice conversion timed out.")
+        except Exception as exc:
+            return PPTPreviewResult(ok=False, converter=converter, error=f"{type(exc).__name__}: {str(exc)[:200]}")
+
+        if proc.returncode != 0:
+            stderr = proc.stderr.decode("utf-8", errors="ignore").strip()
+            return PPTPreviewResult(
+                ok=False,
+                converter=converter,
+                error=f"LibreOffice conversion failed: {stderr[:240] or proc.returncode}",
+            )
+
+        generated = tmpdir / f"{source.stem}.pdf"
+        if not generated.is_file():
+            candidates = sorted(tmpdir.glob("*.pdf"))
+            generated = candidates[0] if candidates else generated
+        if not generated.is_file() or generated.stat().st_size <= 0:
+            return PPTPreviewResult(ok=False, converter=converter, error="LibreOffice did not produce a PDF.")
+
+        shutil.move(str(generated), str(target_pdf))
+        return PPTPreviewResult(ok=True, pdf_path=str(target_pdf), cache_hit=False, converter=converter)

services/ppt_vision_service.py

@@ -21,6 +21,7 @@ import os
 import time
 import base64
 import logging
+import shutil
 from dataclasses import dataclass, field
 from typing import Optional, Dict, Any, List
 
@@ -38,6 +39,27 @@ def is_ppt_vision_enabled() -> bool:
     return os.getenv('PPT_VISION_ENABLED', 'false').strip().lower() in ('true', '1', 'yes', 'on')
 
 
+def get_ppt_vision_runtime_status() -> Dict[str, Any]:
+    """Expose why the PPT vision pipeline is or is not ready."""
+    env_value = os.getenv('PPT_VISION_ENABLED')
+    enabled = is_ppt_vision_enabled()
+    converter = shutil.which('libreoffice') or shutil.which('soffice')
+    blockers = []
+    if not enabled:
+        blockers.append('PPT_VISION_ENABLED 未設定為 true')
+    if not converter:
+        blockers.append('容器內缺少 LibreOffice，無法轉換 PPT 做視覺審核')
+    return {
+        'enabled': enabled,
+        'env_value': env_value if env_value is not None else '未設定（預設 false）',
+        'model': os.getenv('PPT_VISION_MODEL', PPT_VISION_MODEL),
+        'converter': converter,
+        'converter_ready': bool(converter),
+        'blockers': blockers,
+        'ready': enabled and bool(converter),
+    }
+
+
 # ─────────────────────────────────────────────────────────────────────────────
 # 結果容器
 # ─────────────────────────────────────────────────────────────────────────────
@@ -115,11 +137,16 @@ class PPTVisionService:
             result['error'] = f'pptx not found: {pptx_path}'
             return result
 
+        converter = shutil.which('libreoffice') or shutil.which('soffice')
+        if not converter:
+            result['error'] = 'libreoffice not installed (skip vision check)'
+            return result
+
         # 1. LibreOffice 轉 png
         with tempfile.TemporaryDirectory() as tmpdir:
             try:
                 proc = subprocess.run(
-                    ['libreoffice', '--headless', '--convert-to', 'png',
+                    [converter, '--headless', '--convert-to', 'png',
                      '--outdir', tmpdir, pptx_path],
                     capture_output=True, timeout=60,
                 )
@@ -440,6 +467,7 @@ __all__ = [
     'VisionResult',
     'ppt_vision_service',
     'is_ppt_vision_enabled',
+    'get_ppt_vision_runtime_status',
     'PPT_VISION_SYSTEM_PROMPT',
     'audit_recent_ppts',
     'push_ppt_audit_to_telegram',