Skip to content

SBML Dataloader

This module handles loading and simulating SBML models using PyTorch Lightning.

SBMLDataModule

Bases: LightningDataModule

A LightningDataModule for simulating and loading SBML-based time course data.

Source code in vpeleaderboard/data/src/sbml/sbml_dataloader.py 
 16
 17
 18
 19
 20
 21
 22
 23
 24
 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
class SBMLDataModule(LightningDataModule):
    """
    A LightningDataModule for simulating and loading SBML-based time course data.
    """

    class SBMLTimeCourseDataset(IterableDataset):
        """
        Dataset class for iterating over SBML simulation results.
        """

        def __init__(self, data: Any):
            """
            Args:
                data (Any): Data to be used in the dataset (e.g., pandas DataFrame).
            """
            self.data = data

        def __iter__(self):
            """
            Iterator method for the dataset.

            Yields:
                torch.Tensor: Each row in the dataset as a tensor.
            """
            for _, row in self.data.iterrows():
                yield torch.tensor(row.values, dtype=torch.float)

        def __getitem__(self, index: int):
            raise NotImplementedError("Indexing is not supported for this IterableDataset.")

    def __init__(self, file_name: str):
        """
        Initializes the SBMLDataModule with the given SBML model file name.

        Args:
            file_name (str): The name of the SBML model to load.
        """
        super().__init__()
        self.fine_name = file_name

        self.config = None
        self.sbml_file_path = None
        self.copasi_model: Optional[Any] = None

        self._is_prepared = False
        self._is_setup = False

    def prepare_data(self) -> None:
        """
        Loads YAML config and locates the SBML file.

        This method will locate and load the YAML configuration file, validate its contents,
        and check for the necessary SBML model file.

        Raises:
            FileNotFoundError: If the SBML model or YAML config file is not found.
            ValueError: If the YAML config file is empty or missing required keys.
        """
        current_file_path = os.path.abspath(__file__)
        project_root_dir = current_file_path
        # Traverse up the directory tree until 'vpeleaderboard' is found
        while os.path.basename(project_root_dir) != "vpeleaderboard" and \
              project_root_dir != os.path.dirname(project_root_dir):
            project_root_dir = os.path.dirname(project_root_dir)
        if os.path.basename(project_root_dir) != "vpeleaderboard":
            raise RuntimeError(
                f"Could not find 'vpeleaderboard' directory \
                in the parent path of {current_file_path}. "
                "Ensure the script is run from within the vpeleaderboard project structure."
            )
        current_script_dir_abs = os.path.dirname(os.path.abspath(__file__))
        configs_dir_abs = os.path.join(project_root_dir, "configs")
        hydra_config_path = os.path.relpath(configs_dir_abs, current_script_dir_abs)
        if not os.path.isdir(configs_dir_abs):
            raise FileNotFoundError(
                f"Hydra config directory not found: {configs_dir_abs}. "
                "Please ensure the 'configs' directory exists at the project root."
            )
        with hydra.initialize(version_base=None, config_path=hydra_config_path):
            cfg: DictConfig = hydra.compose(config_name="config")

        model_base_dir = os.path.join(project_root_dir, cfg.model_directory)
        model_base_dir = os.path.abspath(model_base_dir)
        sbml_path = os.path.join(model_base_dir, f"{self.fine_name}.xml")
        if not os.path.exists(sbml_path):
            raise FileNotFoundError(
                f"SBML model .xml file not found for model '{self.fine_name}'. "
                f"Expected at: {sbml_path}"
            )
        self.sbml_file_path = sbml_path

        yaml_file = os.path.join(
            project_root_dir,
            "configs",
            "data",
            "sbml",
            self.fine_name, # This is the specific model's subfolder
            "default.yaml"   # The default configuration file within that folder
        )
        if not os.path.exists(yaml_file):
            raise FileNotFoundError(
                f"YAML config file not found for model '{self.fine_name}'. "
                f"Expected at: {yaml_file}"
            )
        with open(yaml_file, 'r', encoding='utf-8') as file:
            self.config = yaml.safe_load(file)
        if not self.config:
            raise ValueError(f"YAML config {yaml_file} is empty or malformed.")

        required_keys = ['train_duration', 'test_duration', 'val_duration']
        missing_keys = [key for key in required_keys
                        if key not in self.config or self.config[key] is None]
        if missing_keys:
            raise ValueError(
                f"Missing required key(s) in YAML config {yaml_file}: {', '.join(missing_keys)}"
            )

        self._is_prepared = True

    def setup(self, stage: Optional[str] = None) -> None:
        """
        Loads the SBML model only after prepare_data has been called.

        Args:
            stage (Optional[str]): The stage of setup, typically used in multi-stage setups.

        Raises:
            RuntimeError: If `prepare_data()` has not been called before `setup()`.
            FileNotFoundError: If the SBML file is not found.
        """
        if not self._is_prepared:
            raise RuntimeError("You must call `prepare_data()` before `setup()`.")

        if not os.path.exists(self.sbml_file_path):
            raise FileNotFoundError(f"SBML model file not found: {self.sbml_file_path}")

        self.copasi_model = basico.load_model(self.sbml_file_path)

    def train_dataloader(self) -> DataLoader:
        """
        Creates the DataLoader for the training dataset based on the SBML simulation results.

        Returns:
            DataLoader: The DataLoader for the training dataset.
        """
        train_df = basico.run_time_course(
            duration=self.config['train_duration'],
            use_initial_values=False
        )
        dataset = self.SBMLTimeCourseDataset(train_df)
        return DataLoader(dataset)
    def val_dataloader(self) -> DataLoader:
        """
        Creates the DataLoader for the validating dataset based on the SBML simulation results.

        Returns:
            DataLoader: The DataLoader for the validating dataset.
        """
        val_df = basico.run_time_course(
            duration=self.config['val_duration'],
            use_initial_values=False
        )
        dataset = self.SBMLTimeCourseDataset(val_df)
        return DataLoader(dataset)

    def test_dataloader(self) -> DataLoader:
        """
        Creates the DataLoader for the test dataset based on the SBML simulation results.

        Returns:
            DataLoader: The DataLoader for the test dataset.
        """
        test_df = basico.run_time_course(
            duration=self.config['test_duration'],
            automatic=False,
            use_initial_values=False,
            update_model=False
        )
        dataset = self.SBMLTimeCourseDataset(test_df)
        return DataLoader(dataset)

SBMLTimeCourseDataset

Bases: IterableDataset

Dataset class for iterating over SBML simulation results.

Source code in vpeleaderboard/data/src/sbml/sbml_dataloader.py 
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
class SBMLTimeCourseDataset(IterableDataset):
    """
    Dataset class for iterating over SBML simulation results.
    """

    def __init__(self, data: Any):
        """
        Args:
            data (Any): Data to be used in the dataset (e.g., pandas DataFrame).
        """
        self.data = data

    def __iter__(self):
        """
        Iterator method for the dataset.

        Yields:
            torch.Tensor: Each row in the dataset as a tensor.
        """
        for _, row in self.data.iterrows():
            yield torch.tensor(row.values, dtype=torch.float)

    def __getitem__(self, index: int):
        raise NotImplementedError("Indexing is not supported for this IterableDataset.")

__init__(data)

Parameters:

Name Type Description Default
data Any

Data to be used in the dataset (e.g., pandas DataFrame).

 required
Source code in vpeleaderboard/data/src/sbml/sbml_dataloader.py 
26
27
28
29
30
31
def __init__(self, data: Any):
    """
    Args:
        data (Any): Data to be used in the dataset (e.g., pandas DataFrame).
    """
    self.data = data

__iter__()

Iterator method for the dataset.

Yields:

Type Description

torch.Tensor: Each row in the dataset as a tensor.
Source code in vpeleaderboard/data/src/sbml/sbml_dataloader.py 
33
34
35
36
37
38
39
40
41
def __iter__(self):
    """
    Iterator method for the dataset.

    Yields:
        torch.Tensor: Each row in the dataset as a tensor.
    """
    for _, row in self.data.iterrows():
        yield torch.tensor(row.values, dtype=torch.float)

__init__(file_name)

Initializes the SBMLDataModule with the given SBML model file name.

Parameters:

Name Type Description Default
file_name str

The name of the SBML model to load.

 required
Source code in vpeleaderboard/data/src/sbml/sbml_dataloader.py 
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
def __init__(self, file_name: str):
    """
    Initializes the SBMLDataModule with the given SBML model file name.

    Args:
        file_name (str): The name of the SBML model to load.
    """
    super().__init__()
    self.fine_name = file_name

    self.config = None
    self.sbml_file_path = None
    self.copasi_model: Optional[Any] = None

    self._is_prepared = False
    self._is_setup = False

prepare_data()

Loads YAML config and locates the SBML file.

This method will locate and load the YAML configuration file, validate its contents, and check for the necessary SBML model file.

Raises:

Type Description
FileNotFoundError

If the SBML model or YAML config file is not found.
ValueError

If the YAML config file is empty or missing required keys.
Source code in vpeleaderboard/data/src/sbml/sbml_dataloader.py 
 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
def prepare_data(self) -> None:
    """
    Loads YAML config and locates the SBML file.

    This method will locate and load the YAML configuration file, validate its contents,
    and check for the necessary SBML model file.

    Raises:
        FileNotFoundError: If the SBML model or YAML config file is not found.
        ValueError: If the YAML config file is empty or missing required keys.
    """
    current_file_path = os.path.abspath(__file__)
    project_root_dir = current_file_path
    # Traverse up the directory tree until 'vpeleaderboard' is found
    while os.path.basename(project_root_dir) != "vpeleaderboard" and \
          project_root_dir != os.path.dirname(project_root_dir):
        project_root_dir = os.path.dirname(project_root_dir)
    if os.path.basename(project_root_dir) != "vpeleaderboard":
        raise RuntimeError(
            f"Could not find 'vpeleaderboard' directory \
            in the parent path of {current_file_path}. "
            "Ensure the script is run from within the vpeleaderboard project structure."
        )
    current_script_dir_abs = os.path.dirname(os.path.abspath(__file__))
    configs_dir_abs = os.path.join(project_root_dir, "configs")
    hydra_config_path = os.path.relpath(configs_dir_abs, current_script_dir_abs)
    if not os.path.isdir(configs_dir_abs):
        raise FileNotFoundError(
            f"Hydra config directory not found: {configs_dir_abs}. "
            "Please ensure the 'configs' directory exists at the project root."
        )
    with hydra.initialize(version_base=None, config_path=hydra_config_path):
        cfg: DictConfig = hydra.compose(config_name="config")

    model_base_dir = os.path.join(project_root_dir, cfg.model_directory)
    model_base_dir = os.path.abspath(model_base_dir)
    sbml_path = os.path.join(model_base_dir, f"{self.fine_name}.xml")
    if not os.path.exists(sbml_path):
        raise FileNotFoundError(
            f"SBML model .xml file not found for model '{self.fine_name}'. "
            f"Expected at: {sbml_path}"
        )
    self.sbml_file_path = sbml_path

    yaml_file = os.path.join(
        project_root_dir,
        "configs",
        "data",
        "sbml",
        self.fine_name, # This is the specific model's subfolder
        "default.yaml"   # The default configuration file within that folder
    )
    if not os.path.exists(yaml_file):
        raise FileNotFoundError(
            f"YAML config file not found for model '{self.fine_name}'. "
            f"Expected at: {yaml_file}"
        )
    with open(yaml_file, 'r', encoding='utf-8') as file:
        self.config = yaml.safe_load(file)
    if not self.config:
        raise ValueError(f"YAML config {yaml_file} is empty or malformed.")

    required_keys = ['train_duration', 'test_duration', 'val_duration']
    missing_keys = [key for key in required_keys
                    if key not in self.config or self.config[key] is None]
    if missing_keys:
        raise ValueError(
            f"Missing required key(s) in YAML config {yaml_file}: {', '.join(missing_keys)}"
        )

    self._is_prepared = True

setup(stage=None)

Loads the SBML model only after prepare_data has been called.

Parameters:

Name Type Description Default
stage Optional[str]

The stage of setup, typically used in multi-stage setups.

 None

Raises:

Type Description
RuntimeError

If prepare_data() has not been called before setup().
FileNotFoundError

If the SBML file is not found.
Source code in vpeleaderboard/data/src/sbml/sbml_dataloader.py 
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
def setup(self, stage: Optional[str] = None) -> None:
    """
    Loads the SBML model only after prepare_data has been called.

    Args:
        stage (Optional[str]): The stage of setup, typically used in multi-stage setups.

    Raises:
        RuntimeError: If `prepare_data()` has not been called before `setup()`.
        FileNotFoundError: If the SBML file is not found.
    """
    if not self._is_prepared:
        raise RuntimeError("You must call `prepare_data()` before `setup()`.")

    if not os.path.exists(self.sbml_file_path):
        raise FileNotFoundError(f"SBML model file not found: {self.sbml_file_path}")

    self.copasi_model = basico.load_model(self.sbml_file_path)

test_dataloader()

Creates the DataLoader for the test dataset based on the SBML simulation results.

Returns:

Name Type Description
DataLoader DataLoader

The DataLoader for the test dataset.
Source code in vpeleaderboard/data/src/sbml/sbml_dataloader.py 
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
def test_dataloader(self) -> DataLoader:
    """
    Creates the DataLoader for the test dataset based on the SBML simulation results.

    Returns:
        DataLoader: The DataLoader for the test dataset.
    """
    test_df = basico.run_time_course(
        duration=self.config['test_duration'],
        automatic=False,
        use_initial_values=False,
        update_model=False
    )
    dataset = self.SBMLTimeCourseDataset(test_df)
    return DataLoader(dataset)

train_dataloader()

Creates the DataLoader for the training dataset based on the SBML simulation results.

Returns:

Name Type Description
DataLoader DataLoader

The DataLoader for the training dataset.
Source code in vpeleaderboard/data/src/sbml/sbml_dataloader.py 
154
155
156
157
158
159
160
161
162
163
164
165
166
def train_dataloader(self) -> DataLoader:
    """
    Creates the DataLoader for the training dataset based on the SBML simulation results.

    Returns:
        DataLoader: The DataLoader for the training dataset.
    """
    train_df = basico.run_time_course(
        duration=self.config['train_duration'],
        use_initial_values=False
    )
    dataset = self.SBMLTimeCourseDataset(train_df)
    return DataLoader(dataset)

val_dataloader()

Creates the DataLoader for the validating dataset based on the SBML simulation results.

Returns:

Name Type Description
DataLoader DataLoader

The DataLoader for the validating dataset.
Source code in vpeleaderboard/data/src/sbml/sbml_dataloader.py 
167
168
169
170
171
172
173
174
175
176
177
178
179
def val_dataloader(self) -> DataLoader:
    """
    Creates the DataLoader for the validating dataset based on the SBML simulation results.

    Returns:
        DataLoader: The DataLoader for the validating dataset.
    """
    val_df = basico.run_time_course(
        duration=self.config['val_duration'],
        use_initial_values=False
    )
    dataset = self.SBMLTimeCourseDataset(val_df)
    return DataLoader(dataset)