Merge pull request #1112 from newrelic/develop-ai-limited-preview-3

Develop ai limited preview 3

newrelic/agent.py

@@ -153,6 +153,10 @@ def __asgi_application(*args, **kwargs):
 from newrelic.api.message_transaction import (
     wrap_message_transaction as __wrap_message_transaction,
 )
+from newrelic.api.ml_model import (
+    record_llm_feedback_event as __record_llm_feedback_event,
+)
+from newrelic.api.ml_model import set_llm_token_count_callback as __set_llm_token_count_callback
 from newrelic.api.ml_model import wrap_mlmodel as __wrap_mlmodel
 from newrelic.api.profile_trace import ProfileTraceWrapper as __ProfileTraceWrapper
 from newrelic.api.profile_trace import profile_trace as __profile_trace
@@ -174,10 +178,10 @@ def __asgi_application(*args, **kwargs):
 from newrelic.api.web_transaction import web_transaction as __web_transaction
 from newrelic.api.web_transaction import wrap_web_transaction as __wrap_web_transaction
 from newrelic.common.object_names import callable_name as __callable_name
+from newrelic.common.object_wrapper import CallableObjectProxy as __CallableObjectProxy
 from newrelic.common.object_wrapper import FunctionWrapper as __FunctionWrapper
 from newrelic.common.object_wrapper import InFunctionWrapper as __InFunctionWrapper
 from newrelic.common.object_wrapper import ObjectProxy as __ObjectProxy
-from newrelic.common.object_wrapper import CallableObjectProxy as __CallableObjectProxy
 from newrelic.common.object_wrapper import ObjectWrapper as __ObjectWrapper
 from newrelic.common.object_wrapper import OutFunctionWrapper as __OutFunctionWrapper
 from newrelic.common.object_wrapper import PostFunctionWrapper as __PostFunctionWrapper
@@ -343,3 +347,5 @@ def __asgi_application(*args, **kwargs):
 insert_html_snippet = __wrap_api_call(__insert_html_snippet, "insert_html_snippet")
 verify_body_exists = __wrap_api_call(__verify_body_exists, "verify_body_exists")
 wrap_mlmodel = __wrap_api_call(__wrap_mlmodel, "wrap_mlmodel")
+record_llm_feedback_event = __wrap_api_call(__record_llm_feedback_event, "record_llm_feedback_event")
+set_llm_token_count_callback = __wrap_api_call(__set_llm_token_count_callback, "set_llm_token_count_callback")

newrelic/api/ml_model.py

@@ -12,11 +12,18 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+import logging
 import sys
+import uuid
+import warnings
 
+from newrelic.api.transaction import current_transaction
 from newrelic.common.object_names import callable_name
+from newrelic.core.config import global_settings
 from newrelic.hooks.mlmodel_sklearn import _nr_instrument_model
 
+_logger = logging.getLogger(__name__)
+
 
 def wrap_mlmodel(model, name=None, version=None, feature_names=None, label_names=None, metadata=None):
     model_callable_name = callable_name(model)
@@ -33,3 +40,87 @@ def wrap_mlmodel(model, name=None, version=None, feature_names=None, label_names
         model._nr_wrapped_label_names = label_names
     if metadata:
         model._nr_wrapped_metadata = metadata
+
+
+def record_llm_feedback_event(trace_id, rating, category=None, message=None, metadata=None):
+    transaction = current_transaction()
+    if not transaction:
+        warnings.warn(
+            "No message feedback events will be recorded. record_llm_feedback_event must be called within the "
+            "scope of a transaction."
+        )
+        return
+
+    feedback_event_id = str(uuid.uuid4())
+    feedback_event = metadata.copy() if metadata else {}
+    feedback_event.update(
+        {
+            "id": feedback_event_id,
+            "trace_id": trace_id,
+            "rating": rating,
+            "category": category,
+            "message": message,
+            "ingest_source": "Python",
+        }
+    )
+
+    transaction.record_custom_event("LlmFeedbackMessage", feedback_event)
+
+
+def set_llm_token_count_callback(callback, application=None):
+    """
+    Set the current callback to be used to calculate LLM token counts.
+
+    Arguments:
+    callback -- the user-defined callback that will calculate and return the total token count as an integer or None if it does not know
+    application -- optional application object to associate call with
+    """
+    if callback and not callable(callback):
+        _logger.error(
+            "callback passed to set_llm_token_count_callback must be a Callable type or None to unset the callback."
+        )
+        return
+
+    from newrelic.api.application import application_instance
+
+    # Check for activated application if it exists and was not given.
+    application = application or application_instance(activate=False)
+
+    # Get application settings if it exists, or fallback to global settings object.
+    settings = application.settings if application else global_settings()
+
+    if not settings:
+        _logger.error(
+            "Failed to set llm_token_count_callback. Settings not found on application or in global_settings."
+        )
+        return
+
+    if not callback:
+        settings.ai_monitoring._llm_token_count_callback = None
+        return
+
+    def _wrap_callback(model, content):
+        if model is None:
+            _logger.debug(
+                "The model argument passed to the user-defined token calculation callback is None. The callback will not be run."
+            )
+            return None
+
+        if content is None:
+            _logger.debug(
+                "The content argument passed to the user-defined token calculation callback is None. The callback will not be run."
+            )
+            return None
+
+        token_count_val = callback(model, content)
+
+        if not isinstance(token_count_val, int) or token_count_val < 0:
+            _logger.warning(
+                "llm_token_count_callback returned an invalid value of %s. This value must be a positive integer and will not be recorded for the token_count."
+                % token_count_val
+            )
+            return None
+
+        return token_count_val
+
+    settings.ai_monitoring._llm_token_count_callback = _wrap_callback

newrelic/api/time_trace.py

@@ -29,7 +29,6 @@
 )
 from newrelic.core.config import is_expected_error, should_ignore_error
 from newrelic.core.trace_cache import trace_cache
-
 from newrelic.packages import six
 
 _logger = logging.getLogger(__name__)
@@ -260,6 +259,11 @@ def _observe_exception(self, exc_info=None, ignore=None, expected=None, status_c
         module, name, fullnames, message_raw = parse_exc_info((exc, value, tb))
         fullname = fullnames[0]
 
+        # In case message is in JSON format for OpenAI models
+        # this will result in a "cleaner" message format
+        if getattr(value, "_nr_message", None):
+            message_raw = value._nr_message
+
         # Check to see if we need to strip the message before recording it.
 
         if settings.strip_exception_messages.enabled and fullname not in settings.strip_exception_messages.allowlist:
@@ -422,23 +426,32 @@ def notice_error(self, error=None, attributes=None, expected=None, ignore=None,
                         input_attributes = {}
                         input_attributes.update(transaction._custom_params)
                         input_attributes.update(attributes)
-                        error_group_name_raw = settings.error_collector.error_group_callback(value, {
-                            "traceback": tb,
-                            "error.class": exc,
-                            "error.message": message_raw,
-                            "error.expected": is_expected,
-                            "custom_params": input_attributes,
-                            "transactionName": getattr(transaction, "name", None),
-                            "response.status": getattr(transaction, "_response_code", None),
-                            "request.method": getattr(transaction, "_request_method", None),
-                            "request.uri": getattr(transaction, "_request_uri", None),
-                        })
+                        error_group_name_raw = settings.error_collector.error_group_callback(
+                            value,
+                            {
+                                "traceback": tb,
+                                "error.class": exc,
+                                "error.message": message_raw,
+                                "error.expected": is_expected,
+                                "custom_params": input_attributes,
+                                "transactionName": getattr(transaction, "name", None),
+                                "response.status": getattr(transaction, "_response_code", None),
+                                "request.method": getattr(transaction, "_request_method", None),
+                                "request.uri": getattr(transaction, "_request_uri", None),
+                            },
+                        )
                         if error_group_name_raw:
                             _, error_group_name = process_user_attribute("error.group.name", error_group_name_raw)
                             if error_group_name is None or not isinstance(error_group_name, six.string_types):
-                                raise ValueError("Invalid attribute value for error.group.name. Expected string, got: %s" % repr(error_group_name_raw))
+                                raise ValueError(
+                                    "Invalid attribute value for error.group.name. Expected string, got: %s"
+                                    % repr(error_group_name_raw)
+                                )
                     except Exception:
-                        _logger.error("Encountered error when calling error group callback:\n%s", "".join(traceback.format_exception(*sys.exc_info())))
+                        _logger.error(
+                            "Encountered error when calling error group callback:\n%s",
+                            "".join(traceback.format_exception(*sys.exc_info())),
+                        )
                         error_group_name = None
 
             transaction._create_error_node(
@@ -595,13 +608,11 @@ def update_async_exclusive_time(self, min_child_start_time, exclusive_duration):
     def process_child(self, node, is_async):
         self.children.append(node)
         if is_async:
-
             # record the lowest start time
             self.min_child_start_time = min(self.min_child_start_time, node.start_time)
 
             # if there are no children running, finalize exclusive time
             if self.child_count == len(self.children):
-
                 exclusive_duration = node.end_time - self.min_child_start_time
 
                 self.update_async_exclusive_time(self.min_child_start_time, exclusive_duration)

newrelic/api/transaction.py

@@ -176,7 +176,7 @@ def __init__(self, application, enabled=None, source=None):
 
         self.thread_id = None
 
-        self._transaction_id = id(self)
+        self._identity = id(self)
         self._transaction_lock = threading.Lock()
 
         self._dead = False
@@ -193,6 +193,7 @@ def __init__(self, application, enabled=None, source=None):
         self._frameworks = set()
         self._message_brokers = set()
         self._dispatchers = set()
+        self._ml_models = set()
 
         self._frozen_path = None
 
@@ -274,6 +275,7 @@ def __init__(self, application, enabled=None, source=None):
         trace_id = "%032x" % random.getrandbits(128)
 
         # 16-digit random hex. Padded with zeros in the front.
+        # This is the official transactionId in the UI.
         self.guid = trace_id[:16]
 
         # 32-digit random hex. Padded with zeros in the front.
@@ -421,7 +423,7 @@ def __exit__(self, exc, value, tb):
         if not self.enabled:
             return
 
-        if self._transaction_id != id(self):
+        if self._identity != id(self):
             return
 
         if not self._settings:
@@ -568,6 +570,10 @@ def __exit__(self, exc, value, tb):
             for dispatcher, version in self._dispatchers:
                 self.record_custom_metric("Python/Dispatcher/%s/%s" % (dispatcher, version), 1)
 
+        if self._ml_models:
+            for ml_model, version in self._ml_models:
+                self.record_custom_metric("Supportability/Python/ML/%s/%s" % (ml_model, version), 1)
+
         if self._settings.distributed_tracing.enabled:
             # Sampled and priority need to be computed at the end of the
             # transaction when distributed tracing or span events are enabled.
@@ -1715,7 +1721,7 @@ def record_custom_event(self, event_type, params):
         if not settings.custom_insights_events.enabled:
             return
 
-        event = create_custom_event(event_type, params)
+        event = create_custom_event(event_type, params, settings=settings)
         if event:
             self._custom_events.add(event, priority=self.priority)
 
@@ -1728,7 +1734,7 @@ def record_ml_event(self, event_type, params):
         if not settings.ml_insights_events.enabled:
             return
 
-        event = create_custom_event(event_type, params)
+        event = create_custom_event(event_type, params, settings=settings, is_ml_event=True)
         if event:
             self._ml_events.add(event, priority=self.priority)
 
@@ -1835,6 +1841,10 @@ def add_dispatcher_info(self, name, version=None):
         if name:
             self._dispatchers.add((name, version))
 
+    def add_ml_model_info(self, name, version=None):
+        if name:
+            self._ml_models.add((name, version))
+
     def dump(self, file):
         """Dumps details about the transaction to the file object."""