feat: Add traceloop translator

util/opentelemetry-util-genai-dev/TRANSLATOR_README.md

@@ -0,0 +1,41 @@
+# Translator
+
+## Automatic Span Processing (Recommended)
+
+Add `TraceloopSpanProcessor` to your TracerProvider to automatically transform all matching spans:
+
+```python
+from opentelemetry import trace
+from opentelemetry.sdk.trace import TracerProvider
+from opentelemetry.util.genai.processors import TraceloopSpanProcessor
+
+# Set up tracer provider
+provider = TracerProvider()
+
+# Add processor - transforms all matching spans automatically
+processor = TraceloopSpanProcessor(
+    attribute_transformations={
+        "remove": ["debug_info"],
+        "rename": {"model_ver": "llm.model.version"}, 
+        "add": {"service.name": "my-llm"}
+    },
+    name_transformations={"chat *": "llm.openai.chat"},
+    traceloop_attributes={
+        "traceloop.entity.name": "MyLLMEntity"
+    }
+)
+provider.add_span_processor(processor)
+trace.set_tracer_provider(provider)
+
+```
+
+## Transformation Rules
+
+### Attributes
+- **Remove**: `"remove": ["field1", "field2"]`
+- **Rename**: `"rename": {"old_name": "new_name"}`  
+- **Add**: `"add": {"key": "value"}`
+
+### Span Names
+- **Direct**: `"old name": "new name"`
+- **Pattern**: `"chat *": "llm.chat"` (wildcard matching)

util/opentelemetry-util-genai-dev/examples/traceloop_span_transformation/traceloop_rules_example.py

@@ -0,0 +1,45 @@
+#!/usr/bin/env python3
+
+from __future__ import annotations
+
+import json
+import os
+from opentelemetry import trace
+from opentelemetry.sdk.trace import TracerProvider
+from opentelemetry.sdk.trace.export import SimpleSpanProcessor, ConsoleSpanExporter
+
+from opentelemetry.util.genai.processors.traceloop_span_processor import TraceloopSpanProcessor
+
+RULE_SPEC = {
+    "rules": [
+        {
+            "attribute_transformations": {
+                "rename": {"llm.provider": "ai.system.vendor"},
+                "add": {"example.rule": "chat"},
+            },
+            "name_transformations": {"chat *": "genai.chat"},
+        },
+        {
+            "attribute_transformations": {"add": {"processed.embedding": True}},
+            "name_transformations": {"*": "genai.embedding"},
+        },
+    ]
+}
+os.environ["OTEL_GENAI_SPAN_TRANSFORM_RULES"] = json.dumps(RULE_SPEC)
+
+# Set up tracing
+provider = TracerProvider()
+provider.add_span_processor(SimpleSpanProcessor(ConsoleSpanExporter()))
+# Add the Traceloop processor
+provider.add_span_processor(TraceloopSpanProcessor())
+trace.set_tracer_provider(provider)
+tracer = trace.get_tracer(__name__)
+
+print("Creating spans ...\n")
+
+with tracer.start_as_current_span("chat gpt-4") as span:
+    span.set_attribute("llm.provider", "openai")
+    span.set_attribute("debug_info", "remove me if rule had remove")
+
+with tracer.start_as_current_span("vector encode") as span:
+    span.set_attribute("custom.kind", "embedding")

util/opentelemetry-util-genai-dev/src/opentelemetry/util/genai/processors/__init__.py

@@ -0,0 +1,6 @@
+from .traceloop_span_processor import TraceloopSpanProcessor, TransformationRule
+
+__all__ = [
+    "TraceloopSpanProcessor",
+    "TransformationRule",
+]

util/opentelemetry-util-genai-dev/src/opentelemetry/util/genai/processors/span_transformer.py

@@ -0,0 +1,104 @@
+# Copyright The OpenTelemetry Authors
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from typing import Any, Dict, Optional
+
+from opentelemetry.sdk.trace import ReadableSpan
+
+from .traceloop_span_generator import TraceloopSpanGenerator
+from ..types import LLMInvocation
+
+
+def _apply_attribute_transformations(
+    base: Dict[str, Any], transformations: Optional[Dict[str, Any]]
+) -> Dict[str, Any]:  # pragma: no cover - trivial helpers
+    if not transformations:
+        return base
+    # Order: remove -> rename -> add (so add always wins)
+    remove_keys = transformations.get("remove") or []
+    for k in remove_keys:
+        base.pop(k, None)
+    rename_map = transformations.get("rename") or {}
+    for old, new in rename_map.items():
+        if old in base:
+            base[new] = base.pop(old)
+    add_map = transformations.get("add") or {}
+    for k, v in add_map.items():
+        base[k] = v
+    return base
+
+
+def _derive_new_name(
+    original_name: str, name_transformations: Optional[Dict[str, str]]
+) -> Optional[str]:  # pragma: no cover - simple matching
+    if not name_transformations:
+        return None
+    import fnmatch
+
+    for pattern, new_name in name_transformations.items():
+        try:
+            if fnmatch.fnmatch(original_name, pattern):
+                return new_name
+        except Exception:  # defensive
+            continue
+    return None
+
+
+def transform_existing_span_to_telemetry(
+    existing_span: ReadableSpan,
+    attribute_transformations: Optional[Dict[str, Any]] = None,
+    name_transformations: Optional[Dict[str, str]] = None,
+    traceloop_attributes: Optional[Dict[str, Any]] = None,
+    generator: Optional[TraceloopSpanGenerator] = None,
+) -> LLMInvocation:
+    """Create a synthetic LLMInvocation span from an ended (or ending) span.
+
+    Returns the synthetic ``LLMInvocation`` used purely as a carrier for the new span.
+    """
+    base_attrs: Dict[str, Any] = (
+        dict(existing_span.attributes) if existing_span.attributes else {}
+    )
+
+    # Apply transformations
+    base_attrs = _apply_attribute_transformations(
+        base_attrs, attribute_transformations
+    )
+    if traceloop_attributes:
+        base_attrs.update(traceloop_attributes)
+
+    # Span name rewrite (store so generator can use & remove later)
+    new_name = _derive_new_name(existing_span.name, name_transformations)
+    if new_name:
+        base_attrs["_traceloop_new_name"] = new_name
+
+    # Determine request_model (best-effort, fallback to unknown)
+    request_model = (
+        base_attrs.get("gen_ai.request.model")
+        or base_attrs.get("llm.request.model")
+        or base_attrs.get("ai.model.name")
+        or "unknown"
+    )
+
+    invocation = LLMInvocation(
+        request_model=str(request_model),
+        attributes=base_attrs,
+        messages=[],  # empty; original content not reconstructed here
+    )
+
+    if generator is None:
+        generator = TraceloopSpanGenerator(capture_content=True)
+    generator.start(invocation)
+    if existing_span.end_time is not None:
+        generator.finish(invocation)
+    return invocation

util/opentelemetry-util-genai-dev/src/opentelemetry/util/genai/processors/traceloop_span_generator.py

@@ -0,0 +1,62 @@
+# Copyright The OpenTelemetry Authors
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from typing import Optional
+from opentelemetry import trace
+from opentelemetry.trace import Tracer
+from ..types import LLMInvocation
+
+class TraceloopSpanGenerator:
+    def __init__(self, tracer: Optional[Tracer] = None, capture_content: bool = False):
+        self._tracer = tracer or trace.get_tracer(__name__)
+        self._capture_content = capture_content
+
+    def start(self, invocation: LLMInvocation):
+        override = getattr(invocation, "attributes", {}).get("_traceloop_new_name")
+        if override:
+            span_name = override
+        else:
+            name = getattr(invocation, "request_model", "llm")
+            span_name = f"chat {name}" if not str(name).startswith("chat ") else str(name)
+        span = self._tracer.start_span(span_name, kind=trace.SpanKind.CLIENT)
+        invocation.span = span
+        invocation.context_token = trace.use_span(span, end_on_exit=False)
+        invocation.context_token.__enter__()
+        # apply starting attributes
+        for k, v in getattr(invocation, "attributes", {}).items():
+            try:
+                span.set_attribute(k, v)
+            except Exception:
+                pass
+
+    def finish(self, invocation: LLMInvocation):
+        span = getattr(invocation, "span", None)
+        if not span:
+            return
+        # re-apply attributes (after transformations)
+        for k, v in getattr(invocation, "attributes", {}).items():
+            try:
+                span.set_attribute(k, v)
+            except Exception:
+                pass
+        token = getattr(invocation, "context_token", None)
+        if token and hasattr(token, "__exit__"):
+            try:
+                token.__exit__(None, None, None)
+            except Exception:
+                pass
+        span.end()
+
+    def error(self, error, invocation: LLMInvocation):  # pragma: no cover - unused in tests now
+        self.finish(invocation)