GitChameleon: Evaluating AI Code Generation Against Python Library Version Incompatibilities
Paper • 2507.12367 • Published • 7
python_version stringclasses 3
values | library stringclasses 26
values | version stringlengths 1 6 | problem stringlengths 34 1.02k | starting_code stringlengths 23 1.55k | example_id stringlengths 1 3 | test stringlengths 66 5.96k | solution stringlengths 7 9.39k | type_of_change stringclasses 21
values | name_of_class_or_func stringlengths 0 63 | additional_dependencies stringclasses 31
values | docs listlengths 1 3 | functional unknown | webdev unknown | solution_api_call bool 1
class | api_calls listlengths 0 47 | release_date stringdate 2014-08-01 00:00:00 2024-01-01 00:00:00 | extra_dependencies stringclasses 3
values |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
3.7 | torch | 1.9.0 | Calculate the logarithm of the cumulative distribution function of the standard normal distribution using available functions. If not available in PyTorch, use another library. | import torch
def log_ndtr(input_tensor: torch.Tensor) -> torch.Tensor:
| 0 | from scipy.stats import norm
input_tensor = torch.linspace(-10, 10, steps=20)
expected_result = torch.tensor([-5.3231e+01, -4.3150e+01, -3.4164e+01, -2.6270e+01, -1.9462e+01,
-1.3734e+01, -9.0731e+00, -5.4610e+00, -2.8617e+00, -1.2062e+00,
-3.5572e-01, -5.8874e-02, -4.2585e-03, -1.1471e-04, -1.0854e-06,... | import numpy as np
from scipy.stats import norm
output = torch.from_numpy(norm.logcdf(input_tensor.numpy()))
return output | other library | log_ndtr | scipy==1.7.3 numpy==1.21.6 | [
"https://docs.scipy.org/doc/scipy-1.8.0/reference/stats.html"
] | 1 | 0 | true | [
"torch.from_numpy",
"input_tensor.numpy",
"scipy.stats.norm.logcdf"
] | 2021-06 | null |
3.7 | torch | 1.9.0 | Calculate the natural logarithm of the absolute value of the gamma function using PyTorch's special functions if available in this version, otherwise you may use another library. | import torch
def gamma_ln(input_tensor: torch.Tensor) -> torch.Tensor:
| 1 | input_tensor = torch.linspace(0, 10, steps=10)
expected_result = torch.Tensor([float('inf'),-0.0545,0.1092,1.0218,2.3770,4.0476,5.9637,8.0806,10.3675,12.8018])
assert torch.allclose(gamma_ln(input_tensor), expected_result, rtol=1e-3, atol=1e-3) | import numpy as np
from scipy.special import gammaln as scipy_gammaln
output = torch.from_numpy(scipy_gammaln(input_tensor.numpy()))
return output | other library | gammaln | scipy==1.7.3 numpy==1.21.6 | [
"https://docs.scipy.org/doc/scipy-1.8.0/reference/special.html"
] | 1 | 0 | true | [
"torch.from_numpy",
"input_tensor.numpy",
"scipy.special.gammaln"
] | 2021-06 | null |
3.7 | torch | 1.9.0 | Calculate the error function using PyTorch's special functions if available in this version, otherwise you may use another library. | import torch
def erf(input_tensor: torch.Tensor) -> torch.Tensor:
| 2 | input_tensor = torch.linspace(0, 10, steps=10)
expected_result = torch.Tensor([0.0000,0.8839,0.9983,1.0000,1.0000,1.0000,1.0000,1.0000,1.0000,1.0000])
assert torch.allclose(erf(input_tensor), expected_result, rtol=1e-3, atol=1e-3) | import numpy as np
from scipy.special import erf as scipy_erf
output = torch.from_numpy(scipy_erf(input_tensor.numpy()))
return output | other library | erf | scipy==1.7.3 numpy==1.21.6 | [
"https://docs.scipy.org/doc/scipy-1.8.0/reference/special.html"
] | 1 | 0 | true | [
"torch.from_numpy",
"input_tensor.numpy",
"scipy.special.erf"
] | 2021-06 | null |
3.7 | torch | 1.9.0 | Calculate the complementary error function using PyTorch's special functions if available in this version, otherwise you may use another library. | import torch
def erfc(input_tensor: torch.Tensor) -> torch.Tensor:
| 3 | input_tensor = torch.linspace(0, 10, steps=10)
expected_result = torch.Tensor([1.0000e+00,1.1610e-01,1.6740e-03,2.4285e-06,3.2702e-10,3.9425e-15,4.1762e-21,3.8452e-28,3.0566e-36,1.4013e-45])
assert torch.allclose(erfc(input_tensor), expected_result, rtol=1e-3, atol=1e-3) | import numpy as np
from scipy.special import erfc as scipy_erfc
output = torch.from_numpy(scipy_erfc(input_tensor.numpy()))
return output | other library | erfc | scipy==1.7.3 numpy==1.21.6 | [
"https://docs.scipy.org/doc/scipy-1.8.0/reference/special.html"
] | 1 | 0 | true | [
"torch.from_numpy",
"input_tensor.numpy",
"scipy.special.erfc"
] | 2021-06 | null |
3.7 | torch | 1.9.0 | Calculate the modified Bessel function of the first kind, order 0 using PyTorch's special functions if available in this version, otherwise you may use another library. | import torch
def bessel_i0(input_tensor: torch.Tensor) -> torch.Tensor:
| 4 | input_tensor = torch.linspace(0, 10, steps=10)
expected_result = torch.Tensor([1.0000e+00,1.3333e+00,2.6721e+00,6.4180e+00,1.6648e+01,4.4894e+01,1.2392e+02,3.4740e+02,9.8488e+02,2.8157e+03])
assert torch.allclose(bessel_i0(input_tensor), expected_result, rtol=1e-3, atol=1e-3) | import numpy as np
from scipy.special import i0 as scipy_i0
output = torch.from_numpy(scipy_i0(input_tensor.numpy()))
return output | other library | bessel_i0 | scipy==1.7.3 numpy==1.21.6 | [
"https://docs.scipy.org/doc/scipy-1.8.0/reference/special.html"
] | 1 | 0 | true | [
"scipy.special.i0",
"torch.from_numpy",
"input_tensor.numpy"
] | 2021-06 | null |
3.7 | torch | 1.9.0 | Calculate the modified Bessel function of the first kind, order 1 using PyTorch's special functions if available in this version, otherwise you may use another library. | import torch
def bessel_i1(input_tensor: torch.Tensor) -> torch.Tensor:
| 5 | input_tensor = torch.linspace(0, 10, steps=10)
expected_result = torch.Tensor([0.0000e+00,6.4581e-01,1.9536e+00,5.3391e+00,1.4628e+01,4.0623e+01,1.1420e+02,3.2423e+02,9.2770e+02,2.6710e+03])
assert torch.allclose(bessel_i1(input_tensor), expected_result, rtol=1e-3, atol=1e-3) | import numpy as np
from scipy.special import i1 as scipy_i1
output = torch.from_numpy(scipy_i1(input_tensor.numpy()))
return output | other library | bessel_i1 | scipy==1.7.3 numpy==1.21.6 | [
"https://docs.scipy.org/doc/scipy-1.8.0/reference/special.html"
] | 1 | 0 | true | [
"torch.from_numpy",
"input_tensor.numpy",
"scipy.special.i1"
] | 2021-06 | null |
3.7 | torch | 1.10.0 | Calculate the natural logarithm of the absolute value of the gamma function using pytorch's special functions if available in this version, otherwise you may use another library. | import torch
def gamma_ln(input_tensor: torch.Tensor) -> torch.Tensor:
| 6 | input_tensor = torch.linspace(0, 10, steps=10)
expected_result = torch.Tensor([torch.inf,-0.0545,0.1092,1.0218,2.3770,4.0476,5.9637,8.0806,10.3675,12.8018])
assert torch.allclose(gamma_ln(input_tensor), expected_result, rtol=1e-3, atol=1e-3) | return torch.special.gammaln(input_tensor) | new func/method/class | gammaln | [
"https://pytorch.org/docs/1.10/special.html"
] | 1 | 0 | true | [
"torch.special.gammaln"
] | 2021-10 | null | |
3.7 | torch | 1.10.0 | Calculate the error function using pytorch's special functions if available in this version, otherwise you may use another library. | import torch
def erf(input_tensor: torch.Tensor) -> torch.Tensor:
| 7 | input_tensor = torch.linspace(0, 10, steps=10)
expected_result = torch.Tensor([0.0000,0.8839,0.9983,1.0000,1.0000,1.0000,1.0000,1.0000,1.0000,1.0000])
assert torch.allclose(erf(input_tensor), expected_result, rtol=1e-3, atol=1e-3) | return torch.special.erf(input_tensor) | new func/method/class | erf | [
"https://pytorch.org/docs/1.10/special.html"
] | 1 | 0 | true | [
"torch.special.erf"
] | 2021-10 | null | |
3.7 | torch | 1.10.0 | Calculate the complementary error function using pytorch's special functions if available in this version, otherwise you may use another library. | import torch
def erfc(input_tensor: torch.Tensor) -> torch.Tensor:
| 8 | input_tensor = torch.linspace(0, 10, steps=10)
expected_result = torch.Tensor([1.0000e+00,1.1610e-01,1.6740e-03,2.4285e-06,3.2702e-10,3.9425e-15,4.1762e-21,3.8452e-28,3.0566e-36,1.4013e-45])
assert torch.allclose(erfc(input_tensor), expected_result, rtol=1e-3, atol=1e-3) | return torch.special.erfc(input_tensor) | new func/method/class | erfc | [
"https://pytorch.org/docs/1.10/special.html"
] | 1 | 0 | true | [
"torch.special.erfc"
] | 2021-10 | null | |
3.7 | torch | 1.10.0 | Calculate the modified Bessel function of the first kind, order 0 using pytorch's special functions if available in this version, otherwise you may use another library. | import torch
def bessel_i0(input_tensor: torch.Tensor) -> torch.Tensor:
| 9 | input_tensor = torch.linspace(0, 10, steps=10)
expected_result = torch.Tensor([1.0000e+00,1.3333e+00,2.6721e+00,6.4180e+00,1.6648e+01,4.4894e+01,1.2392e+02,3.4740e+02,9.8488e+02,2.8157e+03])
assert torch.allclose(bessel_i0(input_tensor), expected_result, rtol=1e-3, atol=1e-3) | return torch.special.i0(input_tensor) | new func/method/class | bessel_i0 | [
"https://pytorch.org/docs/1.10/special.html"
] | 1 | 0 | true | [
"torch.special.i0"
] | 2021-10 | null | |
3.7 | torch | 1.10.0 | Calculate the modified Bessel function of the first kind, order 1 using pytorch's special functions if available in this version, otherwise you may use another library. | import torch
def bessel_i1(input_tensor: torch.Tensor) -> torch.Tensor:
| 10 | input_tensor = torch.linspace(0, 10, steps=10)
expected_result = torch.Tensor([0.0000e+00,6.4581e-01,1.9536e+00,5.3391e+00,1.4628e+01,4.0623e+01,1.1420e+02,3.2423e+02,9.2770e+02,2.6710e+03])
assert torch.allclose(bessel_i1(input_tensor), expected_result, rtol=1e-3, atol=1e-3) | return torch.special.i1(input_tensor) | new func/method/class | bessel_i1 | [
"https://pytorch.org/docs/1.10/special.html"
] | 1 | 0 | true | [
"torch.special.i1"
] | 2021-10 | null | |
3.7 | torch | 1.10.0 | You are given two tensors, `tensor1` and `tensor2`, both of shape `(n,)`. Your task is to create a boolean mask indicating whether each element of `tensor1` is less than the corresponding element of `tensor2`, and then invert this mask. Store the answer in a variable named mask. | import torch
def invert_mask(tensor1: torch.Tensor, tensor2: torch.Tensor) -> torch.BoolTensor:
| 11 | tensor1 = torch.Tensor([1, 2, 3])
tensor2 = torch.Tensor([3, 1, 2])
expected_mask=torch.Tensor([False, True, True])
assert torch.all(torch.eq(invert_mask(tensor1, tensor2), expected_mask)) | return ~(tensor1 < tensor2) | output behaviour | invert_mask_v1_1 | [
"https://pytorch.org/docs/1.10/tensors.html"
] | 1 | 0 | true | [] | 2021-10 | null | |
3.10 | torch | 1.12.0 | Calculate the logarithm of the cumulative distribution function of the standard normal distribution using PyTorch's special functions. | import torch
def log_ndtr(input_tensor: torch.Tensor) -> torch.Tensor:
| 12 | input_tensor = torch.linspace(-10, 10, steps=20)
expected_result = torch.tensor([-5.3231e+01, -4.3150e+01, -3.4164e+01, -2.6270e+01, -1.9462e+01,
-1.3734e+01, -9.0731e+00, -5.4610e+00, -2.8617e+00, -1.2062e+00,
-3.5572e-01, -5.8874e-02, -4.2585e-03, -1.1471e-04, -1.0854e-06,
-3.5303e-09, -3.9019... | return torch.special.log_ndtr(input_tensor) | new func/method/class | log_ndtr | [
"https://pytorch.org/docs/1.12/distributions.html"
] | 1 | 0 | true | [
"torch.special.log_ndtr"
] | 2022-06 | null | |
3.10 | torch | 1.13 | You are given two tensors, `tensor1` and `tensor2`, both of shape `(n,)`. Your task is to create a boolean mask indicating whether each element of `tensor1` is less than the corresponding element of `tensor2`, and then invert this mask. Store the answer in a variable named mask. | import torch
def invert_mask(tensor1: torch.Tensor, tensor2: torch.Tensor) -> torch.BoolTensor:
| 13 | tensor1 = torch.Tensor([1, 2, 3])
tensor2 = torch.Tensor([3, 1, 2])
expected_mask=torch.Tensor([False, True, True])
assert torch.all(torch.eq(invert_mask(tensor1, tensor2), expected_mask)) | return ~(tensor1 < tensor2).bool() | output behaviour | invert_mask_v1_2 | [
"https://pytorch.org/docs/1.13/tensors.html"
] | 1 | 0 | true | [
"bool"
] | 2022-10 | null | |
3.10 | torch | 1.13 | You are given an audio signal represented as a 1D tensor `audio_signal`. Your task is to compute the Short-Time Fourier Transform (STFT) of the signal. Do not return a complex data type. | import torch
def stft(audio_signal: torch.Tensor, n_fft: int) -> torch.Tensor:
| 14 | audio_signal = torch.rand(1024)
n_fft=128
expected_shape = (65, 33, 2)
assert stft(audio_signal, n_fft).shape == expected_shape | return torch.stft(audio_signal, n_fft=n_fft, return_complex=False) | argument change | torch.stft | [
"https://pytorch.org/docs/1.13/torch.html"
] | 1 | 0 | true | [
"torch.stft"
] | 2022-10 | null | |
3.10 | torch | 2 | You are given an audio signal represented as a 1D tensor `audio_signal`. Your task is to compute the Short-Time Fourier Transform (STFT) of the signal. Do not return a complex data type. | import torch
def stft(audio_signal: torch.Tensor, n_fft: int) -> torch.Tensor:
| 15 | audio_signal = torch.rand(1024)
n_fft=128
expected_shape = (65, 33, 2)
assert stft(audio_signal, n_fft).shape == expected_shape | return torch.view_as_real(torch.stft(audio_signal, n_fft=n_fft, return_complex=True)) | argument change | torch.stft | [
"https://pytorch.org/docs/2.0/torch.html"
] | 1 | 0 | true | [
"torch.view_as_real",
"torch.stft"
] | 2023-05 | null | |
3.10 | torch | 1.13 | You are given a spectrogram represented as a 3D tensor `spectrogram` with dimensions `(65, 33, 2)`, where the first dimension represents the frequency bins, the second dimension represents the time frames, and the third dimension represents the real and imaginary parts of the complex values. Your task is to compute the... | import torch
def istft(spectrogram: torch.Tensor, signal: torch.Tensor, n_fft: int, hop_length: int, win_length: int, normalized=False) -> torch.Tensor:
| 16 |
# Sample rate (samples per second)
fs = 8000
# Duration of the signal in seconds
t = 1
# Time axis for the signal
time = torch.linspace(0, t, steps=int(fs * t))
# Frequency of the sine wave in Hz
frequency = 440
# Generate a sine wave
signal = torch.sin(2 * torch.pi * frequency * time)
n_fft = 1024 # Number of ... | return torch.istft(spectrogram, n_fft=n_fft, hop_length=hop_length, win_length=win_length, window=torch.hann_window(win_length), length=signal.shape[0], normalized=False) | argument change | torch.istft | [
"https://pytorch.org/docs/1.13/torch.html"
] | 1 | 0 | true | [
"torch.hann_window",
"torch.istft"
] | 2022-10 | null | |
3.10 | torch | 2 | You are given a spectrogram represented as a 3D tensor `spectrogram` with dimensions `(65, 33, 2)`, where the first dimension represents the frequency bins, the second dimension represents the time frames, and the third dimension represents the real and imaginary parts of the complex values. Your task is to compute the... | import torch
def istft(spectrogram: torch.Tensor, signal: torch.Tensor, n_fft: int, hop_length: int, win_length: int, normalized=False) -> torch.Tensor:
| 17 |
# Sample rate (samples per second)
fs = 8000
# Duration of the signal in seconds
t = 1
# Time axis for the signal
time = torch.linspace(0, t, steps=int(fs * t))
# Frequency of the sine wave in Hz
frequency = 440
# Generate a sine wave
signal = torch.sin(2 * torch.pi * frequency * time)
n_fft = 1024 # Number of ... | return torch.istft(torch.view_as_complex(spectrogram), n_fft=n_fft, hop_length=hop_length, win_length=win_length, window=torch.hann_window(win_length), length=signal.shape[0], normalized=False) | argument change | torch.istft | [
"https://pytorch.org/docs/2.0/torch.html"
] | 1 | 0 | true | [
"torch.hann_window",
"torch.istft",
"torch.view_as_complex"
] | 2023-05 | null | |
3.10 | geopandas | 0.10.0 | Write a function that performs a spatial join. | import geopandas as gpd
from shapely.geometry import Point, Polygon
def spatial_join(gdf1 : gpd.GeoDataFrame, gdf2 : gpd.GeoDataFrame) -> gpd.GeoDataFrame:
return | 18 |
gdf1 = gpd.GeoDataFrame({'geometry': [Point(1, 1), Point(2, 2), Point(3, 3)]})
polygons = [Polygon([(0, 0), (0, 4), (4, 4), (4, 0)]), Polygon([(4, 4), (4, 8), (8, 8), (8, 4)])]
gdf2 = gpd.GeoDataFrame({'geometry': polygons})
expected = gpd.GeoDataFrame({
'geometry': [Point(1, 1), Point(2, 2), Point(3, 3)],
'in... | gpd.sjoin(gdf1, gdf2, predicate='within') | name change | sjoin | rtree==0.9.3 | [
"https://geopandas.org/en/v0.10.0/docs/reference/api/geopandas.GeoDataFrame.html"
] | 1 | 0 | true | [
"geopandas.sjoin"
] | 2021-10 | null |
3.10 | geopandas | 0.9.0 | Write a function that performs a spatial join. | import geopandas as gpd
from shapely.geometry import Point, Polygon
def spatial_join(gdf1 : gpd.GeoDataFrame, gdf2 : gpd.GeoDataFrame) -> gpd.GeoDataFrame:
return | 19 |
gdf1 = gpd.GeoDataFrame({'geometry': [Point(1, 1), Point(2, 2), Point(3, 3)]})
polygons = [Polygon([(0, 0), (0, 4), (4, 4), (4, 0)]), Polygon([(4, 4), (4, 8), (8, 8), (8, 4)])]
gdf2 = gpd.GeoDataFrame({'geometry': polygons})
expected_result = gpd.GeoDataFrame({
'geometry': [Point(1, 1), Point(2, 2), Point(3, 3)],
... | gpd.sjoin(gdf1, gdf2, op='within') | name change | sjoin | rtree==0.9.3 shapely==1.8.5 | [
"https://geopandas.org/en/v0.9.0/docs/reference/api/geopandas.GeoDataFrame.html"
] | 1 | 0 | true | [
"geopandas.sjoin"
] | 2021-02 | null |
3.10 | geopandas | 0.10.0 | Write a function that performs a union. | import geopandas as gpd
from shapely.geometry import box
def perform_union(gdf : gpd.GeoDataFrame) -> gpd.GeoSeries:
return | 20 |
gdf = gpd.GeoDataFrame({'geometry': [box(0, 0, 2, 5), box(0, 0, 2, 1)]})
from shapely.geometry import Polygon
coords = [
(2, 0),
(0, 0),
(0, 1),
(0, 5),
(2, 5),
(2, 1),
(2, 0)
]
expected_result = Polygon(coords)
assert perform_union(gdf).equals(expected_result) | gdf.geometry.unary_union | name change | cascaded_union | [
"https://geopandas.org/en/v0.10.0/docs/reference/api/geopandas.GeoDataFrame.html"
] | 1 | 0 | true | [] | 2021-10 | null | |
3.10 | geopandas | 0.9.0 | Write a function that performs a union. | import geopandas as gpd
from shapely.geometry import box
def perform_union(gdf: gpd.GeoDataFrame) -> gpd.GeoSeries:
return | 21 |
gdf = gpd.GeoDataFrame({'geometry': [box(0, 0, 2, 5), box(0, 0, 2, 1)]})
from shapely.geometry import Polygon
coords = [
(2, 0),
(0, 0),
(0, 1),
(0, 5),
(2, 5),
(2, 1),
(2, 0)
]
expected_result = Polygon(coords)
assert perform_union(gdf) == expected_result
| gdf.geometry.cascaded_union | name change | cascaded_union | shapely==1.8.5 | [
"https://geopandas.org/en/v0.9.0/docs/reference/api/geopandas.GeoDataFrame.html"
] | 1 | 0 | true | [] | 2021-02 | null |
3.10 | geopandas | 0.10.0 | Write a function that creates a GeoSeries from x and y coordinates. | import geopandas as gpd
def create_geoseries(x: list[int], y: list[int]) -> gpd.GeoSeries:
return | 22 |
from shapely.geometry import Point
x, y = [1, 2], [3, 4]
expected_result = gpd.GeoSeries([Point(1, 3), Point(2, 4)])
assert list(create_geoseries(x, y)) == list(expected_result)
| gpd.GeoSeries.from_xy(x, y) | name change | points_from_xy | [
"https://geopandas.org/en/v0.10.0/docs/reference/api/geopandas.GeoSeries.html"
] | 1 | 0 | true | [
"geopandas.GeoSeries.from_xy"
] | 2021-10 | null | |
3.10 | geopandas | 0.9.0 | Write a function that creates a GeoSeries from x and y coordinates. | import geopandas as gpd
def create_geoseries(x:list[int], y:list[int]) -> gpd.GeoSeries:
return | 23 |
from shapely.geometry import Point
x, y = [1, 2], [3, 4]
expected_result = gpd.GeoSeries([Point(1, 3), Point(2, 4)])
assert list(create_geoseries(x, y)) == list(expected_result)
| gpd.points_from_xy(x, y) | name change | points_from_xy | shapely==1.8.5 | [
"https://geopandas.org/en/v0.9.0/docs/reference/api/geopandas.GeoSeries.html"
] | 1 | 0 | true | [
"geopandas.points_from_xy"
] | 2021-02 | null |
3.10 | geopandas | 0.13.0 | Write a function that performs a spatial query. | import geopandas as gpd
from shapely.geometry import Point, Polygon, box
def spatial_query(gdf:gpd.GeoDataFrame, other:gpd.GeoDataFrame) -> gpd.GeoDataFrame:
combined_geometry = other.unary_union
return | 24 |
gdf = gpd.GeoDataFrame({'geometry': [Point(1, 2)]})
other = gpd.GeoDataFrame({'geometry': [Point(1,1)]})
result = spatial_query(gdf, other)
expected_result = []
assert (result == expected_result).all() | gdf.sindex.query(combined_geometry) | name change | query_bulk | rtree==1.0.1 | [
"https://geopandas.org/en/v0.13.0/docs/reference/api/geopandas.GeoDataFrame.html"
] | 1 | 0 | true | [
"gdf.sindex.query"
] | 2023-05 | null |
3.10 | geopandas | 0.10.0 | Write a function that performs a spatial query. | import geopandas as gpd
from shapely.geometry import Point, Polygon
def spatial_query(gdf:gpd.GeoDataFrame, other:gpd.GeoSeries) -> gpd.GeoDataFrame:
return | 25 |
gdf = gpd.GeoDataFrame({'geometry': [Point(1, 1), Point(2, 2), Point(3, 3)]})
other = gpd.GeoSeries([Polygon([(0, 0), (0, 4), (4, 4), (4, 0)])])
result = spatial_query(gdf, other)
import numpy as np
expected_result = np.array([
[0, 0, 0],
[0, 1, 2]
])
assert (result == expected_result).all() | gdf.sindex.query_bulk(other) | name change | query_bulk | rtree==0.9.3 | [
"https://geopandas.org/en/v0.10.0/docs/reference/api/geopandas.GeoDataFrame.html"
] | 1 | 0 | true | [
"gdf.sindex.query_bulk"
] | 2021-10 | null |
3.10 | nltk | 3.6.3 | Write a function that displays usage information of an object. | import nltk
import io
import contextlib
def show_usage(obj:object) -> str:
with io.StringIO() as buf, contextlib.redirect_stdout(buf): | 26 |
assert "LazyModule supports the following operations" in show_usage(nltk.corpus)
|
nltk.usage(obj)
return buf.getvalue() | deprecation | usage | [
"https://www.nltk.org/api/nltk.html"
] | 1 | 0 | true | [
"buf.getvalue",
"nltk.usage"
] | 2021-09 | null | |
3.10 | networkx | 2.8 | Write a function that uses NetworkX's greedy_modularity_communities with the number of communities set at 5. | import networkx as nx
def modularity_communities(G:nx.Graph) -> list:
return nx.community.greedy_modularity_communities(G, | 27 | G = nx.karate_club_graph()
result = [
frozenset({8, 14, 15, 18, 20, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33}),
frozenset({1, 2, 3, 7, 9, 12, 13, 17, 21}),
frozenset({0, 16, 4, 5, 6, 10, 11}),
frozenset({19}),
frozenset({22})
]
assertion_value = len(modularity_communities(G)) > 0 and len(modularit... | cutoff=5) | argument change | [
"https://networkx.org/documentation/networkx-2.8/reference/algorithms/generated/networkx.algorithms.community.modularity_max.greedy_modularity_communities.html#networkx.algorithms.community.modularity_max.greedy_modularity_communities"
] | 1 | 0 | true | [] | 2022-04 | null | ||
3.10 | networkx | 2.7 | Write a function that uses NetworkX's greedy_modularity_communities with the number of communities set at 5. | import networkx as nx
def modularity_communities(G:nx.Graph) -> list:
return nx.community.greedy_modularity_communities(G, | 28 | G = nx.karate_club_graph()
result = [
frozenset({8, 14, 15, 18, 20, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33}),
frozenset({1, 2, 3, 7, 9, 12, 13, 17, 21}),
frozenset({0, 16, 4, 5, 6, 10, 11}),
frozenset({19}),
frozenset({22})
]
assertion_value = len(modularity_communities(G)) > 0 and len(modularit... | n_communities=5) | argument change | numpy==1.21.2 | [
"https://networkx.org/documentation/networkx-2.7/reference/algorithms/generated/networkx.algorithms.community.modularity_max.greedy_modularity_communities.html#networkx.algorithms.community.modularity_max.greedy_modularity_communities"
] | 1 | 0 | true | [] | 2022-02 | null | |
3.10 | networkx | 2.8 | Write a function that calculates the diameters' extreme distance of a graph. | import networkx as nx
def bounding_distance(G:nx.Graph) -> int:
return nx.diameter | 29 | G = nx.path_graph(5)
result = 4
assertion_value = bounding_distance(G) is not None and result == bounding_distance(G)
assert assertion_value | (G, usebounds=True) | name change | [
"https://networkx.org/documentation/networkx-2.8/reference/algorithms/generated/networkx.algorithms.distance_measures.diameter.html"
] | 1 | 0 | true | [] | 2022-04 | null | ||
3.10 | networkx | 2.6 | Write a function that calculates the diameters' extreme distance of a graph. | import networkx as nx
def bounding_distance(G:nx.Graph) -> int:
return nx.algorithms.distance_measures. | 30 |
G = nx.path_graph(5)
result = 4
assert bounding_distance(G) is not None and result == bounding_distance(G) | extrema_bounding(G, "diameter") | name change | [
"https://networkx.org/documentation/networkx-2.6/reference/algorithms/distance_measures.html"
] | 1 | 0 | true | [
"extrema_bounding"
] | 2021-07 | null | ||
3.10 | networkx | 2.5 | Write a function that returns the naive greedy modularity communities for a graph. | import networkx as nx
def naive_modularity_communities(G:nx.Graph) -> list:
return nx.community. | 31 | G = nx.karate_club_graph()
naive_modularity_communities_result = 3
assert len(list(naive_modularity_communities(G))) > 0 and len(list(naive_modularity_communities(G))) == naive_modularity_communities_result | naive_greedy_modularity_communities(G) | name change | [
"https://networkx.org/documentation/networkx-2.5/reference/algorithms/generated/networkx.algorithms.community.modularity_max.naive_greedy_modularity_communities.html#networkx.algorithms.community.modularity_max.naive_greedy_modularity_communities"
] | 1 | 0 | true | [
"naive_greedy_modularity_communities"
] | 2019-10 | null | ||
3.10 | networkx | 2.4 | Write a function that returns the naive greedy modularity communities for a graph. | import networkx as nx
def naive_modularity_communities(G:nx.Graph) -> list:
return nx.community. | 32 | G = nx.karate_club_graph()
naive_modularity_communities_result = 3
assert len(list(naive_modularity_communities(G))) > 0 and len(list(naive_modularity_communities(G))) == naive_modularity_communities_result | _naive_greedy_modularity_communities(G) | name change | [
"https://networkx.org/documentation/networkx-2.4/reference/algorithms/generated/networkx.algorithms.community.modularity_max.naive_greedy_modularity_communities.html#networkx.algorithms.community.modularity_max.naive_greedy_modularity_communities"
] | 1 | 0 | true | [
"_naive_greedy_modularity_communities"
] | 2020-08 | null | ||
3.10 | networkx | 2.5 | Write a function that returns the nodes as a list of NetworkX graph. | import networkx as nx
def get_nodes(G:nx.Graph) -> list:
return | 33 |
G = nx.karate_club_graph()
nodes_result = 34
assert get_nodes(G) is not None and len(get_nodes(G)) > 0 and len(get_nodes(G)) == nodes_result | list(G.nodes) | name change (attribute) | [
"https://networkx.org/documentation/networkx-2.5/reference/classes/graph.html"
] | 1 | 0 | true | [
"list"
] | 2019-10 | null | ||
3.10 | networkx | 2.5 | Write a function that accesses the first edge of a NetworkX graph. | import networkx as nx
def get_first_edge(G:nx.Graph) -> tuple :
return | 34 |
G = nx.karate_club_graph()
first_edge_result = (0, 1)
assert get_first_edge(G) is not None and first_edge_result == get_first_edge(G) | list(G.edges)[0] | name change | [
"https://networkx.org/documentation/networkx-2.5/reference/classes/graph.html"
] | 1 | 0 | true | [
"list"
] | 2019-10 | null | ||
3.10 | networkx | 2.5 | Write a function that computes the shortest path lengths and predecessors on shortest paths in weighted graphs using NetworkX. | import networkx as nx
def shortest_path(G:nx.Graph, source:int) -> list:
return nx. | 35 |
G = nx.path_graph(5)
shortest_path_result = nx.bellman_ford_predecessor_and_distance(G, 0)
assert shortest_path(G, 0) is not None and len(shortest_path(G, 0)) == 2 | bellman_ford_predecessor_and_distance(G, source) | name change | [
"https://networkx.org/documentation/networkx-2.5/reference/classes/graph.html"
] | 1 | 0 | true | [
"bellman_ford_predecessor_and_distance"
] | 2019-10 | null | ||
3.10 | gradio | 3.24.0 | Write a function that renders the quadratic formula in LaTeX using Gradio's Chatbot. The quadratic formula is given by: x = \frac{-b \pm \sqrt{b^2 - 4ac}}{2a} | import gradio as gr
def render_quadratic_formula():
pass
interface = gr.Interface(fn=render_quadratic_formula, inputs=[], outputs = "text")
def render_quadratic_formula():
formula = | 36 | assertion_value = render_quadratic_formula().startswith("$") and render_quadratic_formula().endswith("$")
assert assertion_value | "$x = \\frac{-b \\pm \\sqrt{b^2 - 4ac}}{2a}$"
return formula | argument change | - | [
"https://www.gradio.app/docs/gradio/chatbot",
"https://www.gradio.app/guides/the-interface-class",
"https://www.gradio.app/changelog"
] | 1 | 0 | true | [] | 2023-03 | null | |
3.10 | gradio | 3.36.0 | Write a function that renders the quadratic formula in LaTeX using Gradio's Chatbot. The quadratic formula is given by: x = \frac{-b \pm \sqrt{b^2 - 4ac}}{2a} | import gradio as gr
def render_quadratic_formula():
formula = "x = \\frac{-b \\pm \\sqrt{b^2 - 4ac}}{2a}"
return formula
interface = gr.Chatbot | 37 | assertion_value = not render_quadratic_formula().startswith("$") and not render_quadratic_formula().endswith("$") and "$" in interface.latex_delimiters[0] and "$" in interface.latex_delimiters[1]
assert assertion_value | (fn=render_quadratic_formula, latex_delimiters=("$$", "$$"))
| argument change | - | [
"https://www.gradio.app/docs/gradio/chatbot",
"https://www.gradio.app/changelog"
] | 1 | 0 | true | [] | 2023-07 | null | |
3.10 | gradio | 3.36.0 | Write a function that displays an image using Gradio where you cannot share the image. | import gradio as gr
def display_image():
return "https://image_placeholder.com/42"
iface = gr.Interface | 38 | assertion_value = iface.output_components[0].show_share_button==False
assert assertion_value | (fn=display_image, inputs=[], outputs=gr.Image(show_share_button=False))
| argument change | - | [
"https://www.gradio.app/docs/gradio/image",
"https://www.gradio.app/guides/the-interface-class",
"https://www.gradio.app/changelog"
] | 1 | 0 | true | [
"gradio.Image"
] | 2023-07 | null | |
3.10 | gradio | 3.0.0 | Write a function that displays an image using Gradio where you cannot share the image. | import gradio as gr
def display_image():
return "https://image_placeholder.com/42"
iface = gr.Interface | 39 | assertion_value = type(gr.Image()) == type(iface.output_components[0])
assert assertion_value | (fn=display_image, inputs=[], outputs=gr.Image())
| argument change | - | [
"https://www.gradio.app/docs/gradio/image",
"https://www.gradio.app/guides/the-interface-class",
"https://www.gradio.app/changelog"
] | 1 | 0 | true | [
"gradio.Image"
] | 2022-05 | null | |
3.10 | gradio | 2.9.2 | Write a function that takes an image input and returns a textbox output. | import gradio as gr
def process_image(image):
return "Processed"
iface = gr.Interface | 40 | assertion_value = type(iface.input_components[0])==type(gr.inputs.Image()) and type(iface.output_components[0])==type(gr.outputs.Textbox()) or type(iface.input_components[0])==type(gr.components.Image()) and type(iface.output_components[0])==type(gr.components.Textbox())
assert assertion_value | (fn=process_image, inputs=gr.inputs.Image(), outputs=gr.outputs.Textbox()) | argument change | - | black==22.1.0 | [
"https://www.gradio.app/docs/gradio/image",
"https://www.gradio.app/guides/the-interface-class",
"https://www.gradio.app/changelog"
] | 1 | 0 | true | [
"gradio.outputs.Textbox",
"gradio.inputs.Image"
] | 2022-04 | null |
3.10 | gradio | 3.24.0 | Write a function that takes an image input and returns a label output. | import gradio as gr
def process_image(image):
return "Processed"
iface = gr.Interface | 41 | assertion_value = type(iface.input_components[0])==type(gr.Image()) and type(iface.output_components[0])==type(gr.Label())
assert assertion_value | (fn=process_image, inputs=gr.Image(), outputs=gr.Label()) | argument change | - | [
"https://www.gradio.app/docs/gradio/image",
"https://www.gradio.app/guides/the-interface-class",
"https://www.gradio.app/changelog"
] | 1 | 0 | true | [
"gradio.Label",
"gradio.Image"
] | 2023-03 | null | |
3.10 | gradio | 3.17.0 | Write a function that returns the selected options from a list of options. Users can select multiple options. | import gradio as gr
def get_selected_options(options):
return f"Selected options: {options}"
selection_options = ["angola", "pakistan", "canada"]
iface = gr.Interface(get_selected_options, inputs =
| 42 | assertion_value = (type(iface.input_components[0]) == gr.Dropdown and iface.input_components[0].multiselect == True ) or type(iface.input_components[0]) == gr.CheckboxGroup
assert assertion_value | gr.Dropdown(selection_options, multiselect=True), outputs = 'text') | argument change | - | [
"https://www.gradio.app/4.44.1/docs/gradio/dropdown",
"https://www.gradio.app/guides/the-interface-class",
"https://www.gradio.app/changelog"
] | 1 | 0 | true | [
"gradio.Dropdown"
] | 2023-01 | null | |
3.10 | scikit-learn | 1.1 | Train a Gradient Boosting Classifier from scikit-learn for a binary classification task and get the number of features used in fit. | from sklearn.ensemble import GradientBoostingClassifier
import numpy as np
def get_n_features(clf: GradientBoostingClassifier) -> int:
| 43 | X = np.random.rand(100, 20) # 100 samples, 20 features
y = np.random.randint(0, 2, 100)
clf=GradientBoostingClassifier()
clf.fit(X,y)
expected_n_features=20
assert get_n_features(clf)== expected_n_features | n_features_used = clf.n_features_in_
return n_features_used | output behaviour | GradientBoostingClassifier | numpy==1.23.5 | [
"https://scikit-learn.org/1.1/modules/generated/sklearn.ensemble.GradientBoostingClassifier.html#sklearn.ensemble.GradientBoostingClassifier"
] | 1 | 0 | true | [] | 2022-05 | null |
3.10 | scikit-learn | 1.1 | You are tasked with developing a solution that uses Gradient Boosting Classifier from scikit-learn for a binary classification task with the mean squared error as the criterion. | from sklearn.ensemble import GradientBoostingClassifier
# Initialize the classifier
def init_clf() -> GradientBoostingClassifier:
classifier = GradientBoostingClassifier(criterion= | 44 | expected_clf=GradientBoostingClassifier
assert isinstance(init_clf(), expected_clf) | 'squared_error')
return classifier | argument change | GradientBoostingClassifier | numpy==1.23.5 | [
"https://scikit-learn.org/1.1/modules/generated/sklearn.ensemble.GradientBoostingClassifier.html#sklearn.ensemble.GradientBoostingClassifier"
] | 1 | 0 | true | [] | 2022-05 | null |
3.10 | scikit-learn | 1.2 | Given dummy data, determine the shape of the coef_ attribute of a CCA model fitted with this data. | from sklearn.cross_decomposition import CCA
import numpy as np
def get_coef_shape(cca_model: CCA, X: np.ndarray, Y: np.ndarray) -> tuple:
cca_model.fit(X, Y)
return | 45 | X = np.random.rand(100, 10)
Y = np.random.rand(100, 5)
cca_model = CCA()
correct_shape=(X.shape[1], Y.shape[1])
assert get_coef_shape(cca_model, X, Y) == correct_shape | cca_model.coef_.shape | attribute change | CCA | numpy==1.23.5 | [
"https://scikit-learn.org/1.2/modules/generated/sklearn.cross_decomposition.CCA.html#sklearn.cross_decomposition.CCA"
] | 1 | 0 | true | [] | 2022-12 | null |
3.10 | scikit-learn | 1.3 | Given dummy data, determine the shape of the coef_ attribute of a CCA model fitted with this data. | from sklearn.cross_decomposition import CCA
import numpy as np
def get_coef_shape(cca_model: CCA, X: np.ndarray, Y: np.ndarray) -> tuple:
cca_model.fit(X, Y)
return | 46 | X = np.random.rand(100, 10)
Y = np.random.rand(100, 5)
cca_model = CCA()
correct_shape=(Y.shape[1], X.shape[1])
assert get_coef_shape(cca_model, X, Y) == correct_shape | cca_model.coef_.shape | attribute change | CCA | numpy==1.23.5 | [
"https://scikit-learn.org/1.3/modules/generated/sklearn.cross_decomposition.CCA.html#sklearn.cross_decomposition.CCA"
] | 1 | 0 | true | [] | 2023-06 | null |
3.10 | scikit-learn | 1.1 | Generate a sparse coded signal where the data is transposed. | from sklearn.datasets import make_sparse_coded_signal
def get_signal(n_samples: int, n_features: int, n_components: int, n_nonzero_coefs: int) -> tuple:
return | 47 | n_samples=100
n_features=50
n_components=20
n_nonzero_coefs=10
expected_shape_y = (n_features, n_samples)
expected_shape_d = (n_features, n_components)
expected_shape_c = (n_components, n_samples)
y,d,c = get_signal(n_samples, n_features, n_components, n_nonzero_coefs)
assert y.shape == expected_shape_y
assert d.shape... | make_sparse_coded_signal(n_samples=n_samples, n_features=n_features,n_components=n_components,n_nonzero_coefs=n_nonzero_coefs) | output behaviour | make_sparse_coded_signal | numpy==1.23.5 | [
"https://scikit-learn.org/1.1/modules/generated/sklearn.datasets.make_sparse_coded_signal.html#sklearn.datasets.make_sparse_coded_signal"
] | 1 | 0 | true | [
"sklearn.datasets.make_sparse_coded_signal"
] | 2022-05 | null |
3.10 | scikit-learn | 1.1 | Apply Fast Independent Component Analysis (FastICA) with a specific whiten parameter setting. Store transformed data in a variable transformed_data. | from sklearn.datasets import load_digits
from sklearn.utils import Bunch
from sklearn.decomposition import FastICA
def apply_fast_ica(data: Bunch, n_components: int) -> FastICA:
return | 48 | data, _ = load_digits(return_X_y=True)
n_components=7
expected_shape = (1797, n_components)
assert apply_fast_ica(data, n_components).shape == expected_shape | FastICA(n_components=n_components,random_state=0,whiten=True).fit_transform(data) | argument change | FastICA | numpy==1.23.5 | [
"https://scikit-learn.org/1.1/modules/generated/sklearn.decomposition.FastICA.html#sklearn.decomposition.FastICA"
] | 1 | 0 | true | [
"fit_transform",
"sklearn.decomposition.FastICA"
] | 2022-05 | null |
3.10 | scikit-learn | 1.3 | Apply Fast Independent Component Analysis (FastICA) with a specific whiten parameter setting. Store transformed data in a variable transformed_data. | from sklearn.datasets import load_digits
from sklearn.decomposition import FastICA
from sklearn.utils import Bunch
def apply_fast_ica(data: Bunch, n_components: int) -> FastICA:
return | 49 | data, _ = load_digits(return_X_y=True)
n_components=7
expected_shape = (1797, n_components)
assert apply_fast_ica(data, n_components).shape == expected_shape | FastICA(n_components=n_components,random_state=0,whiten='arbitrary-variance').fit_transform(data) | argument change | FastICA | numpy==1.23.5 | [
"https://scikit-learn.org/1.3/modules/generated/sklearn.decomposition.FastICA.html#sklearn.decomposition.FastICA"
] | 1 | 0 | true | [
"fit_transform",
"sklearn.decomposition.FastICA"
] | 2023-06 | null |
3.10 | scikit-learn | 1.1 | Impute missing values in a dataset using SimpleImputer and return the SimpleImputer instance, including the `verbose` parameter if available. | from sklearn.impute import SimpleImputer
import numpy as np
def get_imputer(data: np.ndarray) -> SimpleImputer:
return | 50 | data = np.array([[1, 2, 3], [4, None, 6], [7, 8, None]], dtype=float)
expected_type=SimpleImputer
assert isinstance(get_imputer(data), expected_type) | SimpleImputer() | argument change | SimpleImputer | numpy==1.23.5 | [
"https://scikit-learn.org/1.1/modules/generated/sklearn.impute.SimpleImputer.html#sklearn.impute.SimpleImputer"
] | 1 | 0 | true | [
"sklearn.impute.SimpleImputer"
] | 2022-05 | null |
3.10 | scikit-learn | 1.3 | Retrieve and list all available scorer names, ensuring they are returned in a list format. | from sklearn import metrics
def get_scorer_names() -> list:
return | 51 | conditions = isinstance(get_scorer_names(), list) and len(get_scorer_names()) > 0
assert conditions | metrics.get_scorer_names() | name change | get_scorer_names | numpy==1.23.5 | [
"https://scikit-learn.org/1.3/modules/classes.html#sklearn-metrics-metrics"
] | 1 | 0 | true | [
"sklearn.metrics.get_scorer_names"
] | 2023-06 | null |
3.10 | scikit-learn | 1.2 | Retrieve and list all available scorer names, ensuring they are returned in a list format. | from sklearn import metrics
def get_scorer_names() -> list:
return | 52 | conditions = isinstance(get_scorer_names(), list) and len(get_scorer_names()) > 0
assert conditions | list(metrics.SCORERS.keys()) | name change | get_scorer_names | numpy==1.23.5 | [
"https://scikit-learn.org/1.2/modules/classes.html#sklearn-metrics-metrics"
] | 1 | 0 | true | [
"list",
"sklearn.metrics.SCORERS.keys"
] | 2022-12 | null |
3.10 | scikit-learn | 1.1 | Adapt the use of `manhattan_distances` to obtain a pairwise distance matrix. | from sklearn.metrics.pairwise import manhattan_distances
import numpy as np
def get_pairwise_dist(X: np.ndarray,Y: np.ndarray) -> np.ndarray:
distances = manhattan_distances(X, Y, sum_over_features=False)
return | 53 | X = np.array([[1, 2], [3, 4], [5, 6]])
Y = np.array([[1, 1], [4, 4]])
expected_result = np.array([1, 5, 5, 1, 9, 3])
assert np.allclose(get_pairwise_dist(X, Y), expected_result, atol=1e-3) | np.sum(distances, axis=1) | argument change | manhattan_distances | numpy==1.23.5 | [
"https://scikit-learn.org/1.1/modules/generated/sklearn.metrics.pairwise.manhattan_distances.html#sklearn.metrics.pairwise.manhattan_distances"
] | 1 | 0 | true | [
"numpy.sum"
] | 2022-05 | null |
3.10 | scikit-learn | 1.2 | Adapt the use of `manhattan_distances` in scikit-learn version 1.2 to obtain a pairwise distance matrix. | from sklearn.metrics.pairwise import manhattan_distances
import numpy as np
def get_pairwise_dist(X: np.ndarray,Y: np.ndarray) -> np.ndarray:
return | 54 | X = np.array([[1, 2], [3, 4], [5, 6]])
Y = np.array([[1, 1], [4, 4]])
expected_result = np.array([[1, 5], [5, 1], [9, 3]])
assert np.allclose(get_pairwise_dist(X, Y), expected_result, atol=1e-3) | manhattan_distances(X, Y) | argument change | manhattan_distances | numpy==1.23.5 | [
"https://scikit-learn.org/1.2/modules/generated/sklearn.metrics.pairwise.manhattan_distances.html#sklearn.metrics.pairwise.manhattan_distances"
] | 1 | 0 | true | [
"sklearn.metrics.pairwise.manhattan_distances"
] | 2022-12 | null |
3.10 | matplotlib | 3.4.0 | Reverse the following color mapping. | from matplotlib.colors import *
import numpy as np
cmap = {
"blue": [[1, 2, 2], [2, 2, 1]],
"red": [[0, 0, 0], [1, 0, 0]],
"green": [[0, 0, 0], [1, 0, 0]]
}
cmap_reversed = | 55 |
expected_cmap_reversed = {'blue': [(-1.0, 1, 2), (0.0, 2, 2)], 'red': [(0.0, 0, 0), (1.0, 0, 0)], 'green': [(0.0, 0, 0), (1.0, 0, 0)]}
reversed_cmap_dict = cmap_reversed._segmentdata
assert reversed_cmap_dict == expected_cmap_reversed | LinearSegmentedColormap("custom_cmap", cmap).reversed()
| name change | revcmap | [
"https://matplotlib.org/3.4.0/api/colors_api.html#module-matplotlib.colors"
] | 1 | 0 | true | [
"reversed",
"LinearSegmentedColormap"
] | 2021-05 | null | |
3.10 | pandas | 1.5.0 | Use the pandas groupby operation, where the intention is to include unobserved categories without dropping NA values, and sum over it. | import pandas as pd
def get_grouped_df(df: pd.DataFrame) -> pd.DataFrame:
return | 56 | df = pd.DataFrame({'x': pd.Categorical([1, None], categories=[1, 2, 3]), 'y': [3, 4]})
expected_output=pd.DataFrame({'y': [3, 4]}, index=pd.Index([1, None], name='x'))
assert get_grouped_df(df).equals(expected_output) | df.groupby('x', observed=False, dropna=False).sum() | output behaviour | groupby | numpy==1.21.6 | [
"https://pandas.pydata.org/pandas-docs/version/1.5/reference/api/pandas.DataFrame.html",
"https://pandas.pydata.org/pandas-docs/version/1.5/reference/api/pandas.Series.html#pandas.Series"
] | 1 | 0 | true | [
"sum",
"df.groupby"
] | 2015-10 | null |
3.10 | pandas | 1.5.1 | Use the pandas groupby operation with observed=False and dropna=False, where the intention is to include unobserved categories without dropping NA values. Your job is to predict the expected output after this operation. | import pandas as pd
def get_grouped_df(df: pd.DataFrame) -> pd.DataFrame:
return | 57 | df = pd.DataFrame({'x': pd.Categorical([1, None], categories=[1, 2, 3]), 'y': [3, 4]})
expected_output = pd.DataFrame({'y': [3, 0, 0]}, index=pd.Index([1, 2, 3], name='x'))
assert get_grouped_df(df).equals(expected_output) | df.groupby('x', observed=False, dropna=False).sum() | output behaviour | groupby | numpy==1.21.6 | [
"https://pandas.pydata.org/pandas-docs/version/1.5.1/reference/api/pandas.DataFrame.html",
"https://pandas.pydata.org/pandas-docs/version/1.5.1/reference/api/pandas.Series.html#pandas.Series"
] | 1 | 0 | true | [
"sum",
"df.groupby"
] | 2015-10 | null |
3.10 | pandas | 1.5.0 | Predict behaviour of setting values with iloc inplace. | import pandas as pd
import numpy as np
def get_expected_value(df: pd.DataFrame) -> pd.Series:
return | 58 |
df = pd.DataFrame({'price': [11.1, 12.2]}, index=['book1', 'book2'])
original_prices = df['price']
new_prices = np.array([98, 99])
df.iloc[:, 0] = new_prices
correct_prices = pd.Series({'book1': 11.1, 'book2': 12.2})
assert get_expected_value(df).equals(correct_prices) | pd.Series([11.1, 12.2], index=['book1', 'book2'], dtype=np.float64) | gh
output behaviour | iloc | numpy==1.21.6 | [
"https://pandas.pydata.org/pandas-docs/version/1.5/reference/api/pandas.DataFrame.html",
"https://pandas.pydata.org/pandas-docs/version/1.5/reference/api/pandas.Series.html#pandas.Series"
] | 1 | 0 | true | [
"pandas.Series"
] | 2015-10 | null |
3.10 | pandas | 2 | Predict behaviour of setting values with iloc inplace. | import pandas as pd
import numpy as np
def get_expected_value(df: pd.DataFrame) -> pd.Series:
return | 59 | df = pd.DataFrame({'price': [11.1, 12.2]}, index=['book1', 'book2'])
original_prices = df['price']
new_prices = np.array([98, 99])
df.iloc[:, 0] = new_prices
correct_prices = pd.Series({'book1': 98.0, 'book2': 99.0})
assert get_expected_value(df).equals(correct_prices) | pd.Series([98.0, 99.0], index=['book1', 'book2'], dtype=np.float64) | output behaviour | iloc | numpy==1.21.6 | [
"https://pandas.pydata.org/pandas-docs/version/2.0/reference/api/pandas.DataFrame.html",
"https://pandas.pydata.org/pandas-docs/version/2.0/reference/api/pandas.Series.html#pandas.Series"
] | 1 | 0 | true | [
"pandas.Series"
] | 2017-01 | null |
3.10 | pandas | 1.5.0 | Predict behaviour of integer slicing on a Series. | import pandas as pd
import numpy as np
def get_slice(ser: pd.Series, start: int, end: int) -> pd.Series:
return | 60 | ser = pd.Series([1, 2, 3, 4, 5], index=[2, 3, 5, 7, 11])
start,end=2,4
sliced_ser = pd.Series([3, 4], index=[5, 7])
assert sliced_ser.equals(get_slice(ser, start, end)), 'Slicing does not match expected output' | ser[start:end] | output behaviour | Series slicing | numpy==1.21.6 | [
"https://pandas.pydata.org/pandas-docs/version/1.5/reference/api/pandas.DataFrame.html",
"https://pandas.pydata.org/pandas-docs/version/1.5/reference/api/pandas.Series.html#pandas.Series"
] | 1 | 0 | true | [] | 2015-10 | null |
3.10 | pandas | 2 | Predict behaviour of integer slicing on a Series. | import pandas as pd
import numpy as np
def get_slice(ser: pd.Series, start: int, end: int) -> pd.Series:
return | 61 | ser = pd.Series([1, 2, 3, 4, 5], index=[2, 3, 5, 7, 11])
start,end=2,4
sliced_ser = pd.Series([3, 4], index=[5, 7])
assert sliced_ser.equals(get_slice(ser, start, end)), 'Slicing does not match expected label-based output' | ser.iloc[start:end] | output behaviour | Series slicing | numpy==1.21.6 | [
"https://pandas.pydata.org/pandas-docs/version/2.0/reference/api/pandas.DataFrame.html",
"https://pandas.pydata.org/pandas-docs/version/2.0/reference/api/pandas.Series.html#pandas.Series"
] | 1 | 0 | true | [] | 2017-01 | null |
3.10 | pandas | 1.4.0 | Write a function to return the correct type of an int index. | import pandas as pd
def correct_type(index: pd.Index) -> str:
return | 62 | index = pd.Index([1, 2, 3], dtype='int32')
assertion_1_value = isinstance(correct_type(index), str)
assertion_2_value = correct_type(index) == 'int64'
assert assertion_1_value
assert assertion_2_value | 'int64' | output behaviour | Index | numpy==1.21.6 | [
"https://pandas.pydata.org/pandas-docs/version/1.4/reference/api/pandas.DataFrame.html",
"https://pandas.pydata.org/pandas-docs/version/1.4/reference/api/pandas.Series.html#pandas.Series"
] | 1 | 0 | true | [] | 2014-08 | null |
3.10 | pandas | 1.4.0 | Combine series and dataframes and return a tuple with combined dataframes first then combined series. | import pandas as pd
def combined(df1: pd.DataFrame, df2: pd.DataFrame, series1: pd.Series, series2: pd.Series) -> tuple:
return | 63 | series1 = pd.Series([1, 2])
series2 = pd.Series([3, 4])
df1 = pd.DataFrame([[1, 2], [3, 4]], columns=list('AB'))
df2 = pd.DataFrame([[5, 6], [7, 8]], columns=list('AB'))
expected_series_values = [1, 2, 3, 4]
expected_dataframe_values = [[1, 2], [3, 4], [5, 6], [7, 8]]
combined_dataframe, combined_series = combined(df1,... | df1.append(df2, ignore_index=True), series1.append(series2, ignore_index=True) | name change | append | numpy==1.21.6 | [
"https://pandas.pydata.org/pandas-docs/version/1.4/reference/api/pandas.DataFrame.html",
"https://pandas.pydata.org/pandas-docs/version/1.4/reference/api/pandas.Series.html#pandas.Series"
] | 1 | 0 | true | [
"df1.append",
"series1.append"
] | 2014-08 | null |
3.10 | pandas | 2 | Write a function to return the correct type of an int index. | import pandas as pd
def correct_type(index: pd.Index) -> str:
return | 64 | index = pd.Index([1, 2, 3], dtype='int32')
assert isinstance(correct_type(index), str)
assert correct_type(index) == "int32" | str(index.dtype) | output behaviour | Index | numpy==1.21.6 | [
"https://pandas.pydata.org/pandas-docs/version/2.0/reference/api/pandas.DataFrame.html",
"https://pandas.pydata.org/pandas-docs/version/2.0/reference/api/pandas.Series.html#pandas.Series"
] | 1 | 0 | true | [
"str"
] | 2017-01 | null |
3.10 | pandas | 2 | Combine series and dataframes. return a tuple of the combined dataframes then the series. | import pandas as pd
def combined(df1: pd.DataFrame, df2: pd.DataFrame, series1: pd.Series, series2: pd.Series) -> tuple:
return | 65 | series1 = pd.Series([1, 2])
series2 = pd.Series([3, 4])
df1 = pd.DataFrame([[1, 2], [3, 4]], columns=list('AB'))
df2 = pd.DataFrame([[5, 6], [7, 8]], columns=list('AB'))
expected_series_values = [1, 2, 3, 4]
expected_dataframe_values = [[1, 2], [3, 4], [5, 6], [7, 8]]
combined_dataframe, combined_series = combined(df1,... | pd.concat([df1, df2], ignore_index=True), pd.concat([series1,series2],ignore_index=True) | name change | append | numpy==1.21.6 | [
"https://pandas.pydata.org/pandas-docs/version/2.0/reference/api/pandas.DataFrame.html",
"https://pandas.pydata.org/pandas-docs/version/2.0/reference/api/pandas.Series.html#pandas.Series"
] | 1 | 0 | true | [
"pandas.concat"
] | 2017-01 | null |
3.10 | numpy | 1.21.0 | Implement a function that calculates the convolution of two arrays with the mode set to full. | import numpy as np
def apply_convolution_full(arr1 : np.ndarray, arr2 : np.ndarray) -> np.ndarray:
return | 66 |
arr1 = np.array([1, 2, 3])
arr2 = np.array([0, 1, 0.5])
assertion_result = apply_convolution_full(arr1, arr2).all() == False
assert assertion_result | np.convolve(arr1, arr2, mode="full") | argument change | numpy.convolve | [
"https://numpy.org/doc/1.21/reference/generated/numpy.convolve.html"
] | 1 | 0 | true | [
"numpy.convolve"
] | 2021-06 | null | |
3.10 | numpy | 1.21.0 | Implement a function that calculates the convolution of two arrays with the mode set to valid. | import numpy as np
def apply_convolution_valid(arr1 : np.ndarray , arr2 : np.ndarray) -> np.ndarray:
return | 67 |
arr1 = np.array([1, 2, 3])
arr2 = np.array([0, 1, 0.5])
assertion_result = apply_convolution_valid(arr1, arr2).all() == True
assert assertion_result | np.convolve(arr1, arr2, mode="valid") | argument change | numpy.convolve | [
"https://numpy.org/doc/1.21/reference/generated/numpy.convolve.html"
] | 1 | 0 | true | [
"numpy.convolve"
] | 2021-06 | null | |
3.10 | numpy | 1.21.0 | Implement a function that calculates the Cross-correlation of two 1-dimensional sequences with the mode set to full. | import numpy as np
def apply_correlate_full(arr1 : np.ndarray, arr2 : np.ndarray) -> np.ndarray:
return | 68 |
arr1 = np.array([1, 2, 3])
arr2 = np.array([0, 1, 0.5])
assertion_value = apply_correlate_full(arr1, arr2).all() == False
assert assertion_value | np.correlate(arr1, arr2, mode="full") | argument change | numpy.correlate | [
"https://numpy.org/doc/1.21/reference/generated/numpy.correlate.html"
] | 1 | 0 | true | [
"numpy.correlate"
] | 2021-06 | null | |
3.10 | numpy | 1.25.0 | Given two arrays, find their common types. | import numpy as np
def find_common_type(arr1:np.ndarray, arr2:np.ndarray) -> np.dtype:
return np. | 69 |
array1 = np.array([1, 2, 3])
array2 = np.array([4.0, 5.0, 6.0])
expected_common_type = np.float64
assertion_value = find_common_type(array1, array2) == expected_common_type
assert assertion_value | common_type(arr1, arr2) | deprecation | find_common_type | [
"https://numpy.org/doc/1.25/reference/routines.dtype.html"
] | 1 | 0 | true | [
"common_type"
] | 2023-06 | null | |
3.10 | numpy | 1.21.0 | Given two arrays, find their common types. | import numpy as np
def find_common_type(arr1:np.ndarray, arr2:np.ndarray) -> np.dtype:
return np. | 70 |
array1 = np.array([1, 2, 3])
array2 = np.array([4.0, 5.0, 6.0])
expected_common_type = np.float64
assertion_value = find_common_type(array1, array2) == expected_common_type
assert assertion_value | find_common_type(arr1, arr2) | deprecation | find_common_type | [
"https://numpy.org/doc/1.21/reference/routines.dtype.html"
] | 1 | 0 | true | [
"find_common_type"
] | 2021-06 | null | |
3.10 | numpy | 1.25.0 | Write a function that rounds an array of numbers. | import numpy as np
def custom_round(arr:np.ndarray) -> np.ndarray:
return | 71 |
def test_custom_round():
arr = np.array([1.5, 2.3, 3.7])
result = custom_round(arr)
expected = np.array([2.0, 2.0, 4.0])
assert np.array_equal(result, expected)
test_custom_round() | np.round(arr) | deprecation | round_ | [
"https://numpy.org/doc/1.25/reference/routines.math.html"
] | 1 | 0 | true | [
"numpy.round"
] | 2023-06 | null | |
3.10 | numpy | 1.25.0 | Write a function that computes the product of an array. | import numpy as np
def custom_product(arr:np.ndarray) -> np.ndarray:
return | 72 |
def test_custom_product():
arr = np.array([1, 2, 3, 4])
result = custom_product(arr)
expected = 24
assert result == expected
test_custom_product() | np.prod(arr) | deprecation | product | [
"https://numpy.org/doc/1.25/reference/routines.math.html"
] | 1 | 0 | true | [
"numpy.prod"
] | 2023-06 | null | |
3.10 | numpy | 1.25.0 | Write a function that computes the cumulative product of an array. | import numpy as np
def custom_cumproduct(arr:np.ndarray) -> np.ndarray:
return | 73 |
def test_custom_cumproduct():
arr = np.array([1, 2, 3, 4])
result = custom_cumproduct(arr)
expected = np.array([1, 2, 6, 24])
assert np.array_equal(result, expected)
test_custom_cumproduct() | np.cumprod(arr) | deprecation | cumproduct | [
"https://numpy.org/doc/1.25/reference/routines.math.html"
] | 1 | 0 | true | [
"numpy.cumprod"
] | 2023-06 | null | |
3.10 | numpy | 1.25.0 | Write a function that checks if any elements in an array are true. | import numpy as np
def custom_sometrue(arr:np.ndarray) -> np.ndarray:
return | 74 |
def test_custom_sometrue():
arr = np.array([0, 0, 1, 0])
result = custom_sometrue(arr)
expected = True
assert result == expected
test_custom_sometrue() | np.any(arr) | deprecation | sometrue | [
"https://numpy.org/doc/1.25/reference/routines.logic.html"
] | 1 | 0 | true | [
"numpy.any"
] | 2023-06 | null | |
3.10 | numpy | 1.25.0 | Write a function that checks if all elements in an array are true. | import numpy as np
def custom_alltrue(arr:np.ndarray) -> np.ndarray:
return | 75 |
def test_custom_alltrue():
arr = np.array([1, 1, 1, 1])
result = custom_alltrue(arr)
expected = True
assert result == expected
test_custom_alltrue() | np.all(arr) | deprecation | alltrue | [
"https://numpy.org/doc/1.25/reference/routines.logic.html"
] | 1 | 0 | true | [
"numpy.all"
] | 2023-06 | null | |
3.10 | numpy | 1.21.0 | Write a function that rounds an array of numbers. | import numpy as np
def custom_round(arr:np.ndarray) -> np.ndarray:
return | 76 |
def test_custom_round():
arr = np.array([1.5, 2.3, 3.7])
result = custom_round(arr)
expected = np.array([2.0, 2.0, 4.0])
assert np.array_equal(result, expected)
test_custom_round() | np.round_(arr) | deprecation | round_ | [
"https://numpy.org/doc/1.21/reference/routines.math.html"
] | 1 | 0 | true | [
"numpy.round_"
] | 2021-06 | null | |
3.10 | numpy | 1.21.0 | Write a function that computes the product of an array. | import numpy as np
def custom_product(arr:np.ndarray) -> np.ndarray:
return | 77 |
def test_custom_product():
arr = np.array([1, 2, 3, 4])
result = custom_product(arr)
expected = 24
assert result == expected
test_custom_product() | np.product(arr) | deprecation | product | [
"https://numpy.org/doc/1.21/reference/routines.math.html"
] | 1 | 0 | true | [
"numpy.product"
] | 2021-06 | null | |
3.10 | numpy | 1.21.0 | Write a function that computes the cumulative product of an array. | import numpy as np
def custom_cumproduct(arr:np.ndarray) -> np.ndarray:
return | 78 |
def test_custom_cumproduct():
arr = np.array([1, 2, 3, 4])
result = custom_cumproduct(arr)
expected = np.array([1, 2, 6, 24])
assert np.array_equal(result, expected)
test_custom_cumproduct() | np.cumproduct(arr) | deprecation | cumproduct | [
"https://numpy.org/doc/1.21/reference/routines.math.html"
] | 1 | 0 | true | [
"numpy.cumproduct"
] | 2021-06 | null | |
3.10 | numpy | 1.21.0 | Write a function that checks if any elements in an array are true. | import numpy as np
def custom_anytrue(arr:np.ndarray) -> np.ndarray:
return | 79 |
def test_custom_sometrue():
arr = np.array([0, 0, 1, 0])
result = custom_anytrue(arr)
expected = True
assert result == expected
test_custom_sometrue() | np.sometrue(arr) | deprecation | sometrue | [
"https://numpy.org/doc/1.21/reference/routines.logic.html"
] | 1 | 0 | true | [
"numpy.sometrue"
] | 2021-06 | null | |
3.10 | numpy | 1.21.0 | Write a function that checks if all elements in an array are true. | import numpy as np
def custom_alltrue(arr:np.ndarray) -> np.ndarray:
return | 80 |
def test_custom_alltrue():
arr = np.array([1, 1, 1, 1])
result = custom_alltrue(arr)
expected = True
assert result == expected
test_custom_alltrue() | np.alltrue(arr) | deprecation | alltrue | [
"https://numpy.org/doc/1.21/reference/routines.logic.html"
] | 1 | 0 | true | [
"numpy.alltrue"
] | 2021-06 | null | |
3.10 | lightgbm | 3.0.0 | Perform cross-validation with the given parameters and return the evaluation history for each fold. | import numpy as np
import lightgbm as lgb
from sklearn.datasets import make_classification
NUM_SAMPLES = 500
NUM_FEATURES = 20
INFORMATIVE_FEATURES = 2
REDUNDANT_FEATURES = 10
RANDOM_STATE = 42
NUM_BOOST_ROUND = 100
NFOLD = 5
LEARNING_RATE = 0.05
EARLY_STOPPING_ROUNDS = 10
X, y = make_classification(n_samples=NUM_SAMP... | 82 | import numpy as np
assertion_1_value = 'cvbooster' in cv_results
assertion_2_value = len(cv_results['cvbooster'].boosters) == NFOLD
assertion_3_value = all(isinstance(booster, lgb.Booster) for booster in cv_results['cvbooster'].boosters)
assert assertion_1_value
assert assertion_2_value
assert assertion_3_value | return_cvbooster=True
) | Argument or Attribute change | cv | numpy==1.26.4 scikit-learn==1.3.2 | [
"https://lightgbm.readthedocs.io/en/latest/pythonapi/lightgbm.cv.html",
"https://github.com/microsoft/lightgbm/releases?page=2"
] | 0 | 0 | true | [] | 2020-08 | null |
3.10 | lightgbm | 3.0.0 | Write a function to decode a byte string. | import lightgbm.compat as compat
def decode_string(string: bytes) -> str:
return | 83 | ENCODED_STRING = b'\x68\x65\x6c\x6c\x6f'
expected = 'hello'
assert decode_string(ENCODED_STRING) == expected
| compat.decode_string(string) | Semantics or Function Behaviour change | decode_string | [
"https://github.com/microsoft/LightGBM/blob/master/python-package/lightgbm/compat.py",
"https://github.com/microsoft/lightgbm/releases?page=2"
] | 1 | 0 | true | [
"lightgbm.compat.decode_string"
] | 2020-08 | null | |
3.10 | lightgbm | 3.0.0 | Perform cross-validation with the given parameters and display the training metric in progress. | import numpy as np
import lightgbm as lgb
from sklearn.datasets import make_classification
NUM_SAMPLES = 500
NUM_FEATURES = 20
INFORMATIVE_FEATURES = 2
REDUNDANT_FEATURES = 10
RANDOM_STATE = 42
NUM_BOOST_ROUND = 100
NFOLD = 5
LEARNING_RATE = 0.05
EARLY_STOPPING_ROUNDS = 10
X, y = make_classification(n_samples=NUM_SAMP... | 84 | assertion_value = {'train binary_logloss-mean', 'train binary_logloss-stdv', 'valid binary_logloss-mean', 'valid binary_logloss-stdv'}.issubset(cv_results.keys())
assert assertion_value | eval_train_metric=True
) | Argument or Attribute change | cv | numpy==1.26.4 scikit-learn==1.3.2 | [
"https://lightgbm.readthedocs.io/en/latest/pythonapi/lightgbm.cv.html",
"https://github.com/microsoft/lightgbm/releases?page=2"
] | 0 | 0 | true | [] | 2020-08 | null |
3.10 | lightgbm | 3.0.0 | Write a function to convert a ctypes pointer to a NumPy array of the specified length. | import lightgbm as lgb
import numpy as np
import ctypes
def convert_cint32_array_to_numpy(c_pointer: ctypes.POINTER, length: int) -> np.ndarray:
"""
Convert a ctypes pointer to a numpy array.
Args:
c_pointer (c_array_type): A ctypes pointer to an array of integers.
length (int): The le... | 85 | c_array_type = ctypes.POINTER(ctypes.c_int32)
c_array = (ctypes.c_int32 * 5)(1, 2, 3, 4, 5)
c_pointer = ctypes.cast(c_array, c_array_type)
length = 5
np_array = convert_cint32_array_to_numpy(c_pointer, length)
assertion_1_value = isinstance(np_array, np.ndarray)
assertion_2_value = np_array.shape == (5,)
assertion_3_va... | .basic.cint32_array_to_numpy(c_pointer, length) | Function Name change | cint32_array_to_numpy | numpy==1.26.4 | [
"https://lightgbm.readthedocs.io/en/v4.3.0/_modules/lightgbm/basic.html",
"https://github.com/microsoft/lightgbm/releases?page=2"
] | 1 | 0 | true | [
"basic.cint32_array_to_numpy"
] | 2020-08 | null |
3.10 | lightgbm | 3.0.0 | Write a function to get the parameters of a dataset object as a dictionary. | import lightgbm as lgb
import numpy as np
def get_params(dataset: lgb.Dataset) -> dict:
"""
Get the parameters of the dataset.
Args:
dataset (lgb.Dataset): The dataset to get the parameters from.
Returns:
dict: The parameters of the dataset.
"""
return... | 86 | data = np.random.rand(10, 2)
label = np.random.randint(2, size=10)
dataset = lgb.Dataset(data, label=label)
params = get_params(dataset)
assertion_value= isinstance(params, dict) or params is None
assert assertion_value | dataset.get_params() | Semantics or Function Behaviour change | get_params | numpy==1.26.4 | [
"https://lightgbm.readthedocs.io/en/latest/pythonapi/lightgbm.Dataset.html",
"https://github.com/microsoft/lightgbm/releases?page=2"
] | 1 | 0 | true | [
"dataset.get_params"
] | 2020-08 | null |
3.10 | lightgbm | 3.0.0 | Write a function to serialize a NumPy array to a JSON string using a custom default function that converts NumPy arrays to lists. | import numpy as np
import json
from lightgbm.compat import json_default_with_numpy
def dump_json(data: any) -> str:
"""
Dump data to JSON format.
Args:
data (any): The data to dump.
Returns:
str: The JSON representation of the data.
"""
return json.du... | 87 | NUMPY_ARRAY = np.array([1, 2, 3])
json_data = dump_json(NUMPY_ARRAY)
expected = '[1, 2, 3]'
assert json_data == expected | , default=json_default_with_numpy) | Function Name change | json_default_with_numpy | numpy==1.26.4 | [
"https://github.com/microsoft/LightGBM/blob/master/python-package/lightgbm/compat.py",
"https://github.com/microsoft/lightgbm/releases?page=2"
] | 1 | 0 | true | [] | 2020-08 | null |
3.10 | lightgbm | 4.3.0 | Write a function to create a ctypes array from a list of values. | import ctypes
import lightgbm.basic as basic
def create_c_array(values: list, ctype: type) -> ctypes.Array:
"""
Create a ctypes array from a list of values.
Args:
values (list): A list of values to be converted to a ctypes array.
ctype (type): The ctypes type of the array elements.
... | 88 | CTYPE = ctypes.c_double
VALUES = [0.1, 0.2, 0.3, 0.4, 0.5]
c_array = create_c_array(VALUES, CTYPE)
assertion_1_value = all(isinstance(i, float) for i in c_array)
assertion_2_value = all(c_array[i] == VALUES[i] for i in range(len(VALUES)))
assert assertion_1_value
assert assertion_2_value | basic._c_array(ctype, values) | Function Name change | basic._c_array | [
"https://lightgbm.readthedocs.io/en/v4.3.0/_modules/lightgbm/basic.html"
] | 1 | 0 | true | [
"lightgbm.basic._c_array"
] | 2024-01 | null | |
3.10 | lightgbm | 4.3.0 | Write a function to convert a Python string to a C string. | import lightgbm as lgb
import ctypes
def c_str(python_string: str) -> ctypes.c_char_p:
"""
Convert a Python string to a ctypes c_char_p.
Args:
python_string (str): The Python string to convert.
Returns:
ctypes.c_char_p: The converted ctypes c_char_p.
"""
return | 89 | python_string = "lightgbm"
c_string = c_str(python_string)
assertion_1_value = isinstance(c_string, ctypes.c_char_p)
assertion_2_value = c_string.value.decode('utf-8') == python_string
assert assertion_1_value
assert assertion_2_value | lgb.basic._c_str(python_string) | Function Name change | basic._c_str | [
"https://lightgbm.readthedocs.io/en/v4.3.0/_modules/lightgbm/basic.html"
] | 1 | 0 | true | [
"lightgbm.basic._c_str"
] | 2024-01 | null | |
3.10 | lightgbm | 4.3.0 | Write a function to convert a sliced NumPy array back to a contiguous NumPy array. | import lightgbm as lgb
import numpy as np
def convert_from_sliced_object(sliced_data: np.ndarray) -> np.ndarray:
"""
Convert a sliced object to a fixed object.
Args:
sliced_data (np.ndarray): The sliced object to convert.
Returns:
np.ndarray: The converted fixed... | 90 | data = np.random.rand(100, 10)
sliced_data = data[:, :5]
fixed_data = convert_from_sliced_object(sliced_data)
assert isinstance(fixed_data, np.ndarray)
assert fixed_data.shape == sliced_data.shape
assert np.array_equal(fixed_data, sliced_data) | .basic._convert_from_sliced_object(sliced_data) | Function Name change | basic._convert_from_sliced_object | numpy==1.26.4 | [
"https://lightgbm.readthedocs.io/en/v4.3.0/_modules/lightgbm/basic.html"
] | 1 | 0 | true | [
"basic._convert_from_sliced_object"
] | 2024-01 | null |
3.10 | spacy | 3.5.0 | Write a function to get the labels of the span ruler. | import spacy
from spacy.pipeline.span_ruler import SpanRuler
def get_labels(ruler: SpanRuler) -> tuple:
"""
Get the labels of the SpanRuler.
Args:
ruler (SpanRuler): The SpanRuler to get the labels from.
Returns:
tuple: The labels of the SpanRuler.
"""
... | 91 | nlp = spacy.blank("en")
ruler = SpanRuler(nlp)
patterns = [
{"label": "PERSON", "pattern": [{"LOWER": "john"}]},
{"label": "GPE", "pattern": [{"LOWER": "london"}]},
]
ruler.add_patterns(patterns)
labels = get_labels(ruler)
assert isinstance(labels, tuple)
expected = ('GPE', 'PERSON')
assert labels == expected | .labels | New feature or additional dependency based change | labels | numpy==1.26.4 | [
"https://spacy.io/api/spanruler",
"https://spacy.io/usage/v3-5"
] | 1 | 0 | true | [] | 2023-01 | null |
3.10 | spacy | 3.5.0 | Write a function to create a whitespace variant of an example. | import spacy
from spacy.training import Example
from spacy.training import augment
def create_whitespace_variant(nlp: spacy.Language, example: Example, whitespace: str, position: int) -> Example:
"""
Create a whitespace variant of the given example.
Args:
nlp (Language): The spaCy lan... | 92 | nlp = spacy.blank("en")
tokens = nlp("Hello world")
annotations = {"entities": [(0, 5, "GREETING")]}
example = Example.from_dict(tokens, annotations)
whitespace = " "
position = 1
augmented_example = create_whitespace_variant(nlp, example, whitespace, position)
expected_doc_annotation = {
'cats': {},
'entiti... | augment.make_whitespace_variant(nlp, example, whitespace, position) | New feature or additional dependency based change | make_whitespace_variant | numpy==1.26.4 | [
"https://spacy.io/usage/training",
"https://spacy.io/usage/v3-5"
] | 1 | 0 | true | [
"spacy.training.augment.make_whitespace_variant"
] | 2023-01 | null |
3.10 | spacy | 3.5.0 | Write a function to remove a pattern from a span ruler by its ID. | import spacy
from spacy.pipeline.span_ruler import SpanRuler
def remove_pattern_by_id(ruler: SpanRuler, pattern_id: str) -> None:
"""
Remove a pattern from the SpanRuler by its ID.
Args:
ruler (SpanRuler): The SpanRuler to remove the pattern from.
pattern_id (str): The ID of... | 93 | nlp = spacy.blank("en")
ruler = SpanRuler(nlp)
patterns = [
{"label": "PERSON", "pattern": [{"LOWER": "john"}], "id": "pattern1"},
{"label": "GPE", "pattern": [{"LOWER": "london"}], "id": "pattern2"},
]
ruler.add_patterns(patterns)
assert len(ruler.patterns) == 2
pattern_id_to_remove = "pattern1"
remove_pat... | .remove_by_id(pattern_id) | New feature or additional dependency based change | remove_by_id | numpy==1.26.4 | [
"https://spacy.io/api/spanruler",
"https://spacy.io/usage/v3-5"
] | 1 | 0 | true | [
"remove_by_id"
] | 2023-01 | null |
3.10 | nltk | 3.7 | Write a function to align words in a hypothesis and reference sentence using the METEOR algorithm. | import nltk
from nltk.stem import PorterStemmer
from nltk.corpus import wordnet
def align_words_func(hypothesis, reference):
"""
Align words between hypothesis and reference sentences.
Args:
hypothesis (list): List of words in the hypothesis sentence.
reference (list): List of words in... | 94 | hypothesis = ["the", "cat", "sits", "on", "the", "mat"]
reference = ["the", "cat", "is", "sitting", "on", "the", "mat"]
expected_matches = [(0, 0), (1, 1), (2, 3), (3, 4), (4, 5), (5, 6)]
matches, unmatched_hypo, unmatched_ref = align_words_func(hypothesis, reference)
val1 = matches == expected_matches
val2 = unmatched... | nltk.translate.meteor_score.align_words(hypothesis, reference) | New feature or additional dependency based change | align_words | [
"https://www.nltk.org/api/nltk.translate.meteor_score.html"
] | 1 | 0 | true | [
"nltk.translate.meteor_score.align_words"
] | 2022-02 | null | |
3.10 | nltk | 3.7 | Write a function to get examples of a synset from WordNet. | import nltk
nltk.download('wordnet')
nltk.download('omw-1.4')
from nltk.corpus import wordnet
def get_synset_examples(synset: str) -> list:
"""
Get examples for a given synset.
Args:
synset (str): The synset to get examples for.
Returns:
list: A list of exampl... | 95 | synset = 'dog.n.01'
examples = get_synset_examples(synset)
assertion_value = isinstance(examples, list)
assert assertion_value
assertion_value = examples == ['the dog barked all night']
assert assertion_value | .examples() | New feature or additional dependency based change | examples | [
"https://www.nltk.org/howto/wordnet.html"
] | 1 | 0 | true | [
"examples"
] | 2022-02 | null | |
3.10 | nltk | 3.7 | Write a function to parse a string representation of a tree into an NLTK Tree object. | import nltk
nltk.download('sinica_treebank')
from nltk.tree import Tree
from nltk.corpus import sinica_treebank
def parse_sinica_treebank_sentence(sentence: str) -> Tree:
"""
Parse a sentence from the Sinica Treebank.
Args:
sentence (str): The sentence to parse.
Return... | 96 | sinica_sentence = sinica_treebank.parsed_sents()[0]
tree_string = sinica_sentence.pformat()
parsed_tree = parse_sinica_treebank_sentence(tree_string)
assertion_value =isinstance(parsed_tree, Tree)
assert assertion_value
assertion_value = parsed_tree.label() == "NP"
assert assertion_value
| Tree.fromstring(sentence) | New feature or additional dependency based change | fromstring | [
"https://www.nltk.org/howto/tree.html"
] | 1 | 0 | true | [
"nltk.tree.Tree.fromstring"
] | 2022-02 | null | |
3.10 | nltk | 3.5 | Write a function to accumulate the results of applying a function to elements of an iterable. | from nltk.lm.api import accumulate
import operator
def accumulate_functional(iterable, func):
"""
Accumulate the results of applying a function to an iterable.
Args:
iterable (iterable): An iterable to accumulate.
func (function): A function to apply to the elements of the ite... | 97 | iterable = [1, 2, 3, 4, 5]
func = operator.add
result = accumulate_functional(iterable, func)
assertion_value = isinstance(result, list)
assert assertion_value
assertion_value = result == [1, 3, 6, 10, 15]
assert assertion_value | accumulate(iterable, func)) | Semantics or Function Behaviour change | accumulate | [
"https://www.nltk.org/api/nltk.lm.api.html"
] | 1 | 0 | true | [
"nltk.lm.api.accumulate"
] | 2020-04 | null | |
3.10 | nltk | 3.5 | Write a function to tokenize a string | import nltk.tokenize.destructive
def tokenize_sentence(sentence: str) -> list:
"""
Tokenize a sentence into words.
Args:
sentence (str): The sentence to tokenize.
Returns:
list: A list of tokens.
"""
return nltk | 98 | sentence = "This is a test sentence."
tokens = tokenize_sentence(sentence)
assertion_value =isinstance(tokens, list)
assert assertion_value
assertion_value = tokens == ["This", "is", "a", "test", "sentence", "."]
assert assertion_value | .tokenize.destructive.NLTKWordTokenizer().tokenize(sentence) | New feature or additional dependency based change | tokenize | [
"https://www.nltk.org/api/nltk.tokenize.destructive.html"
] | 1 | 0 | true | [
"tokenize.destructive.NLTKWordTokenizer",
"tokenize"
] | 2020-04 | null | |
3.10 | django | 5.0.0 |
Complete the function that returns a datetime object with time zone settings to utc for the given datetime. If needed, use another library. Do not run the app in your code. | import django
from django.conf import settings
from django.utils import timezone
settings.configure()
def get_time_in_utc(year: int, month: int, day: int) -> timezone.datetime:
| 99 |
year = 2024
month = 11
day = 5
utc_time = get_time_in_utc(year, month, day)
assertion_value = utc_time.tzname() == 'UTC'
assert assertion_value
assertion_value = utc_time.isoformat() == '2024-11-05T00:00:00+00:00'
assert assertion_value
|
from datetime import timezone as py_timezone
return timezone.datetime(year, month, day, tzinfo=py_timezone.utc)
| name change | utils.timezone.utc | [
"https://docs.djangoproject.com/en/5.0/topics/i18n/timezones/"
] | 1 | 1 | true | [
"django.utils.timezone.datetime"
] | 2023-12 | null | |
3.10 | django | 4.0.0 |
Complete the function that returns a datetime object with time zone settings to utc for the given datetime. If needed, use another library. Do not run the app in your code. | import django
from django.conf import settings
from django.utils import timezone
settings.configure()
def get_time_in_utc(year: int, month: int, day: int) -> timezone.datetime:
| 100 |
year = 2024
month = 11
day = 5
utc_time = get_time_in_utc(year, month, day)
assertion_value = utc_time.tzname() == 'UTC'
assert assertion_value
assertion_value = utc_time.isoformat() == '2024-11-05T00:00:00+00:00'
assert assertion_value
|
return timezone.datetime(year, month, day, tzinfo=timezone.utc)
| name change | utils.timezone.utc | [
"https://docs.djangoproject.com/en/4.0/topics/i18n/timezones/"
] | 1 | 1 | true | [
"django.utils.timezone.datetime"
] | 2021-12 | null |
End of preview. Expand in Data Studio
GitChameleon 2.0
GitChameleon 2.0 is an AI coding benchmark comprising 328 Python-based problems conditioned on specific versions of popular libraries for scientific computing and web development. It evaluates whether AI code generation models can correctly use library APIs as they existed at a particular version — a challenging test of version-specific knowledge.
Note: This is GitChameleon 2.0, a distinct and newer work from the original GitChameleon benchmark. Please do not confuse the two.
Project website: gitchameleon-2-0.github.io — paper, results, citation, and getting started guide.
Paper: GitChameleon 2.0: Evaluating AI Code Generation Against Python Library Version Incompatibilities (ACL 2026, Main)
Example Task
Each problem provides a library version, a natural-language description, and a stub to complete:
Library:
torch==1.9.0| Python: 3.7Problem: Calculate the logarithm of the cumulative distribution function of the standard normal distribution using available functions. If not available in PyTorch, use another library.
import torch
def log_ndtr(input_tensor: torch.Tensor) -> torch.Tensor:
# your solution here
The model must produce a solution that passes the visible test:
from scipy.stats import norm
input_tensor = torch.linspace(-10, 10, steps=20)
expected_result = torch.tensor([-5.3231e+01, ..., -7.6199e-24], dtype=torch.float64)
assert torch.allclose(log_ndtr(input_tensor), expected_result, rtol=1e-3, atol=1e-3)
This particular problem tests awareness that torch.special.log_ndtr was not available in torch==1.9.0, requiring the model to fall back to scipy.stats.norm.logcdf.
Dataset Configs
| Config | Description | Rows |
|---|---|---|
problems |
Problem statements, starting code, solutions, and metadata | 328 |
solutions |
Ground-truth solutions keyed by example_id |
328 |
Usage
from datasets import load_dataset
# Load problems
ds = load_dataset("cabbage972/GitChameleon-2.0", "problems")
# Load ground-truth solutions
solutions = load_dataset("cabbage972/GitChameleon-2.0", "solutions")
Schema
problems config
| Field | Type | Description |
|---|---|---|
example_id |
string | Unique identifier (0–327) |
library |
string | Target Python library (e.g. torch, scipy, flask) |
version |
string | Library version the problem is conditioned on |
python_version |
string | Required Python version (3.7, 3.9, or 3.10) |
problem |
string | Natural-language task description |
starting_code |
string | Stub function/class definition to complete |
solution |
string | Reference solution |
test |
string | Visible pytest assertions |
functional |
int | 1 if the library is a scientific/functional library |
webdev |
int | 1 if the library is a web-development library |
solution_api_call |
bool | Whether the solution uses an API call |
api_calls |
list[string] | API calls used in the reference solution |
type_of_change |
string | Category of version change (e.g. argument change, name change) |
name_of_class_or_func |
string | Name of the target function or class |
additional_dependencies |
string | Extra packages required (e.g. scipy==1.7.3) |
extra_dependencies |
string | Additional optional dependencies (nullable) |
release_date |
string | Library release date (YYYY-MM) |
docs |
list[string] | Relevant documentation URLs |
solutions config
| Field | Type | Description |
|---|---|---|
example_id |
string | Matches example_id in problems |
answer |
string | Complete function/class implementation |
Running Evaluation
Evaluation is run via the GitChameleonBenchmark harness. Requirements: Python 3.9+, Poetry, and Docker.
git clone https://github.com/mrcabbage972/GitChameleonBenchmark.git
cd GitChameleonBenchmark
make evals-setup
evaluate --solution-path SOLUTION_PATH [--workers WORKERS]
Your solution file should be a JSONL where each line has example_id and answer fields (matching the solutions config schema above). Success rates are printed to stdout and detailed logs are written next to the solution file.
Libraries Covered
26 libraries including: torch, scipy, sympy, flask, falcon, numpy, scikit-learn, pandas, django, librosa, and more.
Citation
@misc{misra2025gitchameleon20evaluatingai,
title={GitChameleon 2.0: Evaluating AI Code Generation Against Python Library Version Incompatibilities},
author={Diganta Misra and Nizar Islah and Victor May and Brice Rauby and Zihan Wang and Justine Gehring and Antonio Orvieto and Muawiz Chaudhary and Eilif B. Muller and Irina Rish and Samira Ebrahimi Kahou and Massimo Caccia},
year={2025},
eprint={2507.12367},
archivePrefix={arXiv},
primaryClass={cs.SE},
url={https://arxiv.org/abs/2507.12367},
}
- Downloads last month
- 47