codebuff-agent 2.1.74__py3-none-any.whl

jupyter_ext/handlers/analytics_handler.py

@@ -0,0 +1,170 @@
+"""Analytics events handler - append JSONL to file."""
+
+from __future__ import annotations
+
+import asyncio
+import datetime
+import json
+import logging
+import os
+from pathlib import Path
+from typing import Optional
+
+from jupyter_server.base.handlers import APIHandler
+
+from .._version import __version__ as _EXT_VERSION
+from ..analytics_ingest import enqueue_ux_event
+from ..log_paths import get_analytics_log_path
+from ..resource_usage import get_integrated_resources
+from ..team_access import resolve_team
+
+logger = logging.getLogger(__name__)
+
+VALID_EVENT_TYPES = {"impression", "click", "error"}
+
+# Environment-variable sourced identity fields.
+# If any is missing, the handler is disabled (silent 204 for every request).
+_ENV_TEAM = resolve_team()
+_ENV_PROJECT = os.environ.get("HDSP_PRJ_ID", "")
+_ENV_USER = os.environ.get("JUPYTERHUB_USER", "") or os.environ.get("HDSP_BDP_ID", "")
+print(
+    f"[Analytics] env check: HDSP_USER_TEAM={_ENV_TEAM!r}, "
+    f"HDSP_PRJ_ID={_ENV_PROJECT!r}, "
+    f"JUPYTERHUB_USER={os.environ.get('JUPYTERHUB_USER', '')!r}, "
+    f"HDSP_BDP_ID={os.environ.get('HDSP_BDP_ID', '')!r}, "
+    f"_ENV_USER={_ENV_USER!r}",
+    flush=True,
+)
+_DISABLED = not ((_ENV_TEAM or _ENV_PROJECT) and _ENV_USER)
+print(f"[Analytics] _DISABLED={_DISABLED}", flush=True)
+if _DISABLED:
+    logger.warning(
+        "Analytics disabled: missing identity (team=%r, HDSP_PRJ_ID=%r, "
+        "JUPYTERHUB_USER/HDSP_BDP_ID=%r)",
+        _ENV_TEAM,
+        _ENV_PROJECT,
+        _ENV_USER,
+    )
+
+
+def _get_analytics_log_path_for_type(event_type: str) -> Optional[Path]:
+    """Return today's log file path for a given event type.
+
+    File pattern: ``analytics_{event_type}_{project}_{user}_{YYYYMMDD}.jsonl``
+    """
+    today = datetime.date.today().isoformat()
+    day_cache = _DAILY_PATH_CACHE.get(today)
+    if day_cache is not None:
+        cached = day_cache.get(event_type)
+        if cached is not None:
+            return cached
+    else:
+        day_cache = {}
+        _DAILY_PATH_CACHE[today] = day_cache
+    path = get_analytics_log_path(event_type, today.replace("-", ""))
+    day_cache[event_type] = path
+    return path
+
+
+_DAILY_PATH_CACHE: dict[str, dict[str, Path]] = {}
+
+
+def _collect_hardware_attrs() -> dict[str, object]:
+    try:
+        res = get_integrated_resources()
+        cpu = res.get("cpu", {})
+        mem = res.get("memory", {})
+        gpus = res.get("gpus", [])
+        return {
+            "cpu_cores": cpu.get("cores"),
+            "memory_total_gb": mem.get("total_gb"),
+            "gpu_name": gpus[0]["name"] if gpus else None,
+            "gpu_memory_total_mb": gpus[0]["memory_total_mb"] if gpus else None,
+        }
+    except Exception:
+        return {
+            "cpu_cores": None,
+            "memory_total_gb": None,
+            "gpu_name": None,
+            "gpu_memory_total_mb": None,
+        }
+
+
+_HW_ATTRS: dict[str, object] = _collect_hardware_attrs()
+
+
+def _enqueue_events(events: list) -> None:
+    """RDS 적재 큐에 넣는다. 워커 스레드에서 돈다 — 빌더가 JSON 직렬화를 하므로
+    이벤트루프 위에서 돌리면 큰 context 에서 응답이 밀린다."""
+    for event in events:
+        enqueue_ux_event(event)
+
+
+def _write_events(path: Path, lines: list[str]) -> None:
+    try:
+        with open(path, "a", encoding="utf-8") as f:
+            for line in lines:
+                f.write(line + "\n")
+    except OSError:
+        pass
+
+
+class AnalyticsEventsHandler(APIHandler):
+    """POST /codebuff/analytics/events"""
+
+    def check_xsrf_cookie(self):
+        pass
+
+    async def post(self):
+        try:
+            if _DISABLED:
+                self.set_status(204)
+                self.finish()
+                return
+
+            body = json.loads(self.request.body.decode("utf-8"))
+            if not isinstance(body, list):
+                self.set_status(400)
+                self.finish()
+                return
+
+            lines_by_type: dict[str, list[str]] = {}
+            accepted: list[dict] = []
+            for event in body:
+                if not isinstance(event, dict):
+                    continue
+                if event.get("event_type") not in VALID_EVENT_TYPES:
+                    continue
+                if not event.get("feature"):
+                    continue
+                # Server-side injection
+                event["datetime"] = datetime.datetime.now().strftime(
+                    "%Y-%m-%d %H:%M:%S"
+                )
+                event["team"] = _ENV_TEAM
+                event["project"] = _ENV_PROJECT
+                event["user_id"] = _ENV_USER
+                event["extension_version"] = _EXT_VERSION
+                event.update(_HW_ATTRS)
+                et = event["event_type"]
+                lines_by_type.setdefault(et, []).append(
+                    json.dumps(event, ensure_ascii=False)
+                )
+                accepted.append(event)
+
+            for et, lines in lines_by_type.items():
+                log_path = _get_analytics_log_path_for_type(et)
+                if log_path is not None:
+                    await asyncio.to_thread(_write_events, log_path, lines)
+
+            # 파일 기록 뒤에 RDS 로 한 벌 더 보낸다 (이슈 #30). enqueue 는 예외를
+            # 내지 않고 URL 이 없으면 no-op 이라 여기서 응답이 바뀌지 않는다.
+            if accepted:
+                await asyncio.to_thread(_enqueue_events, accepted)
+
+            self.set_status(204)
+            self.finish()
+        except Exception:
+            logger.exception("Analytics write failed")
+            self.set_status(204)
+            self.finish()