Skip to main content

btsf (for Binary Time Series File) is a package to store your data in a condensed and fast yet flexible way.

Project description

btfs - Binary Time Series File

This is a module to store time series data in a custom binary file format (.btfs) with the following objectives:

  • small file size
  • appendable
  • fast to read and write
  • easy to adapt to new scenarios
  • ability to store metadata (units, annotations, ...)
  • targeted at Python (adding support for other languages should be rather straight-forward)
  • possibility to tune implementation details such as byte order or padding

Quick description of the file structure of a .btfs file:

File Signature (fixed)
(32 bytes)

at least one intro section
(16-byte header + variable length payload + variable length reserved zero-bytes)

end of intro section header
(16 bytes)

concatenated packed structured data
(structure of fixed length * N, thus appendable!)

A typical use case for btfs is:

  • Time series measurements of one or multiple variables (sampled at the same time).
  • Appending (=adding) new measurements to file is always possible, eg. after a restart of the acquisition.
  • The metadata of the file (stored in the intro sections of the file) can later be updated with additional information, such as data annotations, tags or similar.

What btfs is not suited for, at least not perfectly:

  • event-based data with varying number of values or varying value types.

Note: As the size of the introduction sections cannot be expanded, editing the metadata is only possible within the limits of reserved extra space for the intro section, specified when creating the file.

Installation

pip install --upgrade https://github.com/pklaus/btsf/archive/master.zip

Example Usage

from btsf import Metric, MetricType, BinaryTimeSeriesFile

metrics = [
    Metric('time', MetricType.Double),
    Metric('power', MetricType.Float),
    Metric('counter', MetricType.UInt64),
    Metric('flags', MetricType.UInt8),
]

with BinaryTimeSeriesFile.create('test.btsf', metrics) as f:
    f.append(1.1, 2.2, 0xf0f0f0f0f0f0f0f0, 1)
    f.append(3.3, 4.4, 0xffffffffffffffff, 2)
    f.append(5.0, 1.0, 0x123456789ABCDEF1, 3)

Checking what's in a .btsf file is easy with the supplied CLI:

btsf info test.btsf

resulting in the following output:

test.btsf - Number of entries: 3
Metrics:
time (MetricType.Double)   power (MetricType.Float)   counter (MetricType.UInt64)   flags (MetricType.UInt8)
entries:
(1.1, 2.200000047683716, 17361641481138401520, 1)
(3.3, 4.400000095367432, 18446744073709551615, 2)
(5.0, 1.0, 1311768467463790321, 3)

Project details


Download files

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

Source Distribution

btsf-1.0.1.tar.gz (11.5 kB view hashes)

Uploaded Source

Built Distribution

btsf-1.0.1-py2.py3-none-any.whl (9.3 kB view hashes)

Uploaded Python 2 Python 3

Supported by

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