op

Python SDK reference for weave.trace.op

API Overview

---

class DisplayNameFuncError

---

class OpCallError

---

class OpKwargs

TypedDict for op() keyword arguments.

---

class Sentinel

Sentinel(package: ‘str’, path: ‘str’, name: ‘str’)

method __init__

__init__(package: 'str', path: 'str', name: 'str') → None

---

class WeaveKwargs

---

function as_op

as_op(fn: 'Callable[P, R]') → Op[P, R]

Given a @weave.op() decorated function, return its Op.

@weave.op() decorated functions are instances of Op already, so this function should be a no-op at runtime. But you can use it to satisfy type checkers if you need to access OpDef attributes in a typesafe way.

Args:

• fn: A weave.op() decorated function. Returns: The Op of the function.

---

function call

call(
    op: 'Op',
    *args: 'Any',
    __weave: 'WeaveKwargs | None' = None,
    __should_raise: 'bool' = False,
    __require_explicit_finish: 'bool' = False,
    **kwargs: 'Any'
) → tuple[Any, Call] | Coroutine[Any, Any, tuple[Any, Call]]

Executes the op and returns both the result and a Call representing the execution.

This function will never raise. Any errors are captured in the Call object.

This method is automatically bound to any function decorated with @weave.op, allowing for usage like:

@weave.op
def add(a: int, b: int) -> int:
     return a + b

result, call = add.call(1, 2)

---

function calls

calls(op: 'Op') → CallsIter

Get an iterator over all calls to this op.

This method is automatically bound to any function decorated with @weave.op, allowing for usage like:

@weave.op
def add(a: int, b: int) -> int:
     return a + b

calls = add.calls()
for call in calls:
     print(call)

---

function get_captured_code

get_captured_code(op: 'Op') → str

Get the captured code of the op.

This only works when you get an op back from a ref. The pattern is:

ref = weave.publish(func) op = ref.get() captured_code = op.get_captured_code()

---

function is_op

is_op(obj: 'Any') → TypeIs[Op]

Check if an object is an Op.

---

function is_placeholder_call

is_placeholder_call(call: 'Call') → TypeIs[NoOpCall]

---

function is_tracing_setting_disabled

is_tracing_setting_disabled() → bool

---

function maybe_bind_method

maybe_bind_method(func: 'Callable', self: 'Any' = None) → Callable | MethodType

Bind a function to any object (even if it’s not a class).

If self is None, return the function as is.

---

function maybe_unbind_method

maybe_unbind_method(oplike: 'Op | MethodType | partial') → Op

Unbind an Op-like method or partial to a plain Op function.

For:

• methods, remove set self param
• partials, remove any preset params

---

function op

op(
    func: 'Callable[P, R] | None' = None,
    name: 'str | None' = None,
    call_display_name: 'str | CallDisplayNameFunc | None' = None,
    postprocess_inputs: 'PostprocessInputsFunc | None' = None,
    postprocess_output: 'PostprocessOutputFunc | None' = None,
    tracing_sample_rate: 'float' = 1.0,
    enable_code_capture: 'bool' = True,
    accumulator: 'Callable[[Any | None, Any], Any] | None' = None
) → Callable[[Callable[P, R]], Op[P, R]] | Op[P, R]

A decorator to weave op-ify a function or method. Works for both sync and async. Automatically detects iterator functions and applies appropriate behavior.

---

function placeholder_call

placeholder_call() → Call

---

function setup_dunder_weave_dict

setup_dunder_weave_dict(d: 'WeaveKwargs | None' = None) → WeaveKwargs

Sets up a __weave dict used to pass WeaveKwargs to ops.

---

function should_skip_tracing_for_op

should_skip_tracing_for_op(op: 'Op') → bool