Python APIollama

Core

Instrumentation

Python · Instrumentation

openinference-instrumentation-ollama

OpenInference Ollama Instrumentation

Module
openinference.instrumentation.ollama
Version
0.1.20
Python
>=3.10, <3.16
Instruments
ollama >= 0.4.0

Install

pip install openinference-instrumentation-ollama

Usage

Instrument once, at startup, before the library is used. The OllamaInstrumentor reference below lists the options.

from opentelemetry.sdk.trace import TracerProvider
from openinference.instrumentation import TraceConfig
from openinference.instrumentation.ollama import OllamaInstrumentor

tracer_provider = TracerProvider()
# ...add a span processor that exports to Phoenix or any OTLP backend.

OllamaInstrumentor().instrument(
    tracer_provider=tracer_provider,
    config=TraceConfig(hide_inputs=True),  # optional
)

openinference.instrumentation.ollama

OpenInference instrumentation for the Ollama Python client.

Usage:

“python from openinference.instrumentation.ollama import OllamaInstrumentor

OllamaInstrumentor().instrument(tracer_provider=tracer_provider) “

Chat calls made through ollama.chat, ollama.Client.chat, and ollama.AsyncClient.chat (including stream=True) are exported as OpenInference LLM spans. Call instrument() before making chat calls; references captured earlier (e.g. from ollama import chat at import time) keep the unwrapped method and are not traced. uninstrument() restores the original methods and deactivates any wrapper references that were captured while instrumented.

OllamaInstrumentor

classSource ↗
class OllamaInstrumentor(BaseInstrumentor)

An instrumentor for the Ollama Python client.

instrument() accepts two optional keyword arguments: tracer_provider (an OpenTelemetry TracerProvider) and config (an OpenInference TraceConfig for masking sensitive data). It wraps Client.chat and AsyncClient.chat (covering the module-level ollama.chat helper) so chat calls are exported as OpenInference LLM spans; uninstrument() restores the original methods.

instrument

method
def instrument(
    *,
    tracer_provider: Optional[TracerProvider] = None,
    config: Optional[TraceConfig] = None,
    skip_dep_check: bool = False,
) -> None

Start tracing ollama: patch it so its calls make OpenInference spans. Calling it again does nothing. Inherited from OpenTelemetry’s BaseInstrumentor.

Parameters

NameDescription
tracer_provider

The tracer provider the spans are made with. Defaults to the global one, opentelemetry.trace.get_tracer_provider().

config

What the spans record: hide inputs, outputs or messages, or limit image sizes. Defaults to TraceConfig(), which reads the OPENINFERENCE_* environment variables. See TraceConfig.

skip_dep_check

Instrument even if the installed version of ollama is outside the supported range.

uninstrument

method
def uninstrument() -> None

Stop tracing: undo the patches instrument() made. Inherited from OpenTelemetry’s BaseInstrumentor.

instrumentation_dependencies

methodSource ↗
def instrumentation_dependencies() -> Collection[str]