Skip to content

utils

DirectoryTree dataclass 

DirectoryTree(path: str, files: list[str], children: list[DirectoryTree])

This dataclass is used as Storage backends in Django are not guaranteed to support os.walk type functionality.

Attributes:

Name Type Description
path str

The relative path of the directory as a string.
files str

A list of filenames contained in the directory.
children list[DirectoryTree]

A list of DirectoryTree objects representing child directories.

add_child 

add_child(child_name: str, storage: Storage = default_storage) -> None

Adds a child directory to the tree and recursively adds its children.

Parameters:

Name Type Description Default
child_name str

The name of the child directory.

 required
storage Storage

The storage instance to use.

 default_storage
Source code in src/prune_media/utils.py 
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
def add_child(self, child_name: str, storage: Storage = default_storage) -> None:
    """Adds a child directory to the tree and recursively adds its children.

    Args:
        child_name (str): The name of the child directory.
        storage (Storage): The storage instance to use.
    """
    child_path = (
        f"{self.path}/{child_name}" if self.path not in ["", "/"] else child_name
    )
    child = DirectoryTree(path=child_path, files=[], children=[])
    directories, files = storage.listdir(child_path)
    child.files = files
    for directory in directories:
        child.add_child(directory, storage=storage)
    self.children.append(child)

get_empty_child_directories 

get_empty_child_directories() -> list[str]

Get a list of empty child directories within the tree.

Returns:

Type Description
list[str]

list of empty child directory paths as strings
Source code in src/prune_media/utils.py 
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
def get_empty_child_directories(self) -> list[str]:
    """
    Get a list of empty child directories within the tree.

    Returns:
        list of empty child directory paths as strings
    """
    if len(self.children) == 0:
        return []
    empty_directories = []
    for child in self.children:
        if not child.files and not child.children:
            empty_directories.append(child.path)
        else:
            empty_directories += child.get_empty_child_directories()
    return empty_directories

get_file_paths 

get_file_paths() -> list[str]

Get a list of all file paths within the tree

Returns:

Type Description
list[str]

list of file paths as strings
Source code in src/prune_media/utils.py 
63
64
65
66
67
68
69
70
71
72
73
74
75
def get_file_paths(self) -> list[str]:
    """
    Get a list of all file paths within the tree

    Returns:
        list of file paths as strings
    """
    file_paths = [
        f"{self.path}/{file}" if file != "" else file for file in self.files
    ]
    for child in self.children:
        file_paths += child.get_file_paths()
    return file_paths

FieldListing dataclass 

FieldListing(app_label: str, model_name: str, field_name: str)

Represents a given model field within the Django project.

Attributes:

Name Type Description
app_label str

The app label the field belongs to.
model_name str

The model label that the field belongs to.
field_name str

The field name.

get_all_file_fields 

get_all_file_fields() -> list[FieldListing]

Get all fields in the project where a field is an instance or subclass of django.db.models.FileField.

Returns:

Type Description
list[FieldListing]

list of fields as tuples with "app_label", "model_name", "field_name"
Source code in src/prune_media/utils.py 
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
def get_all_file_fields() -> list[FieldListing]:
    """
    Get all fields in the project where a field is an instance or
    subclass of django.db.models.FileField.

    Returns:
        list of fields as tuples with "app_label",
            "model_name", "field_name"
    """
    file_fields = []
    for app, model_dict in apps.all_models.items():
        if model_dict:
            for model_name, model in model_dict.items():
                for field in model._meta.fields:
                    if isinstance(field, FileField):
                        file_fields.append(
                            FieldListing(
                                app_label=app,
                                model_name=model_name,
                                field_name=field.name,
                            )
                        )
    return file_fields

get_empty_media_directories 

get_empty_media_directories(storage_backend: Storage = default_storage) -> list[str]

Get a list of empty media directories found in the supplied storage instance.

Parameters:

Name Type Description Default
storage_backend Storage

a Django Storage instance

 default_storage

Returns:

Type Description
list[str]

list of empty media directory paths as strings
Source code in src/prune_media/utils.py 
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
def get_empty_media_directories(
    storage_backend: Storage = default_storage,
) -> list[str]:
    """Get a list of empty media directories found in the supplied storage instance.

    Args:
        storage_backend (Storage): a Django Storage instance

    Returns:
        list of empty media directory paths as strings
    """
    dirs, files = storage_backend.listdir(".")
    dir_tree = DirectoryTree(path="", files=files, children=[])
    for directory in dirs:
        dir_tree.add_child(directory, storage=storage_backend)
    return dir_tree.get_empty_child_directories()

get_media_paths 

get_media_paths(storage_backend: Storage = default_storage) -> list[str]

Get a list of all media files found in the supplied storage instance.

Parameters:

Name Type Description Default
storage_backend Storage

a Django Storage instance

 default_storage

Returns:

Type Description
list[str]

list of media file paths as strings
Source code in src/prune_media/utils.py 
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
def get_media_paths(storage_backend: Storage = default_storage) -> list[str]:
    """
    Get a list of all media files found in the supplied storage instance.

    Args:
        storage_backend (Storage): a Django Storage instance

    Returns:
        list of media file paths as strings
    """
    dirs, files = storage_backend.listdir(".")
    dir_tree = DirectoryTree(path="", files=files, children=[])
    for directory in dirs:
        dir_tree.add_child(directory, storage=storage_backend)
    return dir_tree.get_file_paths()

get_referenced_file_paths 

get_referenced_file_paths(fields: list[FieldListing]) -> list[str]

Get a list of all file paths from the supplied field data.

Parameters:

Name Type Description Default
fields list[tuple[str, str, str]]

list of fields as tuples with "app_label", "model_name", "field_name"

 required

Returns:

Type Description
list[str]

list of file paths as strings
Source code in src/prune_media/utils.py 
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
def get_referenced_file_paths(fields: list[FieldListing]) -> list[str]:
    """
    Get a list of all file paths from the supplied field data.

    Args:
        fields (list[tuple[str, str, str]]): list of fields as tuples with "app_label",
             "model_name", "field_name"

    Returns:
        list of file paths as strings
    """
    filepaths = []
    for model_spec in fields:
        model = apps.get_model(
            app_label=model_spec.app_label, model_name=model_spec.model_name
        )
        filepaths += model.objects.filter(
            **{f"{model_spec.field_name}__isnull": False}
        ).values_list(model_spec.field_name, flat=True)
    return filepaths

get_unreferenced_media_paths 

get_unreferenced_media_paths(storage_backend: Storage = default_storage) -> list[str]

Get a list of media files that are not referenced by Django FileFields

Parameters:

Name Type Description Default
storage_backend Storage

a Django Storage instance

 default_storage

Returns:

Type Description
list[str]

list of media file paths as strings
Source code in src/prune_media/utils.py 
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
def get_unreferenced_media_paths(
    storage_backend: Storage = default_storage,
) -> list[str]:
    """Get a list of media files that are not referenced by Django FileFields

    Args:
        storage_backend (Storage): a Django Storage instance

    Returns:
        list of media file paths as strings

    """
    media_paths = get_media_paths(storage_backend=storage_backend)
    return [
        path
        for path in media_paths
        if path not in get_referenced_file_paths(get_all_file_fields())
    ]