feat: Send GenAI spans as V2 envelope items

sentry_sdk/client.py

@@ -2,7 +2,7 @@
 import uuid
 import random
 import socket
-from collections.abc import Mapping
+from collections.abc import Mapping, Iterable
 from datetime import datetime, timezone
 from importlib import import_module
 from typing import TYPE_CHECKING, List, Dict, cast, overload
@@ -27,6 +27,7 @@
     get_before_send_metric,
     has_logs_enabled,
     has_metrics_enabled,
+    serialize_attribute,
 )
 from sentry_sdk.serializer import serialize
 from sentry_sdk.tracing import trace
@@ -56,6 +57,8 @@
 )
 from sentry_sdk.scrubber import EventScrubber
 from sentry_sdk.monitor import Monitor
+from sentry_sdk.envelope import Item, PayloadRef
+from sentry_sdk.utils import datetime_from_isoformat
 
 if TYPE_CHECKING:
     from typing import Any
@@ -89,6 +92,125 @@
 }
 
 
+def _serialized_v1_span_to_serialized_v2_span(
+    span: "Dict[str, Any]", event: "Event"
+) -> "dict[str, Any]":
+    # See SpanBatcher._to_transport_format() for analogous population of all entries except "attributes".
+    res: "Dict[str, Any]" = {
+        "status": "ok",
+        "is_segment": False,
+    }
+
+    if "trace_id" in span:
+        res["trace_id"] = span["trace_id"]
+
+    if "span_id" in span:
+        res["span_id"] = span["span_id"]
+
+    if "description" in span:
+        res["name"] = span["description"]
+
+    if "start_timestamp" in span:
+        start_timestamp = None
+        try:
+            start_timestamp = datetime_from_isoformat(span["start_timestamp"])
+        except Exception:
+            pass
+
+        if start_timestamp is not None:
+            res["start_timestamp"] = start_timestamp.timestamp()
+
+    if "timestamp" in span:
+        end_timestamp = None
+        try:
+            end_timestamp = datetime_from_isoformat(span["timestamp"])
+        except Exception:
+            pass
+
+        if end_timestamp is not None:
+            res["end_timestamp"] = end_timestamp.timestamp()
+
+    if "parent_span_id" in span:
+        res["parent_span_id"] = span["parent_span_id"]
+
+    if "status" in span and span["status"] != "ok":
+        res["status"] = "error"
+
+    attributes: "Dict[str, Any]" = {}
+
+    if "op" in span:
+        attributes["sentry.op"] = span["op"]
+    if "origin" in span:
+        attributes["sentry.origin"] = span["origin"]
+
+    span_data = span.get("data")
+    if isinstance(span_data, dict):
+        attributes.update(span_data)
+
+    span_tags = span.get("tags")
+    if isinstance(span_tags, dict):
+        attributes.update(span_tags)
+
+    # See Scope._apply_user_attributes_to_telemetry() for user attributes.
+    user = event.get("user")
+    if isinstance(user, dict):
+        if "id" in user:
+            attributes["user.id"] = user["id"]
+        if "username" in user:
+            attributes["user.name"] = user["username"]
+        if "email" in user:
+            attributes["user.email"] = user["email"]
+
+    # See Scope.set_global_attributes() for release, environment, and SDK metadata.
+    if "release" in event:
+        attributes["sentry.release"] = event["release"]
+    if "environment" in event:
+        attributes["sentry.environment"] = event["environment"]
+    if "transaction" in event:
+        attributes["sentry.segment.name"] = event["transaction"]
+
+    trace_context = event.get("contexts", {}).get("trace", {})
+    if "span_id" in trace_context:
+        attributes["sentry.segment.id"] = trace_context["span_id"]
+
+    sdk_info = event.get("sdk")
+    if isinstance(sdk_info, dict):
+        if "name" in sdk_info:
+            attributes["sentry.sdk.name"] = sdk_info["name"]
+        if "version" in sdk_info:
+            attributes["sentry.sdk.version"] = sdk_info["version"]
+
+    if attributes:
+        res["attributes"] = {k: serialize_attribute(v) for k, v in attributes.items()}
+
+    return res
+
+
+def _split_gen_ai_spans(
+    event_opt: "Event",
+) -> "Optional[tuple[List[Dict[str, object]], List[Dict[str, object]]]]":
+    if "spans" not in event_opt:
+        return None
+
+    spans = event_opt["spans"]
+    if isinstance(spans, AnnotatedValue):
+        spans = spans.value
+
+    if not isinstance(spans, Iterable):
+        return None
+
+    non_gen_ai_spans = []
+    gen_ai_spans = []
+    for span in spans:
+        span_op = span.get("op")
+        if isinstance(span_op, str) and span_op.startswith("gen_ai."):
+            gen_ai_spans.append(span)
+        else:
+            non_gen_ai_spans.append(span)
+
+    return non_gen_ai_spans, gen_ai_spans
+
+
 def _get_options(*args: "Optional[str]", **kwargs: "Any") -> "Dict[str, Any]":
     if args and (isinstance(args[0], (bytes, str)) or args[0] is None):
         dsn: "Optional[str]" = args[0]
@@ -912,7 +1034,37 @@
         if is_transaction:
             if isinstance(profile, Profile):
                 envelope.add_profile(profile.to_json(event_opt, self.options))
-            envelope.add_transaction(event_opt)
+
+            split_spans = _split_gen_ai_spans(event_opt)
+            if split_spans is None or not split_spans[1]:
+                envelope.add_transaction(event_opt)
+            else:
+                non_gen_ai_spans, gen_ai_spans = split_spans
+
+                event_opt["spans"] = non_gen_ai_spans
+                envelope.add_transaction(event_opt)
+
+                envelope.add_item(
+                    Item(
+                        type=SpanBatcher.TYPE,
+                        content_type=SpanBatcher.CONTENT_TYPE,
+                        headers={
+                            "item_count": len(gen_ai_spans),
+                        },
+                        payload=PayloadRef(
+                            json={
+                                "items": [
+                                    _serialized_v1_span_to_serialized_v2_span(
+                                        span, event
+                                    )
+                                    for span in gen_ai_spans
+                                    if isinstance(span, dict)
+                                ]
+                            },
+                        ),
+                    )
+                )
+
         elif is_checkin:
             envelope.add_checkin(event_opt)
         else: