agentx-python 0.6.19__tar.gz → 0.6.21__tar.gz

{agentx_python-0.6.19 → agentx_python-0.6.21}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: agentx-python
-Version: 0.6.19
+Version: 0.6.21
 Summary: Official Python SDK for AgentX (https://www.agentx.so/)
 Home-page: https://github.com/AgentX-ai/AgentX-python
 Author: Robin Wang and AgentX Team

{agentx_python-0.6.19 → agentx_python-0.6.21}/agentx/integrations/__init__.py

@@ -7,3 +7,4 @@
 #   from agentx.integrations.anthropic import patch_anthropic_client
 #   from agentx.integrations.google_adk import AgentXADKPlugin
 #   from agentx.integrations.google_genai import patch_genai_client
+#   from agentx.integrations.moveworks import MoveworksImporter  # Data API pull sync, not in-process

agentx_python-0.6.21/agentx/integrations/moveworks.py

@@ -0,0 +1,525 @@
+"""
+Moveworks integration for AgentX production tracing.
+
+Moveworks agents run inside the Moveworks cloud (Agent Studio + Reasoning Engine), so unlike the
+LangChain/CrewAI integrations there is no in-process callback to hook: Moveworks Python Script
+Actions have outbound network access disabled at the infrastructure level. The sanctioned export
+surface is the read-only Moveworks **Data API** (OData, ``https://api.moveworks.ai/export/v1``),
+which exposes ``conversations``, ``interactions``, ``plugin-calls``, ``plugin-resources`` and
+``users`` tables. This module pulls those records and replays them into AgentX as traces:
+
+- conversation      -> AgentX session (``session_id = mw_<conversation_id>``)
+- interaction       -> one AgentX trace (input/output, timestamps, latency where derivable)
+- plugin call       -> ``tool_calls`` entries on its interaction's trace
+
+Usage::
+
+    from agentx.integrations.moveworks import MoveworksDataAPIClient, MoveworksImporter
+
+    client = MoveworksDataAPIClient(api_key="mw-data-api-key")
+    importer = MoveworksImporter(
+        client,
+        agentx_api_key="agtx_local_...",
+        agentx_base_url="http://localhost:4700/api/v1",   # self-host engine
+    )
+    report = importer.sync(since=datetime.now(timezone.utc) - timedelta(days=1))
+    print(report)
+
+Or from a shell / cron job (keeps an incremental cursor between runs)::
+
+    export MOVEWORKS_API_KEY=... AGENTX_API_KEY=... AGENTX_API_BASE_URL=http://localhost:4700/api/v1
+    agentx-moveworks sync --since 7d
+
+Notes and limits:
+
+- The Data API retains 30 days of history and follows a ~24-hour freshness SLA - this is
+  near-real-time observability, not live tracing.
+- Re-syncing a window is safe: every trace carries a deterministic ``span_id``
+  (``mw:<interaction_id>``) and the engine skips spans it has already ingested.
+- Moveworks does not export token counts, so these traces do not contribute to the LLM cost
+  chart; plugin calls carry served/used flags but no failure semantics, so no tool failures are
+  fabricated.
+- Data API credentials are minted by a Moveworks superadmin (Setup -> Credentials). Field names
+  in Data API records vary between deployments/versions; the importer reads each field from a
+  list of known candidates and tolerates absences.
+"""
+from __future__ import annotations
+
+import argparse
+import json
+import os
+import sys
+from datetime import datetime, timedelta, timezone
+from pathlib import Path
+from typing import Any, Dict, Iterator, List, Optional
+
+import requests
+
+from agentx.tracing.ingest_client import IngestClient
+from agentx.version import VERSION
+
+_DEFAULT_BASE_URL = "https://api.moveworks.ai/export/v1"
+_DEFAULT_TIMESTAMP_FIELD = "created_time"
+_DEFAULT_CURSOR_FILE = Path.home() / ".agentx" / "moveworks_cursor.json"
+
+
+# ----------------------------------------------------------------------------------------------
+# Defensive record readers - Data API field names vary between deployments/versions.
+# ----------------------------------------------------------------------------------------------
+
+def _first(record: Dict[str, Any], *keys: str) -> Any:
+    for key in keys:
+        value = record.get(key)
+        if value not in (None, ""):
+            return value
+    return None
+
+
+def _parse_time(value: Any) -> Optional[datetime]:
+    """ISO-8601 strings and epoch seconds/millis, to aware UTC datetimes."""
+    if value is None:
+        return None
+    if isinstance(value, (int, float)):
+        # Millisecond epochs are > year 33658 when read as seconds; use magnitude to decide.
+        seconds = value / 1000.0 if value > 1e11 else float(value)
+        return datetime.fromtimestamp(seconds, tz=timezone.utc)
+    if isinstance(value, str):
+        text = value.strip().replace("Z", "+00:00")
+        try:
+            parsed = datetime.fromisoformat(text)
+        except ValueError:
+            return None
+        return parsed if parsed.tzinfo else parsed.replace(tzinfo=timezone.utc)
+    return None
+
+
+def _record_time(record: Dict[str, Any]) -> Optional[datetime]:
+    return _parse_time(
+        _first(record, "created_time", "created_at", "timestamp", "event_time", "time", "start_time")
+    )
+
+
+class MoveworksDataAPIClient:
+    """
+    Minimal client for the Moveworks Data API (OData, GET-only).
+
+    ``api_key`` is sent as ``Authorization: Bearer <key>`` by default; override ``auth_header`` /
+    ``auth_scheme`` if your deployment differs. Pagination follows ``odata.nextlink``.
+    """
+
+    def __init__(
+        self,
+        api_key: str,
+        *,
+        base_url: str = _DEFAULT_BASE_URL,
+        auth_header: str = "Authorization",
+        auth_scheme: str = "Bearer",
+        timeout: float = 30.0,
+        session: Optional[requests.Session] = None,
+    ) -> None:
+        if not api_key:
+            raise ValueError("A Moveworks Data API key is required (minted by a Moveworks superadmin)")
+        self._base_url = base_url.rstrip("/")
+        self._timeout = timeout
+        self._session = session or requests.Session()
+        value = f"{auth_scheme} {api_key}".strip() if auth_scheme else api_key
+        self._session.headers.update({auth_header: value, "Accept": "application/json"})
+
+    def records(
+        self,
+        table: str,
+        *,
+        since: Optional[datetime] = None,
+        until: Optional[datetime] = None,
+        timestamp_field: str = _DEFAULT_TIMESTAMP_FIELD,
+        extra_params: Optional[Dict[str, str]] = None,
+        max_pages: int = 1000,
+    ) -> Iterator[Dict[str, Any]]:
+        """Yield records from ``/records/<table>``, following OData pagination."""
+        filters: List[str] = []
+        if since is not None:
+            filters.append(f"{timestamp_field} ge {since.astimezone(timezone.utc).isoformat()}")
+        if until is not None:
+            filters.append(f"{timestamp_field} lt {until.astimezone(timezone.utc).isoformat()}")
+        params: Dict[str, str] = dict(extra_params or {})
+        if filters:
+            params["$filter"] = " and ".join(filters)
+
+        url: Optional[str] = f"{self._base_url}/records/{table}"
+        for _ in range(max_pages):
+            if not url:
+                return
+            response = self._session.get(url, params=params, timeout=self._timeout)
+            response.raise_for_status()
+            body = response.json()
+            for record in body.get("value", []):
+                if isinstance(record, dict):
+                    yield record
+            # nextlink is absolute and already carries its own query string.
+            url = body.get("odata.nextlink") or body.get("@odata.nextLink")
+            params = {}
+
+
+class MoveworksSyncReport:
+    def __init__(self) -> None:
+        self.conversations = 0
+        self.interactions = 0
+        self.ingested = 0
+        self.failed = 0
+        self.plugin_calls_attached = 0
+        self.skipped_no_time = 0
+        # Distinct AgentX session ids touched by this run - what judge_sessions() operates on.
+        self.session_ids: List[str] = []
+        self.sessions_judged = 0
+        self.sessions_judge_skipped = 0
+        self.sessions_judge_failed = 0
+
+    def __repr__(self) -> str:  # also what the CLI prints
+        base = (
+            f"MoveworksSyncReport(conversations={self.conversations}, interactions={self.interactions}, "
+            f"ingested={self.ingested}, failed={self.failed}, "
+            f"plugin_calls_attached={self.plugin_calls_attached}, skipped_no_time={self.skipped_no_time}"
+        )
+        if self.sessions_judged or self.sessions_judge_skipped or self.sessions_judge_failed:
+            base += (
+                f", sessions_judged={self.sessions_judged}, "
+                f"sessions_judge_skipped={self.sessions_judge_skipped}, "
+                f"sessions_judge_failed={self.sessions_judge_failed}"
+            )
+        return base + ")"
+
+
+class MoveworksImporter:
+    """
+    Pulls Moveworks Data API records and replays them into AgentX as traces.
+
+    ``agent_name`` overrides the agent every trace is attributed to; by default traces group per
+    Moveworks domain as ``moveworks-<domain>`` (falling back to ``moveworks-assistant``).
+    """
+
+    def __init__(
+        self,
+        client: MoveworksDataAPIClient,
+        *,
+        agentx_api_key: str,
+        agentx_base_url: Optional[str] = None,
+        agent_name: Optional[str] = None,
+        timestamp_field: str = _DEFAULT_TIMESTAMP_FIELD,
+    ) -> None:
+        self._client = client
+        self._agent_name = agent_name
+        self._timestamp_field = timestamp_field
+        self._ingest = IngestClient(agentx_api_key, sdk_version=VERSION, base_url=agentx_base_url)
+
+    # -- record -> wire mapping ----------------------------------------------------------------
+
+    def _conversation_index(self, since: datetime, until: Optional[datetime]) -> Dict[str, Dict[str, Any]]:
+        index: Dict[str, Dict[str, Any]] = {}
+        try:
+            for conversation in self._client.records(
+                "conversations", since=since, until=until, timestamp_field=self._timestamp_field
+            ):
+                conversation_id = _first(conversation, "conversation_id", "id")
+                if conversation_id:
+                    index[str(conversation_id)] = conversation
+        except requests.RequestException:
+            # Conversations only enrich metadata (domain/route); interactions alone still import.
+            pass
+        return index
+
+    def _plugin_calls_index(self, since: datetime, until: Optional[datetime]) -> Dict[str, List[Dict[str, Any]]]:
+        index: Dict[str, List[Dict[str, Any]]] = {}
+        try:
+            for call in self._client.records(
+                "plugin-calls", since=since, until=until, timestamp_field=self._timestamp_field
+            ):
+                key = _first(call, "interaction_id", "conversation_id")
+                if key:
+                    index.setdefault(str(key), []).append(call)
+        except requests.RequestException:
+            pass
+        return index
+
+    @staticmethod
+    def _tool_call_wire(call: Dict[str, Any]) -> Dict[str, Any]:
+        wire: Dict[str, Any] = {"name": str(_first(call, "plugin_name", "name") or "moveworks-plugin")}
+        served = _first(call, "served", "is_served")
+        used = _first(call, "used", "is_used")
+        # served/used are surfacing metrics, not success/failure - never fabricate a failure.
+        if served is not None:
+            wire["served"] = bool(served)
+        if used is not None:
+            wire["used"] = bool(used)
+        return wire
+
+    def _interaction_wire(
+        self,
+        interaction: Dict[str, Any],
+        conversations: Dict[str, Dict[str, Any]],
+        plugin_calls: Dict[str, List[Dict[str, Any]]],
+    ) -> Optional[Dict[str, Any]]:
+        started = _record_time(interaction)
+        if started is None:
+            return None
+
+        interaction_id = str(_first(interaction, "interaction_id", "id") or "")
+        conversation_id = _first(interaction, "conversation_id")
+        conversation = conversations.get(str(conversation_id)) if conversation_id else None
+
+        domain = _first(interaction, "domain") or (conversation and _first(conversation, "domain"))
+        route = _first(interaction, "route") or (conversation and _first(conversation, "route"))
+        interaction_type = _first(interaction, "interaction_type", "type")
+
+        user_text = _first(interaction, "utterance", "user_input", "text", "message", "query", "content")
+        response_text = _first(interaction, "response", "bot_response", "reply", "answer")
+
+        calls: List[Dict[str, Any]] = []
+        if interaction_id and interaction_id in plugin_calls:
+            calls = plugin_calls[interaction_id]
+        elif conversation_id and str(conversation_id) in plugin_calls:
+            calls = plugin_calls[str(conversation_id)]
+
+        metadata: Dict[str, Any] = {"source": "moveworks"}
+        for key, value in (
+            ("conversationId", conversation_id),
+            ("interactionId", interaction_id or None),
+            ("interactionType", interaction_type),
+            ("domain", domain),
+            ("route", route),
+            ("userId", _first(interaction, "user_id")),
+        ):
+            if value:
+                metadata[key] = value
+
+        latency_ms: Optional[int] = None
+        ended = _parse_time(_first(interaction, "resolved_time", "completed_time", "end_time", "updated_time"))
+        if ended is not None and ended >= started:
+            latency_ms = int((ended - started).total_seconds() * 1000)
+
+        name = self._agent_name or (f"moveworks-{domain}" if domain else "moveworks-assistant")
+
+        wire: Dict[str, Any] = {
+            "name": name,
+            "framework": "moveworks",
+            "metadata": metadata,
+            "started_at_unix_nano": str(int(started.timestamp() * 1_000_000_000)),
+        }
+        if interaction_id:
+            wire["span_id"] = f"mw:{interaction_id}"
+        if conversation_id:
+            wire["session_id"] = f"mw_{conversation_id}"
+        if user_text is not None:
+            wire["input"] = user_text
+        else:
+            wire["input"] = json.dumps({"interaction_type": interaction_type or "unknown"})
+        if response_text is not None:
+            wire["output"] = response_text
+        if latency_ms is not None:
+            wire["latency_ms"] = latency_ms
+        if calls:
+            wire["tool_calls"] = [self._tool_call_wire(call) for call in calls]
+        return wire
+
+    # -- public --------------------------------------------------------------------------------
+
+    def sync(
+        self,
+        since: datetime,
+        until: Optional[datetime] = None,
+        *,
+        monitor: bool = False,
+        judge_sessions: bool = False,
+        dry_run: bool = False,
+        on_payload: Optional[Any] = None,
+    ) -> MoveworksSyncReport:
+        """
+        Import every interaction in ``[since, until)``. Safe to re-run over the same window - the
+        engine deduplicates on the deterministic ``span_id`` (and skips re-judging deduped spans).
+
+        ``monitor=True`` sets ``monitor: true`` on every trace - the engine's explicit opt-in that
+        runs pattern/built-in checks (PII, empty response, tool failure, active patterns) on each
+        imported trace. ``judge_sessions=True`` additionally asks the engine to judge every
+        imported session with each enabled session-scoped evaluator after the sync; the request
+        carries ``ifStale=true`` so a session already scored (e.g. by the engine's own 24h sweep)
+        is never judged twice. Returns a summary report.
+        """
+        report = MoveworksSyncReport()
+        conversations = self._conversation_index(since, until)
+        report.conversations = len(conversations)
+        plugin_calls = self._plugin_calls_index(since, until)
+        session_ids: List[str] = []
+
+        for interaction in self._client.records(
+            "interactions", since=since, until=until, timestamp_field=self._timestamp_field
+        ):
+            report.interactions += 1
+            wire = self._interaction_wire(interaction, conversations, plugin_calls)
+            if wire is None:
+                report.skipped_no_time += 1
+                continue
+            if monitor:
+                wire["monitor"] = True
+            report.plugin_calls_attached += len(wire.get("tool_calls", []))
+            session_id = wire.get("session_id")
+            if session_id and session_id not in session_ids:
+                session_ids.append(session_id)
+            if on_payload is not None:
+                on_payload(wire)
+            if dry_run:
+                continue
+            if self._ingest.send_trace_sync(wire) is not None:
+                report.ingested += 1
+            else:
+                report.failed += 1
+
+        report.session_ids = session_ids
+        if judge_sessions and not dry_run:
+            self.judge_sessions(session_ids, report=report)
+        return report
+
+    def judge_sessions(
+        self, session_ids: List[str], *, report: Optional[MoveworksSyncReport] = None
+    ) -> MoveworksSyncReport:
+        """
+        Judge each session with every enabled session-scoped evaluator via the engine's on-demand
+        judge route, with ``ifStale=true`` so an up-to-date verdict (from the sweep or a previous
+        run) is left alone instead of judged again.
+        """
+        report = report or MoveworksSyncReport()
+        # Same base/key the ingest client already resolved - the judge routes live under
+        # /agent-monitoring on the same engine.
+        base = self._ingest._base_url
+        http = self._ingest._session
+        try:
+            response = http.get(f"{base}/agent-monitoring/online-evaluators", timeout=30)
+            response.raise_for_status()
+            evaluators = [
+                e for e in response.json().get("evaluators", [])
+                if e.get("enabled") and e.get("scope") == "session"
+            ]
+        except requests.RequestException:
+            report.sessions_judge_failed += len(session_ids)
+            return report
+        for session_id in session_ids:
+            for evaluator in evaluators:
+                try:
+                    result = http.post(
+                        f"{base}/agent-monitoring/sessions/{session_id}/judge/{evaluator['_id']}",
+                        params={"ifStale": "true"},
+                        timeout=120,
+                    )
+                    if result.status_code == 200 and result.json().get("skipped"):
+                        report.sessions_judge_skipped += 1
+                    elif result.ok:
+                        report.sessions_judged += 1
+                    else:
+                        report.sessions_judge_failed += 1
+                except requests.RequestException:
+                    report.sessions_judge_failed += 1
+        return report
+
+
+# ----------------------------------------------------------------------------------------------
+# CLI: agentx-moveworks sync
+# ----------------------------------------------------------------------------------------------
+
+def _parse_since(value: str) -> datetime:
+    """Either an ISO timestamp or a relative shorthand like ``7d`` / ``24h`` / ``30m``."""
+    text = value.strip().lower()
+    if text and text[-1] in ("d", "h", "m") and text[:-1].isdigit():
+        amount = int(text[:-1])
+        delta = {"d": timedelta(days=amount), "h": timedelta(hours=amount), "m": timedelta(minutes=amount)}[text[-1]]
+        return datetime.now(timezone.utc) - delta
+    parsed = _parse_time(value)
+    if parsed is None:
+        raise SystemExit(f"agentx-moveworks: could not parse --since {value!r} (use ISO-8601 or e.g. 7d)")
+    return parsed
+
+
+def _read_cursor(path: Path) -> Optional[datetime]:
+    try:
+        return _parse_time(json.loads(path.read_text()).get("last_synced"))
+    except (OSError, ValueError, json.JSONDecodeError):
+        return None
+
+
+def _write_cursor(path: Path, value: datetime) -> None:
+    path.parent.mkdir(parents=True, exist_ok=True)
+    path.write_text(json.dumps({"last_synced": value.isoformat()}))
+
+
+def cli_main(argv: Optional[List[str]] = None) -> None:
+    parser = argparse.ArgumentParser(
+        prog="agentx-moveworks",
+        description="Sync Moveworks Data API activity into AgentX as traces/sessions.",
+    )
+    sub = parser.add_subparsers(dest="command", required=True)
+    sync = sub.add_parser("sync", help="Import a time window of Moveworks activity")
+    sync.add_argument("--since", help="ISO timestamp or relative (7d, 24h). Default: cursor file, else 24h")
+    sync.add_argument("--until", help="ISO timestamp (default: now)")
+    sync.add_argument("--cursor-file", default=str(_DEFAULT_CURSOR_FILE), help="Incremental cursor path")
+    sync.add_argument("--no-cursor", action="store_true", help="Ignore and don't update the cursor file")
+    sync.add_argument("--agent-name", help="Attribute every trace to this agent (default: per Moveworks domain)")
+    sync.add_argument("--timestamp-field", default=_DEFAULT_TIMESTAMP_FIELD, help="Data API record timestamp field")
+    sync.add_argument("--moveworks-base-url", default=os.getenv("MOVEWORKS_BASE_URL", _DEFAULT_BASE_URL))
+    sync.add_argument(
+        "--monitor",
+        action="store_true",
+        help="Run pattern/built-in checks (PII, empty response, tool failure, active patterns) on each imported trace",
+    )
+    sync.add_argument(
+        "--judge-sessions",
+        action="store_true",
+        help="After the sync, judge each imported session with every enabled session-scoped evaluator "
+        "(skips sessions that already have an up-to-date verdict, so the engine's own sweep is never duplicated)",
+    )
+    sync.add_argument("--dry-run", action="store_true", help="Map and print payloads without ingesting")
+
+    args = parser.parse_args(argv)
+
+    moveworks_key = os.getenv("MOVEWORKS_API_KEY", "")
+    agentx_key = os.getenv("AGENTX_API_KEY", "")
+    if not moveworks_key:
+        raise SystemExit("agentx-moveworks: set MOVEWORKS_API_KEY (Data API credential from Moveworks Setup)")
+    if not agentx_key and not args.dry_run:
+        raise SystemExit("agentx-moveworks: set AGENTX_API_KEY (your AgentX project API key)")
+
+    cursor_path = Path(args.cursor_file)
+    since = (
+        _parse_since(args.since)
+        if args.since
+        else ((None if args.no_cursor else _read_cursor(cursor_path)) or (datetime.now(timezone.utc) - timedelta(hours=24)))
+    )
+    until = _parse_time(args.until) if args.until else datetime.now(timezone.utc)
+
+    client = MoveworksDataAPIClient(moveworks_key, base_url=args.moveworks_base_url)
+    importer = MoveworksImporter(
+        client,
+        agentx_api_key=agentx_key or "dry-run",
+        agentx_base_url=os.getenv("AGENTX_API_BASE_URL"),
+        agent_name=args.agent_name,
+        timestamp_field=args.timestamp_field,
+    )
+
+    print(f"agentx-moveworks: syncing {since.isoformat()} -> {until.isoformat()}")
+    on_payload = (lambda wire: print(json.dumps(wire))) if args.dry_run else None
+    report = importer.sync(
+        since,
+        until,
+        monitor=args.monitor,
+        judge_sessions=args.judge_sessions,
+        dry_run=args.dry_run,
+        on_payload=on_payload,
+    )
+    print(report)
+
+    if not args.dry_run and not args.no_cursor and report.failed == 0:
+        _write_cursor(cursor_path, until)
+        print(f"agentx-moveworks: cursor -> {cursor_path}")
+    elif report.failed:
+        print("agentx-moveworks: failures occurred - cursor NOT advanced, re-run to retry", file=sys.stderr)
+        sys.exit(1)
+
+
+if __name__ == "__main__":
+    cli_main()

{agentx_python-0.6.19 → agentx_python-0.6.21}/agentx/tracing/tracer.py

@@ -99,9 +99,12 @@ class _TraceSpan:
         self._parent_span_id: Optional[str] = None
         # Numbers auto-named "LLM Call N"/"Retrieval N" child spans - see _merge_child_run.
         self._child_span_count = 0
-        # Monitor: check this trace against patterns immediately on ingest, no dashboard profile
-        # required. pattern_ids (if given) fully defines what's checked - only those patterns run,
-        # the built-in checks are skipped. See Tracer.trace()'s monitor/pattern_ids params.
+        # Monitor: True checks this trace against patterns immediately on ingest, no dashboard
+        # profile required; False explicitly OPTS OUT of every ingest-time check (patterns,
+        # built-ins, online/custom evaluators, topics) - what eval-run traces send, since the
+        # run's own evaluator already judges each case. None (default) leaves the server's
+        # standard behavior. pattern_ids (if given) fully defines what's checked. See
+        # Tracer.trace()'s monitor/pattern_ids params.
         self._monitor = monitor
         self._pattern_ids = pattern_ids
 
@@ -152,7 +155,7 @@ class _TraceSpan:
 
         self._trace_id = self._tracer._send(
             sync=self._sync,
-            monitor=self._monitor or None,
+            monitor=self._monitor,
             pattern_ids=self._pattern_ids,
             name=self.name,
             agent_id=self._agent_id,
@@ -769,7 +772,7 @@ class Tracer:
         model: Optional[str] = None,
         session_id: Optional[str] = None,
         sync: bool = False,
-        monitor: bool = False,
+        monitor: Optional[bool] = None,
         pattern_ids: Optional[List[str]] = None,
         agent_id: Optional[str] = None,
     ) -> _TraceSpan:
@@ -799,7 +802,9 @@ class Tracer:
             return {"output": resp, "trace_id": span.trace_id}
 
         Pass ``monitor=True`` to check this trace against Monitor patterns immediately, with no
-        dashboard profile required. ``pattern_ids`` (ids from ``client.monitor.patterns.builder(
+        dashboard profile required. Pass ``monitor=False`` to explicitly skip EVERY ingest-time
+        check (patterns, built-ins, online/custom evaluators, topics) - use this for traces
+        created inside an evaluation run, which the run's own evaluator already judges. ``pattern_ids`` (ids from ``client.monitor.patterns.builder(
         ...).publish()``) restricts detection to exactly those patterns; omit it to run the full
         default sweep (built-in checks plus every enabled workspace pattern)::
 

agentx_python-0.6.21/agentx/version.py

@@ -0,0 +1 @@
+VERSION = "0.6.21"

{agentx_python-0.6.19 → agentx_python-0.6.21}/agentx_python.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: agentx-python
-Version: 0.6.19
+Version: 0.6.21
 Summary: Official Python SDK for AgentX (https://www.agentx.so/)
 Home-page: https://github.com/AgentX-ai/AgentX-python
 Author: Robin Wang and AgentX Team

{agentx_python-0.6.19 → agentx_python-0.6.21}/agentx_python.egg-info/SOURCES.txt

@@ -36,6 +36,7 @@ agentx/integrations/google_genai.py
 agentx/integrations/langchain.py
 agentx/integrations/litellm.py
 agentx/integrations/llamaindex.py
+agentx/integrations/moveworks.py
 agentx/integrations/openai.py
 agentx/integrations/openai_agents.py
 agentx/monitor/__init__.py

agentx_python-0.6.21/agentx_python.egg-info/entry_points.txt

@@ -0,0 +1,3 @@
+[console_scripts]
+agentx-moveworks = agentx.integrations.moveworks:cli_main
+agentx-trace-eval = agentx.cli:main

{agentx_python-0.6.19 → agentx_python-0.6.21}/setup.py

@@ -30,7 +30,12 @@ setup(
         # Thin launcher for the self-hostable governance engine (AgentX-ai/AgentX-trace-eval),
         # not this SDK itself, see agentx/cli.py's module docstring for why it lazily downloads
         # rather than bundling that engine in this package.
-        "console_scripts": ["agentx-trace-eval=agentx.cli:main"],
+        "console_scripts": [
+            "agentx-trace-eval=agentx.cli:main",
+            # Moveworks Data API -> AgentX traces sync (see agentx/integrations/moveworks.py) -
+            # a pull importer, since Moveworks agents run in their cloud with no in-process hook.
+            "agentx-moveworks=agentx.integrations.moveworks:cli_main",
+        ],
     },
     install_requires=[
         "urllib3>=1.26.11",

agentx_python-0.6.19/agentx/version.py

@@ -1 +0,0 @@
-VERSION = "0.6.19"

agentx_python-0.6.19/agentx_python.egg-info/entry_points.txt

@@ -1,2 +0,0 @@
-[console_scripts]
-agentx-trace-eval = agentx.cli:main