SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2005.00160 · 2020

PipelineProfiler: A Visual Analytics Tool for the Exploration of AutoML Pipelines

arXiv · PDF · Open in the Atlas

Code that ran

18 of 22 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 18 have no contract verdict; 0 failures of our instrument; 4 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2005.00160.json · badge: /api/ran/2005.00160.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
VIDA-NYU/PipelineVis canonical (classifier label)
classifier basis: census
18 ran · 4 no recorded run
FunctionStatusWhere it lives
comm_export_pipelines Ran
ran on a synthesized input
VIDA-NYU/PipelineVis/PipelineProfiler/_plot_pipeline_matrix.py
code served (permissive licence) · get_code("cb700fb13b74cafb")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned dict. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

# The function expects msg to be a dict with a 'pipelines' key
# The value should be a list of pipelines (or any serializable structure)
pipelines = [
    {
        'name': 'pipeline_1',
        'stages': ['stage_a', 'stage_b']
    },
    {
        'name': 'pipeline_2',
        'stages': ['stage_c']
    }
]

msg = {
    'pipelines': pipelines
}

ARGS = {
    'msg': msg
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
compute_edge_type Ran
ran on a synthesized input
VIDA-NYU/PipelineVis/PipelineProfiler/_similarity_flooding.py
code served (permissive licence) · get_code("468aa06640e32104")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned str. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import networkx as nx

torch.manual_seed(0)

# Build a simple directed graph with nodes that have the required 'data' structure
g = nx.DiGraph()

# Add nodes with 'data' attribute containing python_path information
g.add_node(0, data=[{'python_path': 'module.submodule.ClassA'}])
g.add_node(1, data=[{'python_path': 'module.submodule.ClassB'}])
g.add_node(2, data=[{'python_path': 'other.Class'}])

# Add edges to the graph
g.add_edge(0, 1)
g.add_edge(1, 2)

# Define the edge to compute the type for
edge = (0, 1)

ARGS = {
    'g': g,
    'edge': edge
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
networkx, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
create_pairwise_conn_graph Ran
ran on a synthesized input
VIDA-NYU/PipelineVis/PipelineProfiler/_similarity_flooding.py
code served (permissive licence) · get_code("093743a6ad664cab")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned DiGraph. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import networkx as nx

def create_graph_with_nodes_and_edges(num_nodes, num_edges):
    """Helper to create a graph with proper node structure"""
    g = nx.DiGraph()
    
    # Add nodes with the required 'data' structure
    for i in range(num_nodes):
        # Vary python_path to create different node types
        if i == 0:
            python_path = "module.submodule.NodeTypeA"
        elif i == 1:
            python_path = "module.submodule.NodeTypeB"
        else:
            python_path = "module.submodule.NodeTypeC"
        
        g.add_node(i, data=[{'python_path': python_path}])
    
    # Add edges between nodes
    edges_added = 0
    for src in range(num_nodes):
        for dst in range(num_nodes):
            if src != dst and edges_added < num_edges:
                g.add_edge(src, dst)
                edges_added += 1
            if edges_added >= num_edges:
                break
        if edges_added >= num_edges:
            break
    
    return g

# Create two graphs with matching edge types
g1 = create_graph_with_nodes_and_edges(num_nodes=4, num_edges=2)
g2 = create_graph_with_nodes_and_edges(num_nodes=4, num_edges=2)

# Ensure g1 and g2 have matching edge types by controlling their structure
# g1: edges between nodes with specific types
g1 = nx.DiGraph()
g1.add_node(0, data=[{'python_path': 'module.submodule.TypeA'}])
g1.add_node(1, data=[{'python_path': 'module.submodule.TypeB'}])
g1.add_node(2, data=[{'python_path': 'module.submodule.TypeC'}])
g1.add_edge(0, 1)  # TypeA -> TypeB
g1.add_edge(1, 2)  # TypeB -> TypeC

# g2: same edge type pattern
g2 = nx.DiGraph()
g2.add_node(0, data=[{'python_path': 'module.submodule.TypeA'}])
g2.add_node(1, data=[{'python_path': 'module.submodule.TypeB'}])
g2.add_node(2, data=[{'python_path': 'module.submodule.TypeC'}])
g2.add_edge(0, 1)  # TypeA -> TypeB
g2.add_edge(1, 2)  # TypeB -> TypeC

ARGS = {
    'g1': g1,
    'g2': g2,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
networkx — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DiGraph — a DiGraph output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DiGraph output the recorded ndim 2 and dtype family int describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
extract_coef_matrix Ran
ran on a synthesized input
VIDA-NYU/PipelineVis/PipelineProfiler/Experimental/FeatureWeights.py
code served (permissive licence) · get_code("981f88dd9636a989")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

from sklearn import preprocessing
import numpy as np

def extract_primitive_names(pipeline):
    return [s['primitive']['name'] for s in pipeline['steps']]

# Create sample pipelines with the expected structure
pipelines = [
    {
        'steps': [
            {'primitive': {'name': 'preprocessor'}},
            {'primitive': {'name': 'scaler'}},
        ]
    },
    {
        'steps': [
            {'primitive': {'name': 'scaler'}},
            {'primitive': {'name': 'classifier'}},
        ]
    },
    {
        'steps': [
            {'primitive': {'name': 'preprocessor'}},
            {'primitive': {'name': 'classifier'}},
        ]
    },
]

ARGS = {
    'pipelines': pipelines
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
extract_primitive_matrix Ran
ran on a synthesized input
VIDA-NYU/PipelineVis/PipelineProfiler/_powerset_analysis.py
code served (permissive licence) · get_code("02b0b27a25321f61")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np
from sklearn import preprocessing

# Create sample pipelines with the expected structure
pipelines = [
    {
        'steps': [
            {'primitive': {'python_path': 'sklearn.preprocessing.StandardScaler'}},
            {'primitive': {'python_path': 'sklearn.tree.DecisionTreeClassifier'}},
        ]
    },
    {
        'steps': [
            {'primitive': {'python_path': 'sklearn.preprocessing.MinMaxScaler'}},
            {'primitive': {'python_path': 'sklearn.ensemble.RandomForestClassifier'}},
            {'primitive': {'python_path': 'sklearn.preprocessing.StandardScaler'}},
        ]
    },
    {
        'steps': [
            {'primitive': {'python_path': 'sklearn.tree.DecisionTreeClassifier'}},
        ]
    },
]

ARGS = {
    'pipelines': pipelines
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
extract_primitive_names Ran
ran on a synthesized input
VIDA-NYU/PipelineVis/PipelineProfiler/_powerset_analysis.py
code served (permissive licence) · get_code("8f8cace5414ad119")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

def extract_primitive_names(pipeline):
    return [s['primitive']['python_path'] for s in pipeline['steps']]

ARGS = {
    'pipeline': {
        'steps': [
            {
                'primitive': {
                    'python_path': 'module.primitive_1'
                }
            },
            {
                'primitive': {
                    'python_path': 'module.primitive_2'
                }
            },
            {
                'primitive': {
                    'python_path': 'module.primitive_3'
                }
            }
        ]
    }
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
extract_primitive_names Ran
ran on a synthesized input
VIDA-NYU/PipelineVis/PipelineProfiler/Experimental/FeatureWeights.py
code served (permissive licence) · get_code("e83e9f9bff219c2f")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

# Construct a valid pipeline dict that matches the structure expected by extract_primitive_names
pipeline = {
    'steps': [
        {
            'primitive': {
                'name': 'preprocessing'
            }
        },
        {
            'primitive': {
                'name': 'feature_extraction'
            }
        },
        {
            'primitive': {
                'name': 'model_training'
            }
        }
    ]
}

ARGS = {
    'pipeline': pipeline
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
extract_scores Ran
ran on a synthesized input
VIDA-NYU/PipelineVis/PipelineProfiler/_powerset_analysis.py
code served (permissive licence) · get_code("2bdd440663e3b6c7")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

# The function expects pipelines to be a list of dicts
# Each dict should have a 'scores' key that maps to a list of dicts
# Each score dict should have a 'value' key

pipelines = [
    {
        'scores': [
            {'value': 0.85}
        ]
    },
    {
        'scores': [
            {'value': 0.92}
        ]
    },
    {
        'scores': [
            {'value': 0.78}
        ]
    }
]

ARGS = {
    'pipelines': pipelines
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
find_ensemble_weights Ran
ran on a synthesized input
VIDA-NYU/PipelineVis/PipelineProfiler/_import_autosklearn.py
code served (permissive licence) · get_code("78e46d80267263a0")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 1, dtype family float, shape (3,) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

class MockConfig:
    def __init__(self, config_dict):
        self._dict = config_dict
    
    def get_dictionary(self):
        return self._dict

class MockPipeline:
    def __init__(self, config_dict):
        self.config = MockConfig(config_dict)

# Create some sample parameters (can be any hashable objects)
param1 = {"layer": "conv1", "param": "weight"}
param2 = {"layer": "conv2", "param": "bias"}
param3 = {"layer": "fc", "param": "weight"}

all_params = [param1, param2, param3]

# Create models_with_weights list containing (weight, pipeline) tuples
pipeline1 = MockPipeline({"layer": "conv1", "param": "weight"})
pipeline2 = MockPipeline({"layer": "fc", "param": "weight"})

models_with_weights = [
    (0.5, pipeline1),
    (0.3, pipeline2),
]

ARGS = {
    "all_params": all_params,
    "models_with_weights": models_with_weights,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family float, shape (3,) — shape (3,) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
find_metric_name Ran
ran on a synthesized input
VIDA-NYU/PipelineVis/PipelineProfiler/_import_autosklearn.py
code served (permissive licence) · get_code("af652d3f85cb2cd0")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned str. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

class MockAutoML:
    def sprint_statistics(self):
        return """
        Some statistics here
        Metric: AUC
        Other info: value
        """

automl = MockAutoML()

ARGS = {
    'automl': automl
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
find_pipeline_digest Ran
ran on a synthesized input
VIDA-NYU/PipelineVis/PipelineProfiler/_helpers.py
code served (permissive licence) · get_code("02ed464496dd4171")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

# Construct valid arguments for find_pipeline_digest

# The function expects:
# 1. pipelines: a list of dicts, each containing a 'pipeline_digest' key
# 2. digest: a single value or a list of values to search for

pipelines = [
    {'pipeline_digest': 'digest_a', 'name': 'pipeline_1'},
    {'pipeline_digest': 'digest_b', 'name': 'pipeline_2'},
    {'pipeline_digest': 'digest_c', 'name': 'pipeline_3'},
    {'pipeline_digest': 'digest_a', 'name': 'pipeline_4'},
]

digest = 'digest_a'

ARGS = {
    'pipelines': pipelines,
    'digest': digest,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
graph_data_accessor Ran
ran on a synthesized input
VIDA-NYU/PipelineVis/PipelineProfiler/_graph_matching.py
code served (permissive licence) · get_code("171e8b337efff784")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned dict. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

# Construct the arguments for graph_data_accessor
node = {
    'primitive': {
        'python_path': 'some.module.path.ClassA',
        'name': 'node_primitive_name'
    },
    'hyperparams': {
        'learning_rate': 0.001,
        'batch_size': 32,
        'hidden_dim': 64
    }
}

graph_name = 'my_graph'

ARGS = {
    'node': node,
    'graph_name': graph_name
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
id_generator Ran
ran on a synthesized input
VIDA-NYU/PipelineVis/PipelineProfiler/_plot_pipeline_node_link.py
code served (permissive licence) · get_code("9e6f0e878723c786")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned str. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import string
import numpy as np

# Set seed for reproducibility
np.random.seed(0)

# The function id_generator takes one optional parameter: size
# It generates a random string of uppercase ASCII letters of specified length
# The default size is 15, but we can pass any integer value

ARGS = {
    "size": 15
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
import_autosklearn Ran
ran on a synthesized input
VIDA-NYU/PipelineVis/PipelineProfiler/_import_autosklearn.py
code served (permissive licence) · get_code("bbed82fca7a991d2")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

class MockAutoML:
    """Mock AutoML object that simulates auto-sklearn behavior"""
    
    def __init__(self):
        # Create mock cv_results_ with typical structure
        self.cv_results_ = {
            'params': [
                {
                    'balancing:strategy': 'none',
                    'data_preprocessing:__choice__': 'data_preprocessing_choice_0',
                    'feature_preprocessor:__choice__': 'feature_preprocessor_choice_0',
                    'feature_preprocessor:polynomial:degree': 2,
                    'classifier:__choice__': 'classifier_choice_0',
                    'classifier:random_forest:n_estimators': 100,
                },
                {
                    'balancing:strategy': 'weighting',
                    'data_preprocessing:__choice__': 'data_preprocessing_choice_1',
                    'feature_preprocessor:__choice__': 'feature_preprocessor_choice_1',
                    'feature_preprocessor:pca:n_components': 5,
                    'classifier:__choice__': 'classifier_choice_1',
                    'classifier:svm:C': 1.0,
                }
            ],
            'mean_test_score': np.array([0.85, 0.90]),
            'mean_fit_time': np.array([1.5, 2.3])
        }
        
        self._models_with_weights = [
            (0.6, MockPipeline({'balancing:strategy': 'none', 'data_preprocessing:__choice__': 'data_preprocessing_choice_0', 'feature_preprocessor:__choice__': 'feature_preprocessor_choice_0', 'feature_preprocessor:polynomial:degree': 2, 'classifier:__choice__': 'classifier_choice_0', 'classifier:random_forest:n_estimators': 100})),
            (0.4, MockPipeline({'balancing:strategy': 'weighting', 'data_preprocessing:__choice__': 'data_preprocessing_choice_1', 'feature_preprocessor:__choice__': 'feature_preprocessor_choice_1', 'feature_preprocessor:pca:n_components': 5, 'classifier:__choice__': 'classifier_choice_1', 'classifier:svm:C': 1.0}))
        ]
    
    def sprint_statistics(self):
        return "Some Statistics\nMetric: accuracy\nOther info"
    
    def get_models_with_weights(self):
        return self._models_with_weights


class MockPipeline:
    """Mock Pipeline object"""
    
    def __init__(self, config_dict):
        self.config = MockConfig(config_dict)


class MockConfig:
    """Mock Config object"""
    
    def __init__(self, config_dict):
        self._dict = config_dict
    
    def get_dictionary(self):
        return self._dict


ARGS = {
    'automl': MockAutoML(),
    'source': 'auto-sklearn'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
merge_pipeline_files Ran
ran on a synthesized input
VIDA-NYU/PipelineVis/PipelineProfiler/pipeline_merge.py
code served (permissive licence) · get_code("cdf74ce85a1cf951")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import json
import tempfile
import os

# Create temporary files with valid JSON data matching the function's expectations
temp_dir = tempfile.mkdtemp()
pipelines_file = os.path.join(temp_dir, 'pipelines.jsonl')
pipeline_runs_file = os.path.join(temp_dir, 'pipeline_runs.jsonl')

# Write pipelines file
with open(pipelines_file, 'w', encoding='utf8') as f:
    pipeline1 = {
        'digest': 'digest_1',
        'id': 'pipeline_1',
        'source': 'source_1',
        'created': '2023-01-01',
        'inputs': [{'name': 'input_1', 'type': 'table'}],
        'outputs': [{'name': 'output_1', 'type': 'table'}],
        'steps': [{'step_id': 0, 'primitive': 'primitive_1'}]
    }
    f.write(json.dumps(pipeline1) + '\n')
    
    pipeline2 = {
        'digest': 'digest_2',
        'id': 'pipeline_2',
        'source': 'source_2',
        'created': '2023-01-02',
        'inputs': [{'name': 'input_2', 'type': 'table'}],
        'outputs': [{'name': 'output_2', 'type': 'table'}],
        'steps': [{'step_id': 0, 'primitive': 'primitive_2'}]
    }
    f.write(json.dumps(pipeline2) + '\n')

# Write pipeline runs file
with open(pipeline_runs_file, 'w', encoding='utf8') as f:
    run1 = {
        'pipeline': {'digest': 'digest_1'},
        'run': {
            'phase': 'PRODUCE',
            'results': {'scores': [0.95]}
        },
        'problem': 'problem_1',
        'start': '2023-01-01T10:00:00',
        'end': '2023-01-01T10:05:00'
    }
    f.write(json.dumps(run1) + '\n')
    
    run2 = {
        'pipeline': {'digest': 'digest_2'},
        'run': {
            'phase': 'PRODUCE',
            'results': {'scores': [0.87]}
        },
        'problem': 'problem_2',
        'start': '2023-01-02T10:00:00',
        'end': '2023-01-02T10:05:00'
    }
    f.write(json.dumps(run2) + '\n')

ARGS = {
    'pipelines_file': pipelines_file,
    'pipeline_runs_file': pipeline_runs_file,
    'n': -1,
    'verbose': False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
node_type Ran
ran on a synthesized input
VIDA-NYU/PipelineVis/PipelineProfiler/_similarity_flooding.py
code served (permissive licence) · get_code("e257df5f0fbcb877")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned str. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import networkx as nx

torch.manual_seed(0)

# Create a simple directed graph with networkx
g = nx.DiGraph()

# Add a node with the required 'data' structure
# The function accesses node['data'][0]['python_path']
node_data = {
    'python_path': 'module.submodule.NodeType'
}

g.add_node('node_0', data=[node_data])

# Also add another node with a shorter path
node_data_2 = {
    'python_path': 'simple.path'
}

g.add_node('node_1', data=[node_data_2])

ARGS = {
    'g': g,
    'node': 'node_0'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
networkx, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
pipeline_to_graph Ran
ran on a synthesized input
VIDA-NYU/PipelineVis/PipelineProfiler/_graph_matching.py
code served (permissive licence) · get_code("5d81e8b2d80be745")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned DiGraph. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import networkx as nx

ARGS = {
    'pipeline': {
        'pipeline_digest': 'test_pipeline_v1',
        'steps': [
            {
                'primitive': {
                    'python_path': 'sklearn.preprocessing.StandardScaler',
                    'name': 'StandardScaler'
                },
                'hyperparams': {
                    'with_mean': True,
                    'with_std': True
                },
                'arguments': {
                    'inputs': {
                        'data': 'inputs.0'
                    }
                }
            },
            {
                'primitive': {
                    'python_path': 'sklearn.linear_model.LogisticRegression',
                    'name': 'LogisticRegression'
                },
                'hyperparams': {
                    'max_iter': 100
                },
                'arguments': {
                    'inputs': {
                        'data': 'steps.0'
                    }
                }
            }
        ],
        'outputs': [
            {
                'data': 'steps.1'
            }
        ]
    },
    'name': 'my_test_pipeline'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
networkx — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DiGraph — a DiGraph output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DiGraph output the recorded ndim 1 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
python_path_similarity Ran
ran on a synthesized input
VIDA-NYU/PipelineVis/PipelineProfiler/_graph_matching.py
code served (permissive licence) · get_code("64cc10732a5bf54b")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned float, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

# The function takes two python_path strings and compares them
# It checks equality, splits by '.', and compares elements at index 2

ARGS = {
    'python_path1': 'my.module.submodule.function',
    'python_path2': 'my.module.submodule.other_function'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
comm_merge_graphs Not yet run
unverified (harvested, no recorded run)
VIDA-NYU/PipelineVis/PipelineProfiler/_plot_pipeline_matrix.py
code served (permissive licence) · get_code("fd412c5826f4338f")
How it ran

no invocation recorded.

comm_powerset_analysis Not yet run
unverified (harvested, no recorded run)
VIDA-NYU/PipelineVis/PipelineProfiler/_plot_pipeline_matrix.py
code served (permissive licence) · get_code("ce3a9407ef5f2b82")
How it ran

no invocation recorded.

make_html Not yet run
unverified (harvested, no recorded run)
VIDA-NYU/PipelineVis/PipelineProfiler/_plot_pipeline_node_link.py
code served (permissive licence) · get_code("f03844004cdb90bc")
How it ran

no invocation recorded.

setup_comm_colab Not yet run
unverified (harvested, no recorded run)
VIDA-NYU/PipelineVis/PipelineProfiler/_comm_api.py
code served (permissive licence) · get_code("1996a25026aaf84a")
How it ran

no invocation recorded.

Also present in other papers' code

Of 22 samples lifted from this paper's linked repositories, 22 carry a function identity and 0 do not. Of the 22 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 22 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

In recent years, a wide variety of automated machine learning (AutoML) methods have been proposed to search and generate end-to-end learning pipelines. While these techniques facilitate the creation of models for real-world applications, given their black-box nature, the complexity of the underlying algorithms, and the large number of pipelines they derive, it is difficult for their developers to debug these systems. It is also challenging for machine learning experts to select an AutoML system that is well suited for a given problem or class of problems. In this paper, we present the PipelineProfiler, an interactive visualization tool that allows the exploration and comparison of the solution space of machine learning (ML) pipelines produced by AutoML systems. PipelineProfiler is integrated with Jupyter Notebook and can be used together with common data science tools to enable a rich set of analyses of the ML pipelines and provide insights about the algorithms that generated them. We demonstrate the utility of our tool through several use cases where PipelineProfiler is used to better understand and improve a real-world AutoML system. Furthermore, we validate our approach by presenting a detailed analysis of a think-aloud experiment with six data scientists who develop and evaluate AutoML tools.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2005.00160")
get_code_for_paper("2005.00160")
have("2005.00160")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2005.00160.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2005.00160.svg)](https://syntology.ai/paper/2005.00160)

Connect an agent — have() is free.