Skip to main content

pandas readers on steroids (remote files, glob patterns, cache, etc.)

Project description

Pypi-v Pypi-pyversions Pypi-l Pypi-wheel GitHub Actions codecov

Pea Kina aka 'Giant Panda'

Wrapper around pandas library, which detects separator, encoding and type of the file. It allows to get a group of files with a matching pattern (python or glob regex). It can read both local and remote files (HTTP/HTTPS, FTP/FTPS/SFTP or S3/S3N/S3A).

The supported file types are csv, excel, json, parquet and xml.

:information_source: If the desired type is not yet supported, feel free to open an issue or to directly open a PR with the code !

Please, read the documentation for more information


pip install peakina


Considering a file file.csv


Just type

>>> import peakina as pk
>>> pk.read_pandas('file.csv')
   a  b
0  0  0
1  0  1

Or files on a FTPS server:

  • my_data_2015.csv
  • my_data_2016.csv
  • my_data_2017.csv
  • my_data_2018.csv

You can just type

>>> pk.read_pandas('ftps://<path>/my_data_\\d{4}\\.csv$', match='regex', dtype={'a': 'str'})
    a   b     __filename__
0  '0'  0  'my_data_2015.csv'
1  '0'  1  'my_data_2015.csv'
2  '1'  0  'my_data_2016.csv'
3  '1'  1  'my_data_2016.csv'
4  '3'  0  'my_data_2017.csv'
5  '3'  1  'my_data_2017.csv'
6  '4'  0  'my_data_2018.csv'
7  '4'  1  'my_data_2018.csv'

Using cache

You may want to keep the last result in cache, to avoid downloading and extracting the file if it didn't change:

>>> from peakina.cache import Cache
>>> cache = Cache.get_cache('memory')  # in-memory cache
>>> df = pk.read_pandas('file.csv', expire=3600, cache=cache)

In this example, the resulting dataframe will be fetched from the cache, unless file.csv modification time has changed on disk, or unless the cache is older than 1 hour.

For persistent caching, use: cache = Cache.get_cache('hdf', cache_dir='/tmp')

Use only downloading feature

If you just want to download a file, without converting it to a pandas dataframe:

>>> uri = ''
>>> f = pk.fetch(uri)
>>> f.get_str_mtime()
>>> with as stream:
...     print('Image size:', len(, 'bytes')
Image size: 60284 bytes

Installation on macOS M1 chipset

install everything

brew install hdf5 snappy
HDF5_DIR="/opt/homebrew/Cellar/hdf5/1.12.1/" CPPFLAGS="-I/opt/homebrew/Cellar/snappy/1.1.9/include -L/opt/homebrew/Cellar/snappy/1.1.9/lib" poetry install

For more details, here is what is needed:

install pytables

brew install hdf5
HDF5_DIR="/opt/homebrew/Cellar/hdf5/1.12.1/" poetry run pip install tables

install python-snappy

brew install snappy
CPPFLAGS="-I/opt/homebrew/Cellar/snappy/1.1.9/include -L/opt/homebrew/Cellar/snappy/1.1.9/lib" poetry run pip install python-snappy

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

peakina-0.12.1.tar.gz (21.9 kB view hashes)

Uploaded source

Built Distribution

peakina-0.12.1-py3-none-any.whl (27.1 kB view hashes)

Uploaded py3

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