@xmarts/genius-setup 1.2.0 → 1.3.0

package/hooks/genius_capture.py

@@ -18,6 +18,7 @@ Hard rules:
 Input: Claude Code passes a JSON event on stdin including `transcript_path` and
 `session_id`. We tolerate missing fields.
 """
+import datetime as _dt
 import json
 import os
 import re
@@ -63,6 +64,75 @@ def _read_transcript(path):
         return ""
 
 
+_SESSION_GAP_SECONDS = 30 * 60     # idle gap that splits work into blocks
+_METRIC_MAX_LINES = 40000
+
+
+def _parse_ts(s):
+    try:
+        return _dt.datetime.fromisoformat(str(s).replace("Z", "+00:00"))
+    except Exception:
+        return None
+
+
+def _session_metrics(path):
+    """Derive honest session timing from the transcript JSONL: gap-clustered
+    ACTIVE minutes (idle gaps removed), start/end, and tool-event count. This is
+    the human-work signal the server turns into hours — computed locally so no
+    transcript content is needed to know the TIME spent."""
+    if not path or not os.path.isfile(path):
+        return None
+    times, tool_events = [], 0
+    try:
+        with open(path, "r", encoding="utf-8", errors="replace") as fh:
+            for i, line in enumerate(fh):
+                if i >= _METRIC_MAX_LINES:
+                    break
+                line = line.strip()
+                if not line:
+                    continue
+                try:
+                    obj = json.loads(line)
+                except Exception:
+                    continue
+                t = _parse_ts(obj.get("timestamp")) if obj.get("timestamp") else None
+                if t:
+                    times.append(t)
+                msg = obj.get("message") or {}
+                content = msg.get("content") if isinstance(msg, dict) else None
+                if isinstance(content, list):
+                    for blk in content:
+                        if isinstance(blk, dict) and blk.get("type") == "tool_use":
+                            tool_events += 1
+    except Exception:
+        return None
+    if not times:
+        return None
+    times.sort()
+    active = 0.0
+    start = prev = times[0]
+    blocks = []
+    for t in times[1:]:
+        if (t - prev).total_seconds() > _SESSION_GAP_SECONDS:
+            blocks.append((start, prev))
+            start = t
+        prev = t
+    blocks.append((start, prev))
+    for a, b in blocks:
+        active += max((b - a).total_seconds(), 60.0)   # >=1 min per block
+
+    def _fmt(d):
+        if d.tzinfo:
+            d = d.astimezone(_dt.timezone.utc).replace(tzinfo=None)
+        return d.strftime("%Y-%m-%d %H:%M:%S")
+    return {
+        "started_at": _fmt(times[0]),
+        "ended_at": _fmt(times[-1]),
+        "active_minutes": round(active / 60.0, 2),
+        "tool_events": tool_events,
+    }
+
+
 _CRED_PATTERNS = [
     re.compile(r'-----BEGIN[A-Z ]*PRIVATE KEY-----.*?-----END[A-Z ]*PRIVATE KEY-----', re.DOTALL),
     re.compile(r'(?i)\bBearer\s+[A-Za-z0-9._\-]{20,}'),
@@ -95,15 +165,16 @@ def _spool_files():
     return sorted(f for f in os.listdir(OUTBOX_DIR) if f.endswith(".json"))
 
 
-def _spool(payload):
-    """Persist a payload. (#pkg-2) One file per session_ref — repeated Stop
+def _spool(payload, prefix="sess"):
+    """Persist a payload. (#pkg-2) One file per (prefix, session_ref) — repeated
     firings of the SAME session OVERWRITE one file instead of creating N — and a
     true on-disk RING BUFFER cap, dropping the oldest when at MAX_OUTBOX_FILES,
-    so the directory can never grow without bound."""
+    so the directory can never grow without bound. The prefix keeps the transcript
+    payload ('sess') and the session-metadata payload ('meta') as distinct files."""
     os.makedirs(OUTBOX_DIR, exist_ok=True)
     sref = (payload.get("session_ref") or "s")[:40].replace("/", "_").replace(".", "_")
-    # session dedup: a stable name per session -> overwrite, not append
-    name = "sess_%s.json" % sref if sref != "s" else "%d_anon.json" % int(time.time() * 1000)
+    # session dedup: a stable name per (prefix, session) -> overwrite, not append
+    name = "%s_%s.json" % (prefix, sref) if sref != "s" else "%d_anon.json" % int(time.time() * 1000)
     path = os.path.join(OUTBOX_DIR, name)
     if not os.path.exists(path):
         existing = _spool_files()
@@ -122,6 +193,7 @@ def _spool(payload):
 
 
 def _post(endpoint, token, payload, timeout=15):
+    payload = {k: v for k, v in payload.items() if k != "_endpoint"}
     body = json.dumps({"jsonrpc": "2.0", "method": "call",
                        "params": payload, "id": 1}).encode()
     req = urllib.request.Request(endpoint, data=body)
@@ -154,7 +226,7 @@ def _flush_outbox(endpoint, token):
         try:
             with open(path) as fh:
                 payload = json.load(fh)
-            _post(endpoint, token, payload)
+            _post(payload.get("_endpoint") or endpoint, token, payload)
             os.remove(path)  # delivered
         except PermanentError:
             _deadletter(path)            # bad payload -> set aside, keep going
@@ -186,20 +258,58 @@ def main():
     if not endpoint or not token:
         return  # not configured -> silent no-op
 
+    # Derive the metadata-only session endpoint (no content) from the configured
+    # capture endpoint or base_url.
+    session_endpoint = cfg.get("session_endpoint")
+    if not session_endpoint:
+        if "/capture/transcript" in endpoint:
+            session_endpoint = endpoint.replace("/capture/transcript", "/capture/session")
+        elif cfg.get("base_url"):
+            session_endpoint = cfg["base_url"].rstrip("/") + "/xma/genius/v1/capture/session"
+
     event = _read_stdin_event()
-    transcript = _redact(_read_transcript(event.get("transcript_path")))
+    tpath = event.get("transcript_path")
+    transcript = _redact(_read_transcript(tpath))
     session_ref = event.get("session_id") or event.get("session_ref") or ""
 
-    if transcript.strip():
+    # Off-platform / sensitive work keeps CONTENT out of the Brain entirely — only
+    # the TIME (session metadata) is sent. Inferred from an explicit pii_safe flag
+    # or a topic set with no Genius client (e.g. Ateneai insurance/medical work).
+    offplatform = bool(cfg.get("pii_safe")) or (
+        bool(cfg.get("topic")) and not cfg.get("target_client"))
+
+    if transcript.strip() and not offplatform:
         payload = {
             "transcript": transcript,
             "session_ref": session_ref,
             "source": "session_end_hook",
+            "_endpoint": endpoint,
         }
         if cfg.get("target_client"):
             payload["target_client"] = cfg["target_client"]
         # persist first (durability), then try to deliver
-        _spool(payload)
+        _spool(payload, prefix="sess")
+
+    # Claude-session TIME metadata (no content) -> /capture/session. This is the
+    # honest human-work signal; it also covers off-platform sessions (no client).
+    if session_endpoint and session_ref:
+        metrics = _session_metrics(tpath)
+        if metrics and metrics.get("active_minutes", 0) > 0:
+            meta = {
+                "session_ref": session_ref,
+                "source": "transcript_hook",
+                "is_offplatform": offplatform,
+                "_endpoint": session_endpoint,
+            }
+            meta.update(metrics)
+            if cfg.get("target_client"):
+                meta["target_client"] = cfg["target_client"]
+            ws = cfg.get("workspace") or event.get("cwd")
+            if ws:
+                meta["workspace"] = ws
+            if cfg.get("topic"):
+                meta["topic"] = cfg["topic"]
+            _spool(meta, prefix="meta")
 
     # Always attempt to flush the spool (delivers this + any backlog)
     try:

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@xmarts/genius-setup",
-  "version": "1.2.0",
-  "description": "One-command frictionless onboarding for Xmarts Genius (Brain MCP + per-turn injection + automatic session capture).",
+  "version": "1.3.0",
+  "description": "One-command frictionless onboarding for Xmarts Genius (Brain MCP + per-turn injection + automatic session capture + Claude-session time telemetry).",
   "bin": {
     "genius-setup": "bin/genius-setup.js"
   },