Skip to main content

Neptune Fetcher

Project description

Neptune Fetcher

Neptune Fetcher is a read-only API for querying metadata logged with the Neptune Python client. The separation makes it safer and more efficient to fetch data from Neptune.

With the Fetcher API, you can:

  • List experiments, runs, and attributes of a project.
  • Fetch experiment or run metadata as a data frame.
  • Define filters to fetch experiments, runs, and attributes that meet certain criteria.

[!NOTE] For the Python API corresponding to Neptune 2.x, see neptune-client.

Documentation

Installation

pip install -U neptune-fetcher

Set your Neptune API token and project name as environment variables:

export NEPTUNE_API_TOKEN="h0dHBzOi8aHR0cHM.4kl0jvYh3Kb8...ifQ=="
export NEPTUNE_PROJECT="workspace-name/project-name"

For help, see Get started in the Neptune documentation.

Note: To change the token or project, you can set the context directly in the code. This way, you can work with multiple projects at the same time.

Usage

Import the alpha module:

import neptune_fetcher.alpha as npt

To fetch experiment metadata from your project, use the fetch_experiments_table() function.

  • To filter experiments to return, use the experiments parameter.
  • To specify attributes to include as columns, use the attributes parameter.
npt.fetch_experiments_table(
    experiments=["exp_ergwq", "exp_qgguv"],
    attributes=["metrics/train_accuracy", "metrics/train_loss"],
)
           metrics/train_accuracy   metrics/train_loss
                             last                 last
experiment
exp_ergwq                0.278149             0.336344
exp_qgguv                0.160260             0.790268

To fetch values at each step, use fetch_metrics():

npt.fetch_metrics(
    experiments=r"exp.*",
    attributes=r".*metric.*/val_.+",
)
  experiment  step  metrics/val_accuracy  metrics/val_loss
0  exp_dczjz     0              0.432187          0.823375
1  exp_dczjz     1              0.649685          0.971732
2  exp_dczjz     2              0.760142          0.154741
3  exp_dczjz     3              0.719508          0.504652
4  exp_dczjz     4              0.180321          0.503800
5  exp_hgctc     0              0.012390          0.171790
6  exp_hgctc     1              0.392041          0.768675
7  exp_hgctc     2                  None          0.847072
8  exp_hgctc     3              0.479945          0.323537
9  exp_hgctc     4              0.102646          0.055511

You can define detailed criteria for which experiments to search or which attributes to return.

For instructions, see the how-to guides in the Neptune documentation:


Old Fetcher API

For documentation related to the previous version of the Fetcher API, see the docs/old/ directory:

To update your code to the new version, see Migrate to Fetcher Alpha in the Neptune documentation.


License

This project is licensed under the Apache License Version 2.0. For details, see Apache License Version 2.0.

Project details


Release history Release notifications | RSS feed

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

neptune_fetcher-0.17.0.tar.gz (66.9 kB view details)

Uploaded Source

Built Distribution

neptune_fetcher-0.17.0-py3-none-any.whl (104.9 kB view details)

Uploaded Python 3

File details

Details for the file neptune_fetcher-0.17.0.tar.gz.

File metadata

  • Download URL: neptune_fetcher-0.17.0.tar.gz
  • Upload date:
  • Size: 66.9 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.1.0 CPython/3.12.8

File hashes

Hashes for neptune_fetcher-0.17.0.tar.gz
Algorithm Hash digest
SHA256 1c9fdfbe0681ccd12025c79d60c36e3ba4fde653bfe9a4f518b88cece3ff2475
MD5 f64a87877a4519285bc35b18eec9dbad
BLAKE2b-256 451a86aaf0affd4ff472a83a1cc56fbc924b0cdfc7f6bbd3518ba5e72f440b07

See more details on using hashes here.

File details

Details for the file neptune_fetcher-0.17.0-py3-none-any.whl.

File metadata

File hashes

Hashes for neptune_fetcher-0.17.0-py3-none-any.whl
Algorithm Hash digest
SHA256 b23f6607f5de4df1fe2f9c584e8557c3b00ee8b533fa48f90f9f4652275d95a2
MD5 9d9020c044aea26f868d6f499f0bd7e0
BLAKE2b-256 f0ab5427768eb2effb339e92c19e6b5e7a1d8894a59c4cbc704347b11a0a4795

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page