Mozilla Data Collective Python Client Library

The official Python SDK for accessing and contributing to the Mozilla Data Collective platform.

[!WARNING] Our platform is evolving rapidly. Expect breaking changes while the Python SDK is on 0.X.X versions. Please ensure you are always on the latest version available.

Installation

pip install datacollective

Quick Start

IMPORTANT NOTE: Before trying to access any dataset, make sure you have thoroughly read and agreed to the specific dataset's conditions & licensing terms.

1. Get your API key from the Mozilla Data Collective dashboard

2. Set the API key in your environment variable:

Option A: Run this command in your terminal (replace your-api-key-here with your actual API key):

export MDC_API_KEY=your-api-key-here

Option B: Create a .env file in your project directory and add this line:

MDC_API_KEY=your-api-key-here

3. Get your dataset ID from the last section of the dataset URL at the MDC website.

[!TIP] You can find the dataset-id by looking at the URL of the dataset's page on MDC platform. The ID is the unique string of characters located at the very end of the URL, after the /datasets/ path. For example, for URL https://mozilladatacollective.com/datasets/cminc35no007no707hql26lzk dataset id will be cminc35no007no707hql26lzk.

4. Save a dataset locally:

from datacollective import download_dataset

dataset_path = download_dataset("your-dataset-id")

[!NOTE] download_dataset was previously called save_dataset_to_disk. The old name still works for backward compatibility, but it is deprecated and new code should use download_dataset.

[!TIP] Automatic Resume: If a download is interrupted (e.g., due to a network error or it gets stopped it manually), the next time you try download the same dataset at the same folder location, we will automatically resume from where the download left off!

[!TIP] Set enable_logging=True to emit detailed SDK logs to the console and a local log file at ~/.mozdata/datacollective.log with timestamped entries, a per-session id, and retention of 5 backup files at 10 MB each.

5. Get information & metadata about a dataset:

from datacollective import get_dataset_details

details = get_dataset_details("your-dataset-id")

6. Load the dataset into a pandas DataFrame (Alpha version: Only certain MDC datasets are supported right now):

from datacollective import load_dataset

dataset = load_dataset("your-dataset-id")

7. Or load it as a HuggingFace Dataset object (requires the optional hf extra: pip install / uv add "datacollective[hf]"):

from datacollective import load_dataset

dataset = load_dataset("your-dataset-id", return_format="hf")

Returns a Dataset, or a DatasetDict keyed by split name for datasets with multiple splits. See our docs for more details, including how to lazily decode audio with the Audio() feature.

Programmatic submissions and uploads

You can create dataset submissions and upload files with resumable uploads into the MDC platform programmatically using our Python SDK:

from datacollective import DatasetSubmission, License, Task, create_submission_with_upload

submission = DatasetSubmission(
    name="Dataset Name",
    longDescription="A detailed description of the dataset.",
    shortDescription="A brief description of the dataset.",
    locale="en-US",
    task=Task.ASR,
    format="TSV",
    licenseAbbreviation=License.CC_BY_4_0,
    other="This text should provide a detailed description of the dataset, "
          "including its contents, structure, and any relevant information "
          "that would help users understand what the dataset is about "
          "and how it can be used.",
    restrictions="Any restrictions you want to impose on the dataset",
    forbiddenUsage="Use cases that are not allowed with this dataset",
    additionalConditions="Any additional conditions for using the dataset",
    pointOfContactFullName="Jane Doe",
    pointOfContactEmail="jane@example.com",
    fundedByFullName="Funder Name",
    fundedByEmail="funder@example.com",
    legalContactFullName="Legal Name",
    legalContactEmail="legal@example.com",
    createdByFullName="Creator Name",
    createdByEmail="creator@example.com",
    intendedUsage="Describe the intended usage of the dataset, including "
                  "potential applications and use cases.",
    ethicalReviewProcess="Describe the ethical review process that was "
                         "followed for this dataset, including any approvals "
                         "or considerations related to data collection and usage.",
    exclusivityOptOut=False,  # True = This dataset is non-exclusive to Mozilla Data Collective, 
                              # False = Dataset is exclusively hosted in Mozilla Data Collective
    agreeToSubmit=True,  # True = You confirm that you have the right to submit this dataset and 
                         # that all information provided in the datasheet is accurate. 
                         # Required to be True to complete the submission process
)

response = create_submission_with_upload(
    file_path="/path/to/dataset.tar.gz",
    submission=submission
)

print(response)

For predefined licenses, pass licenseAbbreviation=License.<VALUE> and leave licenseUrl and license unset. For custom licenses, pass a custom string to license and optionally include licenseUrl and licenseAbbreviation.

[!TIP] To upload a new .tar.gz version to an already approved dataset, call upload_dataset_file(file_path=..., submission_id=...) directly. Find the submission under Profile → Uploads, open the approved dataset, and copy the value after /profile/submissions/ in the URL. Note that this value is the submission ID, which is different from the public dataset ID.

For more details, visit our docs

License

This project is released under MPL (Mozilla Public License) 2.0.