neptune-fetcher

What it is and what it does

Neptune Fetcher is a read-only API client for retrieving experiment metadata, run data, and logged attributes from Neptune ML tracking projects. It converts query results into pandas DataFrames, making it easy to analyze experiment outcomes, compare runs, and extract metrics for downstream analysis. The package depends on azure-storage-blob, neptune-api, pandas, and tqdm to handle data retrieval and presentation.

However, this package is deprecated and no longer actively maintained. The maintainers explicitly recommend uninstalling it and migrating to neptune-query instead. The package has not been released in over 330 days and is classified as Inactive. Unless you are maintaining legacy code, you should use neptune-query for new projects.

Use it for:

• Export experiment metrics and metadata from Neptune into a pandas DataFrame for offline analysis or reporting.
• Query multiple runs by ID or regex pattern to compare their logged attributes and performance metrics.
• Fetch time-series metric data (e.g., validation accuracy at each training step) for visualization or post-hoc analysis.
• Build read-only dashboards or reports that pull the latest experiment results from a Neptune project.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Queries and fetches experiment metadata, runs, and attributes from Neptune ML tracking projects as pandas DataFrames, with filtering support.

No. This package is deprecated and explicitly superseded by neptune-query. The maintainers recommend uninstalling it immediately and migrating to neptune-query (1.0.0+). Do not use for new projects; only retain if maintaining legacy code that cannot yet be migrated.

Install

neptune-fetcher on PyPI

pip install neptune-fetcher

uv add neptune-fetcher

poetry add neptune-fetcher

Installing neptune-fetcher

Before you install

Low install friction with four straightforward runtime dependencies. However, the package is marked as aging with no release in 330 days and classified as Inactive (Development Status 7). The maintainers explicitly recommend migrating to neptune-query instead.

License in practice

Licensed under Apache License 2.0 (permissive), which allows commercial use, modification, and distribution with minimal restrictions—suitable for most projects.

Quickstart

pip install neptune-fetcher

import neptune_fetcher.alpha as npt

npt.fetch_experiments_table(
    experiments=["exp_ergwq"],
    attributes=["metrics/train_accuracy"]
)

Requires NEPTUNE_API_TOKEN and NEPTUNE_PROJECT environment variables to be set, or passed via context in code.

Verify before relying

• Whether neptune-query is a true drop-in replacement or requires significant code changes beyond the migration guide.
• Current state of the neptune-api dependency and whether it remains actively maintained.

Package facts

Evidence: neptune_fetcher-0.23.0-py3-none-any.whl

Keywords: MLOps, ML Experiment Tracking, ML Model Registry, ML Model Store, ML Metadata Store

Tags

More Python Modules packages