Basic Usage¶

Here, we demonstrate how to filter conditions where we suspect the model to be undefined

In [1]:

import math

import numpy as np

from autora.experimentalist.prediction_filter import prediction_filter
import pandas as pd

import math import numpy as np from autora.experimentalist.prediction_filter import prediction_filter import pandas as pd

Create a model:

In [2]:

class ModelWithNans:
    def predict(self, x):
        try:
            return math.sqrt((x**2 - 4))
        except:
            return None
model = ModelWithNans()

class ModelWithNans: def predict(self, x): try: return math.sqrt((x**2 - 4)) except: return None model = ModelWithNans()

Let's create our pool:

In [3]:

pool = pd.DataFrame({'x': np.linspace(-5, 5, 11)})
pool

pool = pd.DataFrame({'x': np.linspace(-5, 5, 11)}) pool

Out[3]:

	x
0	-5.0
1	-4.0
2	-3.0
3	-2.0
4	-1.0
5	0.0
6	1.0
7	2.0
8	3.0
9	4.0
10	5.0

Let's create our filter function. This is a function that returns True if a value should be included

In [4]:

def filter_function(x):
    return x is not None

def filter_function(x): return x is not None

Now, let's use our filter on the conditions:

In [5]:

filtered_conditions = prediction_filter(pool, model, filter_function)
filtered_conditions

filtered_conditions = prediction_filter(pool, model, filter_function) filtered_conditions

---------------------------------------------------------------------------
AxisError                                 Traceback (most recent call last)
Cell In[5], line 1
----> 1 filtered_conditions = prediction_filter(pool, model, filter_function)
      2 filtered_conditions

File ~/Documents/GitHub/AutoResearch/autora-experimentalist-prediction-filter/src/autora/experimentalist/prediction_filter/__init__.py:71, in filter(conditions, model, filter_function)
     16 """
     17 Filter conditions based on the expected outcome io the mdeol
     18 
   (...)
     68     0  1  1
     69 """
     70 _pred = model.predict(conditions)
---> 71 _filter = np.apply_along_axis(filter_function, 1, _pred)
     72 _filter = _filter.reshape(1, -1)
     74 new_conditions = conditions[list(_filter[0])]

File ~/Documents/GitHub/AutoResearch/autora-experimentalist-prediction-filter/venv/lib/python3.11/site-packages/numpy/lib/shape_base.py:361, in apply_along_axis(func1d, axis, arr, *args, **kwargs)
    359 arr = asanyarray(arr)
    360 nd = arr.ndim
--> 361 axis = normalize_axis_index(axis, nd)
    363 # arr, with the iteration axis at the end
    364 in_dims = list(range(nd))

AxisError: axis 1 is out of bounds for array of dimension 0

In [5]:

In [ ]: