Skip to content

Pipeline Module Reference

This section provides the API reference for modules that handle the creation, management, and execution of reproducible data processing workflows. Pipelines support load, read, process, split, export, and write steps.

On This Page

Minimal usage

from datarec.pipeline import Pipeline

pipeline = Pipeline()
pipeline.add_step("load", "registry_dataset", {"dataset_name": "movielens", "version": "1m"})
pipeline.add_step("process", "Binarize", {"threshold": 4})
pipeline.add_step("split", "RandomHoldOut", {"test_ratio": 0.2, "val_ratio": 0.1})
pipeline.add_step("export", "Elliot", {"filename": "elliot/"})
pipeline.apply(output_folder="./out")

Pipeline API

Pipeline

Pipeline for reproducible data transformations.

Source code in datarec/pipeline/pipeline.py 
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
class Pipeline:
    """Pipeline for reproducible data transformations."""
    def __init__(self):
        self.steps: List[PipelineStep] = []

    def __str__(self) -> str:
        """Return a readable summary of pipeline steps."""
        if not self.steps:
            return "Pipeline(steps=0)"
        lines = ["Pipeline:"]
        for idx, step in enumerate(self.steps, start=1):
            params = ", ".join(f"{k}={v!r}" for k, v in step.params.items())
            lines.append(f"{idx}. {step.name} -> {step.operation}({params})")
        return "\n".join(lines)

    def add_step(self, name: str, operation: str, params: Dict[str, Any]) -> None:
        """Add a pipeline step.

        Args:
            name (str): Step name (load/read/process/split/export/write).
            operation (str): Operation or class name for the step.
            params (Dict[str, Any]): Parameters for the step.
        """
        self.steps.append(PipelineStep(name, operation, params))

    def to_yaml(self, file_path: str) -> None:
        """Serialize pipeline to YAML.

        Args:
            file_path (str): Path to the output YAML file.
        """
        with open(file_path, "w") as f:
            yaml.dump({"pipeline": [step.to_dict() for step in self.steps]}, f)

    @classmethod
    def from_yaml(cls, file_path: str) -> "Pipeline":
        """Load a pipeline from YAML.

        Args:
            file_path (str): Path to a YAML pipeline file.

        Returns:
            Pipeline: The loaded pipeline instance.
        """
        with open(file_path, "r") as f:
            config = yaml.safe_load(f)

        pipeline = cls()
        for step in config.get("pipeline", []):
            pipeline.add_step(step['name'], step["operation"], step["params"])

        return pipeline

    def copy(self) -> "Pipeline":
        """Deep-copy the pipeline steps.

        Returns:
            Pipeline: A copy of this pipeline with duplicated steps.
        """
        pipeline = Pipeline()
        pipeline.steps = [step.copy() for step in self.steps]
        return pipeline

    def apply(self, input_folder: Optional[str] = None, output_folder: Optional[str] = None) -> Any:
        """Execute the pipeline.

        Args:
            input_folder (Optional[str]): Base folder for file-based read steps.
            output_folder (Optional[str]): Base folder for export/write steps.

        Returns:
            Any: The final DataRec or split dict, depending on pipeline steps.
        """
        frameworks = {
            'Elliot': 'to_elliot',
            'ClayRS': 'to_clayrs',
            'Cornac': 'to_cornac',
            'DaisyRec': 'to_daisyrec',
            'LensKit': 'to_lenskit',
            'RecBole': 'to_recbole',
            'ReChorus': 'to_rechorus',
            'RecPack': 'to_recpack',
            'Recommenders': 'to_recommenders'
        }
        split_required = {'Elliot', 'ReChorus'}

        if not self.steps:
            raise ValueError("Pipeline is empty. Add at least a load step.")
        if self.steps[0].name not in {'load', 'read'}:
            raise ValueError(f"The first pipeline step must be a load or read, not {self.steps[0].name}")

        print(f"\n\n --- Reproducing Pipeline --- \n\n")

        for step in self.steps:
            print(f"\n--- Step: {step.name} -> {step.operation} ---\n")
            func = self.get_transformation_class(step.name, step.operation)
            if not func:
                raise ValueError(f"Unknown operation: {step.operation}")

            if step.name == 'load':
                result = self._apply_load(step, func, input_folder)

            elif step.name == 'read':
                result = self._apply_read(step, func, input_folder)

            elif step.name == 'export':
                if step.operation not in frameworks:
                    raise ValueError(f"Export step requires a framework operation. Unknown: {step.operation}")
                self._apply_export(step, func, result, output_folder, frameworks, split_required)
                return

            elif step.name == 'write':
                self._apply_write(step, func, result, output_folder)
                return
            else:
                result = self._apply_transform(step, func, result)

        print(f"\n\n --- Finished Pipeline --- \n\n")
        return result

    def _apply_load(self, step: PipelineStep, func, input_folder: Optional[str]) -> Any:
        print(f"Pipeline step {step.name}.")
        print(f"Loading dataset via: {step.operation}.")
        params = dict(step.params)
        if step.operation == "registry_dataset":
            loaded = func(**params)
        else:
            raise ValueError("Load step supports only 'registry_dataset'. Use 'read' for file-based inputs.")
        if hasattr(loaded, "prepare_and_load"):
            return loaded.prepare_and_load()
        if hasattr(loaded, "data"):
            return loaded
        raise ValueError(f"Loader '{step.operation}' did not return a DataRec or an object with prepare_and_load.")

    def _apply_read(self, step: PipelineStep, func, input_folder: Optional[str]) -> Any:
        print(f"Pipeline step {step.name}.")
        print(f"Reading dataset via: {step.operation}.")
        params = dict(step.params)
        if input_folder is None:
            raise ValueError("Read step requires input_folder and 'filename' in params.")
        if "filepath" in params:
            raise ValueError("Read step does not accept 'filepath' in pipelines. Use 'filename' and pass input_folder to apply().")
        filename = params.pop("filename", None)
        if not filename:
            raise ValueError("Read step requires 'filename' in params.")
        params["filepath"] = str(Path(input_folder) / filename)
        loaded = func(**params)
        if hasattr(loaded, "data"):
            return loaded
        raise ValueError(f"Reader '{step.operation}' did not return a DataRec/RawData.")

    def _apply_transform(self, step: PipelineStep, func, result):
        print(f"Pipeline step {step.name}.")
        print(f"Applying {func}.")
        return func(**step.params).run(result)

    def _apply_export(
        self,
        step: PipelineStep,
        func,
        result,
        output_folder: Optional[str],
        frameworks: Dict[str, str],
        split_required: set,
    ) -> None:
        print(f"Pipeline step {step.name}.")
        print(f"Exporting dataset for {step.operation}.")
        params = dict(step.params)
        if output_folder is None:
            if "filename" in params and "output_path" in params:
                raise ValueError("Framework export should not include both 'filename' and 'output_path'.")
            if "filename" in params and "output_path" not in params:
                raise ValueError("Framework export requires output_folder when using 'filename' in params.")
            if "output_path" not in params:
                raise ValueError("Framework export requires 'output_path' or 'filename' in params.")
        else:
            if "output_path" in params:
                raise ValueError("Framework export does not accept 'output_path' when output_folder is provided. Use 'filename'.")
            filename = params.pop("filename", None)
            if not filename:
                raise ValueError("Framework export requires 'filename' in params.")
            params["output_path"] = str(Path(output_folder) / filename)
        exporter = func(**params)
        function_name = frameworks[step.operation]
        if isinstance(result, dict):
            if step.operation in split_required:
                train = result.get('train')
                test = result.get('test')
                val = result.get('val')
                if train is None or test is None:
                    raise ValueError(f"Export '{step.operation}' requires train/test splits.")
                getattr(exporter, function_name)(train, test, val)
            else:
                export_data = result.get('train')
                if export_data is None and len(result) == 1:
                    export_data = next(iter(result.values()))
                if export_data is None:
                    raise ValueError(f"Export '{step.operation}' does not support split dicts without a train split.")
                getattr(exporter, function_name)(export_data)
        else:
            if step.operation in split_required:
                raise ValueError(f"Export '{step.operation}' requires a split result (train/test[/val]).")
            getattr(exporter, function_name)(result)
        return

    def _apply_write(
        self,
        step: PipelineStep,
        func,
        result,
        output_folder: Optional[str],
    ) -> None:
        print(f"Pipeline step {step.name}.")
        print(f"Writing dataset via: {step.operation}.")
        params = dict(step.params)
        if output_folder is None:
            if "filename" in params and "filepath" not in params:
                raise ValueError("Write step requires output_folder when using 'filename' in params.")
            if "filename" in params and "filepath" in params:
                raise ValueError("Write step should not include both 'filename' and 'filepath'.")
            if "filepath" not in params:
                raise ValueError("Write step requires 'filepath' or 'filename' in params.")
            base_path = Path(params.get("filepath"))
        else:
            if "filepath" in params:
                raise ValueError("Write step does not accept 'filepath' when output_folder is provided. Use 'filename'.")
            filename = params.pop("filename", None)
            if not filename:
                raise ValueError("Write step requires 'filename' in params.")
            base_path = Path(output_folder) / filename
        if isinstance(result, dict):
            split_param = params.pop("split", None)
            if split_param:
                split_data = result.get(split_param)
                if split_data is None:
                    raise ValueError(f"Split '{split_param}' not found for writer export.")
                params["filepath"] = str(base_path)
                func(split_data, **params)
            else:
                for split_name, split_data in result.items():
                    split_path = _with_split_suffix(base_path, split_name)
                    params["filepath"] = str(split_path)
                    func(split_data, **params)
        else:
            params["filepath"] = str(base_path)
            func(result, **params)

    @staticmethod
    def get_transformation_class(package_name: str, class_name: str) -> Any:
        """Resolve a pipeline step to its callable class/function.

        Args:
            package_name (str): Pipeline step name (load/process/split/export).
            class_name (str): Class name or special loader name.

        Returns:
            Any: Callable class/function implementing the step.
        """
        mapping = {
            'load': 'datasets',
            'process': 'processing',
            'split': 'splitters',
            'export': 'io',
            'read': 'io',
            'write': 'io',
        }

        if package_name not in mapping.keys():
            raise ValueError(f"Unknown package name '{package_name}'")

        module_name = "datarec." + mapping[package_name]

        if package_name == 'load':
            if class_name == "registry_dataset":
                return _load_registry_dataset
            readers_module = importlib.import_module("datarec.io.readers")
            if hasattr(readers_module, class_name):
                return getattr(readers_module, class_name)
            raise ImportError(
                f"Could not find reader '{class_name}' in datarec.io.readers."
            )
        if package_name == 'read':
            readers_module = importlib.import_module("datarec.io.readers")
            if hasattr(readers_module, class_name):
                return getattr(readers_module, class_name)
            raise ImportError(
                f"Could not find reader '{class_name}' in datarec.io.readers."
            )
        if package_name == 'export':
            module = importlib.import_module(module_name)
            return getattr(module, "FrameworkExporter")
        if package_name == 'write':
            writers_module = importlib.import_module("datarec.io.writers")
            if hasattr(writers_module, class_name):
                return getattr(writers_module, class_name)
            raise ImportError(
                f"Could not find writer '{class_name}' in datarec.io.writers."
            )
        try:
            module = importlib.import_module(module_name)
            return getattr(module, class_name)
        except (ModuleNotFoundError, AttributeError) as e:
            raise ImportError(f"Could not find class {class_name} in module {module_name}") from e

__str__()

Return a readable summary of pipeline steps.

Source code in datarec/pipeline/pipeline.py 
30
31
32
33
34
35
36
37
38
def __str__(self) -> str:
    """Return a readable summary of pipeline steps."""
    if not self.steps:
        return "Pipeline(steps=0)"
    lines = ["Pipeline:"]
    for idx, step in enumerate(self.steps, start=1):
        params = ", ".join(f"{k}={v!r}" for k, v in step.params.items())
        lines.append(f"{idx}. {step.name} -> {step.operation}({params})")
    return "\n".join(lines)

add_step(name, operation, params)

Add a pipeline step.

Parameters:

Name Type Description Default
name str

Step name (load/read/process/split/export/write).

 required
operation str

Operation or class name for the step.

 required
params Dict[str, Any]

Parameters for the step.

 required
Source code in datarec/pipeline/pipeline.py 
40
41
42
43
44
45
46
47
48
def add_step(self, name: str, operation: str, params: Dict[str, Any]) -> None:
    """Add a pipeline step.

    Args:
        name (str): Step name (load/read/process/split/export/write).
        operation (str): Operation or class name for the step.
        params (Dict[str, Any]): Parameters for the step.
    """
    self.steps.append(PipelineStep(name, operation, params))

to_yaml(file_path)

Serialize pipeline to YAML.

Parameters:

Name Type Description Default
file_path str

Path to the output YAML file.

 required
Source code in datarec/pipeline/pipeline.py 
50
51
52
53
54
55
56
57
def to_yaml(self, file_path: str) -> None:
    """Serialize pipeline to YAML.

    Args:
        file_path (str): Path to the output YAML file.
    """
    with open(file_path, "w") as f:
        yaml.dump({"pipeline": [step.to_dict() for step in self.steps]}, f)

from_yaml(file_path) classmethod

Load a pipeline from YAML.

Parameters:

Name Type Description Default
file_path str

Path to a YAML pipeline file.

 required

Returns:

Name Type Description
Pipeline Pipeline

The loaded pipeline instance.
Source code in datarec/pipeline/pipeline.py 
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
@classmethod
def from_yaml(cls, file_path: str) -> "Pipeline":
    """Load a pipeline from YAML.

    Args:
        file_path (str): Path to a YAML pipeline file.

    Returns:
        Pipeline: The loaded pipeline instance.
    """
    with open(file_path, "r") as f:
        config = yaml.safe_load(f)

    pipeline = cls()
    for step in config.get("pipeline", []):
        pipeline.add_step(step['name'], step["operation"], step["params"])

    return pipeline

copy()

Deep-copy the pipeline steps.

Returns:

Name Type Description
Pipeline Pipeline

A copy of this pipeline with duplicated steps.
Source code in datarec/pipeline/pipeline.py 
78
79
80
81
82
83
84
85
86
def copy(self) -> "Pipeline":
    """Deep-copy the pipeline steps.

    Returns:
        Pipeline: A copy of this pipeline with duplicated steps.
    """
    pipeline = Pipeline()
    pipeline.steps = [step.copy() for step in self.steps]
    return pipeline

apply(input_folder=None, output_folder=None)

Execute the pipeline.

Parameters:

Name Type Description Default
input_folder Optional[str]

Base folder for file-based read steps.

 None
output_folder Optional[str]

Base folder for export/write steps.

 None

Returns:

Name Type Description
Any Any

The final DataRec or split dict, depending on pipeline steps.
Source code in datarec/pipeline/pipeline.py 
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
def apply(self, input_folder: Optional[str] = None, output_folder: Optional[str] = None) -> Any:
    """Execute the pipeline.

    Args:
        input_folder (Optional[str]): Base folder for file-based read steps.
        output_folder (Optional[str]): Base folder for export/write steps.

    Returns:
        Any: The final DataRec or split dict, depending on pipeline steps.
    """
    frameworks = {
        'Elliot': 'to_elliot',
        'ClayRS': 'to_clayrs',
        'Cornac': 'to_cornac',
        'DaisyRec': 'to_daisyrec',
        'LensKit': 'to_lenskit',
        'RecBole': 'to_recbole',
        'ReChorus': 'to_rechorus',
        'RecPack': 'to_recpack',
        'Recommenders': 'to_recommenders'
    }
    split_required = {'Elliot', 'ReChorus'}

    if not self.steps:
        raise ValueError("Pipeline is empty. Add at least a load step.")
    if self.steps[0].name not in {'load', 'read'}:
        raise ValueError(f"The first pipeline step must be a load or read, not {self.steps[0].name}")

    print(f"\n\n --- Reproducing Pipeline --- \n\n")

    for step in self.steps:
        print(f"\n--- Step: {step.name} -> {step.operation} ---\n")
        func = self.get_transformation_class(step.name, step.operation)
        if not func:
            raise ValueError(f"Unknown operation: {step.operation}")

        if step.name == 'load':
            result = self._apply_load(step, func, input_folder)

        elif step.name == 'read':
            result = self._apply_read(step, func, input_folder)

        elif step.name == 'export':
            if step.operation not in frameworks:
                raise ValueError(f"Export step requires a framework operation. Unknown: {step.operation}")
            self._apply_export(step, func, result, output_folder, frameworks, split_required)
            return

        elif step.name == 'write':
            self._apply_write(step, func, result, output_folder)
            return
        else:
            result = self._apply_transform(step, func, result)

    print(f"\n\n --- Finished Pipeline --- \n\n")
    return result

get_transformation_class(package_name, class_name) staticmethod

Resolve a pipeline step to its callable class/function.

Parameters:

Name Type Description Default
package_name str

Pipeline step name (load/process/split/export).

 required
class_name str

Class name or special loader name.

 required

Returns:

Name Type Description
Any Any

Callable class/function implementing the step.
Source code in datarec/pipeline/pipeline.py 
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
@staticmethod
def get_transformation_class(package_name: str, class_name: str) -> Any:
    """Resolve a pipeline step to its callable class/function.

    Args:
        package_name (str): Pipeline step name (load/process/split/export).
        class_name (str): Class name or special loader name.

    Returns:
        Any: Callable class/function implementing the step.
    """
    mapping = {
        'load': 'datasets',
        'process': 'processing',
        'split': 'splitters',
        'export': 'io',
        'read': 'io',
        'write': 'io',
    }

    if package_name not in mapping.keys():
        raise ValueError(f"Unknown package name '{package_name}'")

    module_name = "datarec." + mapping[package_name]

    if package_name == 'load':
        if class_name == "registry_dataset":
            return _load_registry_dataset
        readers_module = importlib.import_module("datarec.io.readers")
        if hasattr(readers_module, class_name):
            return getattr(readers_module, class_name)
        raise ImportError(
            f"Could not find reader '{class_name}' in datarec.io.readers."
        )
    if package_name == 'read':
        readers_module = importlib.import_module("datarec.io.readers")
        if hasattr(readers_module, class_name):
            return getattr(readers_module, class_name)
        raise ImportError(
            f"Could not find reader '{class_name}' in datarec.io.readers."
        )
    if package_name == 'export':
        module = importlib.import_module(module_name)
        return getattr(module, "FrameworkExporter")
    if package_name == 'write':
        writers_module = importlib.import_module("datarec.io.writers")
        if hasattr(writers_module, class_name):
            return getattr(writers_module, class_name)
        raise ImportError(
            f"Could not find writer '{class_name}' in datarec.io.writers."
        )
    try:
        module = importlib.import_module(module_name)
        return getattr(module, class_name)
    except (ModuleNotFoundError, AttributeError) as e:
        raise ImportError(f"Could not find class {class_name} in module {module_name}") from e