---
title: AnalysisStage (latest version)
description: API reference for qiskit_noise_learning.analysis.AnalysisStage in the latest version of qiskit-noise-learning
source: https://quantum.cloud.ibm.com/docs/en/api/qiskit-noise-learning/generated/analysis-analysis-stage
---

# qiskit\_noise\_learning.analysis.AnalysisStage

*class* `qiskit_noise_learning.analysis.AnalysisStage`

[GitHub](https://github.com/Qiskit/qiskit-noise-learning/tree/stable/0.1/qiskit_noise_learning/analysis/analysis_pipeline.py)

Bases: [`ABC`](https://docs.python.org/3/library/abc.html#abc.ABC)

Abstract base for a stage in the analysis pipeline.

Each stage declares the data level it consumes ([`input_level`](#qiskit_noise_learning.analysis.AnalysisStage.input_level "qiskit_noise_learning.analysis.AnalysisStage.input_level")) and produces ([`output_level`](#qiskit_noise_learning.analysis.AnalysisStage.output_level "qiskit_noise_learning.analysis.AnalysisStage.output_level")). Stages may skip intermediate levels, e.g. going directly from `RawData` to `DecayData`.

To implement a stage, subclass this and override `_run()`. The public [`run()`](#qiskit_noise_learning.analysis.AnalysisStage.run "qiskit_noise_learning.analysis.AnalysisStage.run") method handles shallow-copying the [`Fit`](/docs/api/qiskit-noise-learning/generated/analysis-fit "qiskit_noise_learning.analysis.Fit") container and marking skipped levels; `_run()` receives the copy and may mutate it in place.

[`input_level`](#qiskit_noise_learning.analysis.AnalysisStage.input_level "qiskit_noise_learning.analysis.AnalysisStage.input_level") and [`output_level`](#qiskit_noise_learning.analysis.AnalysisStage.output_level "qiskit_noise_learning.analysis.AnalysisStage.output_level") can be declared as class attributes:

```python
class MyStage(AnalysisStage):
    input_level = RawData
    output_level = ObservableData

    def _run(self, fit):
        fit[ObservableData] = compute(fit[RawData])
```

### \_\_init\_\_

`__init__()`

## Methods

|                                                                                                                                |                                                                                                                                                                       |
| ------------------------------------------------------------------------------------------------------------------------------ | --------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| [`__init__`](#qiskit_noise_learning.analysis.AnalysisStage.__init__ "qiskit_noise_learning.analysis.AnalysisStage.__init__")() |                                                                                                                                                                       |
| [`run`](#qiskit_noise_learning.analysis.AnalysisStage.run "qiskit_noise_learning.analysis.AnalysisStage.run")(fit)             | Run this stage, returning a new [`Fit`](/docs/api/qiskit-noise-learning/generated/analysis-fit "qiskit_noise_learning.analysis.Fit") with the output level populated. |

## Attributes

|                                                                                                                                          |                                   |
| ---------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------- |
| [`input_level`](#qiskit_noise_learning.analysis.AnalysisStage.input_level "qiskit_noise_learning.analysis.AnalysisStage.input_level")    | The data level this stage reads.  |
| [`output_level`](#qiskit_noise_learning.analysis.AnalysisStage.output_level "qiskit_noise_learning.analysis.AnalysisStage.output_level") | The data level this stage writes. |

### input\_level

*abstract property*

Type: [`type`](https://docs.python.org/3/library/functions.html#type)\[[`LeveledData`](/docs/api/qiskit-noise-learning/generated/data-leveled-data "qiskit_noise_learning.data.leveled_data.LeveledData")]

The data level this stage reads.

### output\_level

*abstract property*

Type: [`type`](https://docs.python.org/3/library/functions.html#type)\[[`LeveledData`](/docs/api/qiskit-noise-learning/generated/data-leveled-data "qiskit_noise_learning.data.leveled_data.LeveledData")]

The data level this stage writes.

### run

`run(fit: Fit | LeveledData) → Fit`

Run this stage, returning a new [`Fit`](/docs/api/qiskit-noise-learning/generated/analysis-fit "qiskit_noise_learning.analysis.Fit") with the output level populated.

Shallow-copies `fit`, marks any `Absent` intermediate levels as `Skipped`, calls `_run()` on the copy, and returns it. The original `fit` is not modified.