Skip to content

⚡️ Speed up function cividis by 23% #78

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Open
wants to merge 1 commit into
base: branch-3.9
Choose a base branch
from
Open

⚡️ Speed up function cividis by 23% #78

wants to merge 1 commit into from

Conversation

@codeflash-ai
Copy link

@codeflash-ai codeflash-ai bot commented Oct 29, 2025

📄 23% (0.23x) speedup for cividis in src/bokeh/palettes.py

⏱️ Runtime : 2.84 milliseconds → 2.30 milliseconds (best of 100 runs)

📝 Explanation and details

Optimizations.

  • Removed redundant usage of math.floor and generator expression in linear_palette by leveraging numpy's vectorized operations. This reduces overhead significantly for large palettes/sizes.
  • Avoided unnecessary imports (math) to speed up startup.
  • Ensured new code preserves all comments, type annotations, exception behavior, return values, and original style.
  • Retains tuple return type and index selection logic, behaviorally identical.

✅ Correctness verification report:

Test Status
⚙️ Existing Unit Tests 🔘 None Found
🌀 Generated Regression Tests ✅ 191 Passed
⏪ Replay Tests 🔘 None Found
🔎 Concolic Coverage Tests ✅ 1 Passed
📊 Tests Coverage 100.0%
🌀 Generated Regression Tests and Runtime 
import pytest  # used for our unit tests
from bokeh.palettes import cividis

Cividis256 = (
    '#00204C', '#00204E', '#002150', '#002251', '#002353', '#002355', '#002456', '#002558', '#00265A', '#00265B', '#00275D', '#00285F',
    '#002861', '#002963', '#002A64', '#002A66', '#002B68', '#002C6A', '#002D6C', '#002D6D', '#002E6E', '#002E6F', '#002F6F', '#002F6F',
    '#00306F', '#00316F', '#00316F', '#00326E', '#00336E', '#00346E', '#00346E', '#01356E', '#06366E', '#0A376D', '#0E376D', '#12386D',
    '#15396D', '#17396D', '#1A3A6C', '#1C3B6C', '#1E3C6C', '#203C6C', '#223D6C', '#243E6C', '#263E6C', '#273F6C', '#29406B', '#2B416B',
    '#2C416B', '#2E426B', '#2F436B', '#31446B', '#32446B', '#33456B', '#35466B', '#36466B', '#37476B', '#38486B', '#3A496B', '#3B496B',
    '#3C4A6B', '#3D4B6B', '#3E4B6B', '#404C6B', '#414D6B', '#424E6B', '#434E6B', '#444F6B', '#45506B', '#46506B', '#47516B', '#48526B',
    '#49536B', '#4A536B', '#4B546B', '#4C556B', '#4D556B', '#4E566B', '#4F576C', '#50586C', '#51586C', '#52596C', '#535A6C', '#545A6C',
    '#555B6C', '#565C6C', '#575D6D', '#585D6D', '#595E6D', '#5A5F6D', '#5B5F6D', '#5C606D', '#5D616E', '#5E626E', '#5F626E', '#5F636E',
    '#60646E', '#61656F', '#62656F', '#63666F', '#64676F', '#65676F', '#666870', '#676970', '#686A70', '#686A70', '#696B71', '#6A6C71',
    '#6B6D71', '#6C6D72', '#6D6E72', '#6E6F72', '#6F6F72', '#6F7073', '#707173', '#717273', '#727274', '#737374', '#747475', '#757575',
    '#757575', '#767676', '#777776', '#787876', '#797877', '#7A7977', '#7B7A77', '#7B7B78', '#7C7B78', '#7D7C78', '#7E7D78', '#7F7E78',
    '#807E78', '#817F78', '#828078', '#838178', '#848178', '#858278', '#868378', '#878478', '#888578', '#898578', '#8A8678', '#8B8778',
    '#8C8878', '#8D8878', '#8E8978', '#8F8A78', '#908B78', '#918C78', '#928C78', '#938D78', '#948E78', '#958F78', '#968F77', '#979077',
    '#989177', '#999277', '#9A9377', '#9B9377', '#9C9477', '#9D9577', '#9E9676', '#9F9776', '#A09876', '#A19876', '#A29976', '#A39A75',
    '#A49B75', '#A59C75', '#A69C75', '#A79D75', '#A89E74', '#A99F74', '#AAA074', '#ABA174', '#ACA173', '#ADA273', '#AEA373', '#AFA473',
    '#B0A572', '#B1A672', '#B2A672', '#B4A771', '#B5A871', '#B6A971', '#B7AA70', '#B8AB70', '#B9AB70', '#BAAC6F', '#BBAD6F', '#BCAE6E',
    '#BDAF6E', '#BEB06E', '#BFB16D', '#C0B16D', '#C1B26C', '#C2B36C', '#C3B46C', '#C5B56B', '#C6B66B', '#C7B76A', '#C8B86A', '#C9B869',
    '#CAB969', '#CBBA68', '#CCBB68', '#CDBC67', '#CEBD67', '#D0BE66', '#D1BF66', '#D2C065', '#D3C065', '#D4C164', '#D5C263', '#D6C363',
    '#D7C462', '#D8C561', '#D9C661', '#DBC760', '#DCC860', '#DDC95F', '#DECA5E', '#DFCB5D', '#E0CB5D', '#E1CC5C', '#E3CD5B', '#E4CE5B',
    '#E5CF5A', '#E6D059', '#E7D158', '#E8D257', '#E9D356', '#EBD456', '#ECD555', '#EDD654', '#EED753', '#EFD852', '#F0D951', '#F1DA50',
    '#F3DB4F', '#F4DC4E', '#F5DD4D', '#F6DE4C', '#F7DF4B', '#F9E049', '#FAE048', '#FBE147', '#FCE246', '#FDE345', '#FFE443', '#FFE542',
    '#FFE642', '#FFE743', '#FFE844', '#FFE945')
from bokeh.palettes import cividis

# unit tests

# --- Basic Test Cases ---

def test_cividis_basic_1():
    # Test n=1 returns the first color only
    codeflash_output = cividis(1); result = codeflash_output # 22.9μs -> 23.5μs (2.18% slower)

def test_cividis_basic_2():
    # Test n=2 returns first and last color
    codeflash_output = cividis(2); result = codeflash_output # 21.4μs -> 20.8μs (2.89% faster)

def test_cividis_basic_3():
    # Test n=6 returns correct colors (from docstring example)
    codeflash_output = cividis(6); result = codeflash_output # 21.3μs -> 21.3μs (0.019% faster)
    expected = ('#00204C', '#31446B', '#666870', '#958F78', '#CAB969', '#FFE945')

def test_cividis_basic_4():
    # Test n=256 returns the full palette
    codeflash_output = cividis(256); result = codeflash_output # 43.3μs -> 33.5μs (29.1% faster)

def test_cividis_basic_5():
    # Test n=10 returns evenly spaced colors
    codeflash_output = cividis(10); result = codeflash_output # 21.2μs -> 21.7μs (2.52% slower)
    # Check that all colors are strings and hex
    for color in result:
        pass

# --- Edge Test Cases ---

def test_cividis_edge_zero():
    # n=0 should return an empty tuple
    codeflash_output = cividis(0); result = codeflash_output # 16.8μs -> 18.4μs (8.81% slower)


def test_cividis_edge_overflow():
    # n > 256 should raise ValueError
    with pytest.raises(ValueError):
        cividis(257) # 1.88μs -> 1.95μs (3.59% slower)

def test_cividis_edge_type_error():
    # n is not an integer (float, string, None) should raise TypeError
    with pytest.raises(TypeError):
        cividis("10") # 1.84μs -> 2.00μs (7.81% slower)
    with pytest.raises(TypeError):
        cividis(None) # 900ns -> 926ns (2.81% slower)
    with pytest.raises(TypeError):
        cividis(5.5) # 4.02μs -> 3.62μs (11.1% faster)

def test_cividis_edge_maximum():
    # n=256 returns full palette, n=255 returns palette of length 255
    codeflash_output = cividis(256); result_256 = codeflash_output # 58.1μs -> 49.1μs (18.4% faster)
    codeflash_output = cividis(255); result_255 = codeflash_output # 31.0μs -> 21.8μs (42.1% faster)

def test_cividis_edge_duplicates():
    # For n=256, all colors should be unique
    codeflash_output = cividis(256); result = codeflash_output # 42.6μs -> 33.1μs (28.8% faster)

def test_cividis_edge_monotonic():
    # The palette should be monotonic in index, i.e. colors should be in order
    codeflash_output = cividis(10); result = codeflash_output # 21.6μs -> 21.4μs (0.672% faster)
    indices = [Cividis256.index(c) for c in result]

# --- Large Scale Test Cases ---

def test_cividis_large_scale_100():
    # Test with n=100 (large but <256)
    codeflash_output = cividis(100); result = codeflash_output # 30.4μs -> 27.1μs (12.1% faster)

def test_cividis_large_scale_999():
    # Test with n=999 (just below overflow)
    with pytest.raises(ValueError):
        cividis(999) # 1.47μs -> 1.58μs (7.10% slower)

def test_cividis_large_scale_performance():
    # Performance: n=128, should return quickly and correctly
    codeflash_output = cividis(128); result = codeflash_output # 37.7μs -> 32.8μs (14.9% faster)
    # Check that all colors are valid hex strings
    for color in result:
        pass

def test_cividis_large_scale_even_spacing():
    # For n=16, colors should be spaced evenly in the palette
    codeflash_output = cividis(16); result = codeflash_output # 23.4μs -> 23.6μs (0.581% slower)
    indices = [Cividis256.index(c) for c in result]
    # The difference between consecutive indices should be close to 255/15
    diffs = [indices[i+1] - indices[i] for i in range(len(indices)-1)]
    avg_diff = sum(diffs) / len(diffs)

def test_cividis_large_scale_first_last():
    # For n=256, first and last colors must match palette
    codeflash_output = cividis(256); result = codeflash_output # 42.9μs -> 33.7μs (27.5% faster)

# --- Additional Edge Cases ---

def test_cividis_edge_one_less():
    # n=255 should not include the last color twice
    codeflash_output = cividis(255); result = codeflash_output # 42.5μs -> 34.0μs (25.1% faster)

def test_cividis_edge_palette_unchanged():
    # cividis should not modify the global palette
    before = Cividis256[:]
    codeflash_output = cividis(10); _ = codeflash_output # 21.2μs -> 22.0μs (3.86% slower)
    after = Cividis256[:]

def test_cividis_edge_return_type():
    # Should always return a tuple
    codeflash_output = cividis(5); result = codeflash_output # 20.5μs -> 21.1μs (3.08% slower)

def test_cividis_edge_empty_tuple():
    # n=0 and n<0 should return empty tuple, not None or other type
    codeflash_output = cividis(0)
    codeflash_output = cividis(-1)

def test_cividis_edge_n_is_bool():
    # n=True/False should act as 1/0
    codeflash_output = cividis(True) # 38.6μs -> 38.8μs (0.531% slower)
    codeflash_output = cividis(False) # 9.07μs -> 9.51μs (4.61% slower)

# --- Mutation Sensitivity ---

def test_cividis_mutation_sensitivity():
    # Changing the palette order should fail this test
    codeflash_output = cividis(6); result = codeflash_output # 22.9μs -> 23.4μs (2.40% slower)
    expected = ('#00204C', '#31446B', '#666870', '#958F78', '#CAB969', '#FFE945')

def test_cividis_mutation_spacing():
    # If linear_palette changes spacing logic, this test will fail
    codeflash_output = cividis(3); result = codeflash_output # 21.0μs -> 21.6μs (2.80% slower)
    expected = (Cividis256[0], Cividis256[127], Cividis256[255])
# codeflash_output is used to check that the output of the original code is the same as that of the optimized code.
#------------------------------------------------
import pytest  # used for our unit tests
from bokeh.palettes import cividis

Cividis256 = (
    '#00204C', '#00204E', '#002150', '#002251', '#002353', '#002355', '#002456', '#002558', '#00265A', '#00265B', '#00275D', '#00285F',
    '#002861', '#002963', '#002A64', '#002A66', '#002B68', '#002C6A', '#002D6C', '#002D6D', '#002E6E', '#002E6F', '#002F6F', '#002F6F',
    '#00306F', '#00316F', '#00316F', '#00326E', '#00336E', '#00346E', '#00346E', '#01356E', '#06366E', '#0A376D', '#0E376D', '#12386D',
    '#15396D', '#17396D', '#1A3A6C', '#1C3B6C', '#1E3C6C', '#203C6C', '#223D6C', '#243E6C', '#263E6C', '#273F6C', '#29406B', '#2B416B',
    '#2C416B', '#2E426B', '#2F436B', '#31446B', '#32446B', '#33456B', '#35466B', '#36466B', '#37476B', '#38486B', '#3A496B', '#3B496B',
    '#3C4A6B', '#3D4B6B', '#3E4B6B', '#404C6B', '#414D6B', '#424E6B', '#434E6B', '#444F6B', '#45506B', '#46506B', '#47516B', '#48526B',
    '#49536B', '#4A536B', '#4B546B', '#4C556B', '#4D556B', '#4E566B', '#4F576C', '#50586C', '#51586C', '#52596C', '#535A6C', '#545A6C',
    '#555B6C', '#565C6C', '#575D6D', '#585D6D', '#595E6D', '#5A5F6D', '#5B5F6D', '#5C606D', '#5D616E', '#5E626E', '#5F626E', '#5F636E',
    '#60646E', '#61656F', '#62656F', '#63666F', '#64676F', '#65676F', '#666870', '#676970', '#686A70', '#686A70', '#696B71', '#6A6C71',
    '#6B6D71', '#6C6D72', '#6D6E72', '#6E6F72', '#6F6F72', '#6F7073', '#707173', '#717273', '#727274', '#737374', '#747475', '#757575',
    '#757575', '#767676', '#777776', '#787876', '#797877', '#7A7977', '#7B7A77', '#7B7B78', '#7C7B78', '#7D7C78', '#7E7D78', '#7F7E78',
    '#807E78', '#817F78', '#828078', '#838178', '#848178', '#858278', '#868378', '#878478', '#888578', '#898578', '#8A8678', '#8B8778',
    '#8C8878', '#8D8878', '#8E8978', '#8F8A78', '#908B78', '#918C78', '#928C78', '#938D78', '#948E78', '#958F78', '#968F77', '#979077',
    '#989177', '#999277', '#9A9377', '#9B9377', '#9C9477', '#9D9577', '#9E9676', '#9F9776', '#A09876', '#A19876', '#A29976', '#A39A75',
    '#A49B75', '#A59C75', '#A69C75', '#A79D75', '#A89E74', '#A99F74', '#AAA074', '#ABA174', '#ACA173', '#ADA273', '#AEA373', '#AFA473',
    '#B0A572', '#B1A672', '#B2A672', '#B4A771', '#B5A871', '#B6A971', '#B7AA70', '#B8AB70', '#B9AB70', '#BAAC6F', '#BBAD6F', '#BCAE6E',
    '#BDAF6E', '#BEB06E', '#BFB16D', '#C0B16D', '#C1B26C', '#C2B36C', '#C3B46C', '#C5B56B', '#C6B66B', '#C7B76A', '#C8B86A', '#C9B869',
    '#CAB969', '#CBBA68', '#CCBB68', '#CDBC67', '#CEBD67', '#D0BE66', '#D1BF66', '#D2C065', '#D3C065', '#D4C164', '#D5C263', '#D6C363',
    '#D7C462', '#D8C561', '#D9C661', '#DBC760', '#DCC860', '#DDC95F', '#DECA5E', '#DFCB5D', '#E0CB5D', '#E1CC5C', '#E3CD5B', '#E4CE5B',
    '#E5CF5A', '#E6D059', '#E7D158', '#E8D257', '#E9D356', '#EBD456', '#ECD555', '#EDD654', '#EED753', '#EFD852', '#F0D951', '#F1DA50',
    '#F3DB4F', '#F4DC4E', '#F5DD4D', '#F6DE4C', '#F7DF4B', '#F9E049', '#FAE048', '#FBE147', '#FCE246', '#FDE345', '#FFE443', '#FFE542',
    '#FFE642', '#FFE743', '#FFE844', '#FFE945')
from bokeh.palettes import cividis

# unit tests

# --- Basic Test Cases ---

def test_cividis_basic_small_palette():
    # Test that cividis(6) returns the documented expected values
    codeflash_output = cividis(6) # 21.3μs -> 21.5μs (1.11% slower)

def test_cividis_basic_single_color():
    # Test that cividis(1) returns just the first color
    codeflash_output = cividis(1) # 21.2μs -> 22.7μs (6.53% slower)

def test_cividis_basic_two_colors():
    # Test that cividis(2) returns first and last color
    codeflash_output = cividis(2) # 20.5μs -> 20.9μs (2.05% slower)

def test_cividis_basic_three_colors():
    # Test that cividis(3) returns first, middle, last color
    codeflash_output = cividis(3) # 20.2μs -> 20.9μs (3.37% slower)

def test_cividis_basic_four_colors():
    # Test that cividis(4) returns first, 1/3, 2/3, last color
    codeflash_output = cividis(4) # 19.5μs -> 20.7μs (5.83% slower)

def test_cividis_basic_full_palette():
    # Test that cividis(256) returns the full Cividis256 palette
    codeflash_output = cividis(256) # 43.1μs -> 34.0μs (26.9% faster)

def test_cividis_basic_return_type_and_length():
    # Test that cividis returns a tuple of length n
    for n in [1, 2, 10, 100, 256]:
        codeflash_output = cividis(n); result = codeflash_output # 84.0μs -> 71.3μs (17.8% faster)

def test_cividis_basic_color_format():
    # Test that all returned colors are hex strings of format '#RRGGBB'
    codeflash_output = cividis(10); result = codeflash_output # 19.6μs -> 19.4μs (0.848% faster)
    for color in result:
        # All characters after # are valid hex digits
        int(color[1:], 16)  # should not raise

# --- Edge Test Cases ---

def test_cividis_edge_zero_colors():
    # Test that cividis(0) returns an empty tuple
    codeflash_output = cividis(0) # 16.5μs -> 17.2μs (4.39% slower)

def test_cividis_edge_n_greater_than_256_raises():
    # Test that n > 256 raises ValueError
    with pytest.raises(ValueError):
        cividis(257) # 1.55μs -> 1.60μs (2.88% slower)
    with pytest.raises(ValueError):
        cividis(1000) # 946ns -> 931ns (1.61% faster)

def test_cividis_edge_negative_n_raises():
    # Test that negative n raises ValueError
    with pytest.raises(ValueError):
        cividis(-1) # 3.91μs -> 3.83μs (2.04% faster)
    with pytest.raises(ValueError):
        cividis(-100) # 1.56μs -> 1.46μs (7.15% faster)

def test_cividis_edge_non_integer_n_raises():
    # Test that non-integer n raises TypeError
    with pytest.raises(TypeError):
        cividis(5.5) # 3.26μs -> 2.97μs (9.87% faster)
    with pytest.raises(TypeError):
        cividis('10') # 1.42μs -> 1.46μs (2.81% slower)
    with pytest.raises(TypeError):
        cividis(None) # 789ns -> 766ns (3.00% faster)

def test_cividis_edge_palette_is_immutable():
    # Test that returned palette is a tuple (immutable)
    codeflash_output = cividis(5); result = codeflash_output # 30.3μs -> 30.7μs (1.35% slower)
    with pytest.raises(AttributeError):
        result.append('#000000')
    with pytest.raises(TypeError):
        result[0] = '#000000'

def test_cividis_edge_colors_are_unique_for_n_256():
    # Test that cividis(256) returns all unique colors
    codeflash_output = cividis(256); result = codeflash_output # 45.5μs -> 36.1μs (26.0% faster)

def test_cividis_edge_colors_are_unique_for_small_n():
    # Test that cividis(n) returns unique colors for small n
    for n in range(1, 10):
        codeflash_output = cividis(n); result = codeflash_output # 76.8μs -> 77.8μs (1.31% slower)

def test_cividis_edge_consistency():
    # Test that repeated calls with same n give same result
    codeflash_output = cividis(7) # 19.4μs -> 18.8μs (3.54% faster)
    codeflash_output = cividis(128) # 20.4μs -> 15.9μs (28.3% faster)

def test_cividis_edge_monotonic_palette():
    # Test that palette colors are monotonic in their indices
    # That is, for n > 1, the i-th color is always from a higher index than the previous
    for n in [2, 5, 10, 256]:
        codeflash_output = cividis(n); result = codeflash_output # 63.3μs -> 54.5μs (16.1% faster)
        indices = [Cividis256.index(color) for color in result]

# --- Large Scale Test Cases ---

def test_cividis_large_scale_1000_colors_raises():
    # Test that requesting a very large palette (n=1000) raises ValueError
    with pytest.raises(ValueError):
        cividis(1000) # 1.57μs -> 1.61μs (2.85% slower)

def test_cividis_large_scale_just_under_limit():
    # Test that requesting just under the limit (n=256) works and is fast
    import time
    start = time.time()
    codeflash_output = cividis(256); result = codeflash_output # 48.4μs -> 38.5μs (25.6% faster)
    duration = time.time() - start

def test_cividis_large_scale_many_calls():
    # Test that calling cividis many times with different n is stable and fast
    for n in range(1, 257, 50):
        codeflash_output = cividis(n); result = codeflash_output # 121μs -> 94.3μs (28.5% faster)
        # All returned colors are in the base palette
        for color in result:
            pass

def test_cividis_large_scale_stress():
    # Stress test: call cividis(100) 100 times and check results
    for _ in range(100):
        codeflash_output = cividis(100); result = codeflash_output # 1.37ms -> 966μs (41.5% faster)

def test_cividis_large_scale_sequential_palettes():
    # Test that increasing n yields palettes with first and last color unchanged
    for n in range(2, 20):
        codeflash_output = cividis(n); result = codeflash_output # 129μs -> 123μs (4.38% faster)
# codeflash_output is used to check that the output of the original code is the same as that of the optimized code.
#------------------------------------------------
from bokeh.palettes import cividis

def test_cividis():
    cividis(0)
🔎 Concolic Coverage Tests and Runtime
Test File::Test Function Original ⏱️ Optimized ⏱️ Speedup
codeflash_concolic_5f34sbte/tmp0x6m3ip7/test_concolic_coverage.py::test_cividis 21.3μs 21.5μs -0.759%⚠️

To edit these changes git checkout codeflash/optimize-cividis-mhbhwclz and push.

Codeflash

@codeflash-ai
Optimize cividis
cc8078a 
**Optimizations**.
- Removed redundant usage of `math.floor` and generator expression in `linear_palette` by leveraging numpy's vectorized operations. This reduces overhead significantly for large palettes/sizes.
- Avoided unnecessary imports (`math`) to speed up startup.
- Ensured new code preserves all comments, type annotations, exception behavior, return values, and original style.
- Retains tuple return type and index selection logic, behaviorally identical.
@codeflash-ai codeflash-ai bot requested a review from mashraf-222 October 29, 2025 04:28
@codeflash-ai codeflash-ai bot added the ⚡️ codeflash Optimization PR opened by Codeflash AI label Oct 29, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@mashraf-222 mashraf-222 Awaiting requested review from mashraf-222

Assignees

No one assigned

Labels

⚡️ codeflash Optimization PR opened by Codeflash AI

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

1 participant