Easy access to the DES DB. Enhanced command line SQL interpreter client for DES


Keywords
astronomy, interpreter, python, sql-query
License
Other
Install
pip install easyaccess==1.4.7

Documentation

easyaccess latest release License pypi version DOI

Enhanced command line SQL interpreter client for astronomical surveys. help_screen

Description

easyaccess is an enhanced command line interpreter and Python package created to facilitate access to astronomical catalogs stored in SQL Databases. It provides a custom interface with custom commands and was specifically designed to access data from the Dark Energy Survey Oracle database, including autocompletion of tables, columns, users and commands, simple ways to upload and download tables using csv, fits and HDF5 formats, iterators, search and description of tables among others. It can easily be extended to another surveys or SQL databases. The package was completely written in Python and support customized addition of commands and functionalities.

For a short tutorial check here

Current version = 1.4.7

DES DR1 users

For DES public data release, you can start easyaccess with:

easyaccess -s desdr

To create an account click here.

Requirements

  • Oracle Client > 11g.2 (External library, no python) Check here for instructions on how to install these libraries
  • cx_Oracle
    • Note that cx_Oracle needs libaio on some Linux systems
    • Note that cx_Oracle needs libbz2 on some Linux systems
  • fitsio >= 0.9.6
  • pandas >= 0.14
  • numpy
  • termcolor
  • PyTables (optional, for hdf5 output)
  • future (for python 2/3 compatibility)
  • requests
  • gnureadline (optional, for better console behavior in OS X)

Installation

Installing easyaccess can be a little bit tricky given the external libraries required, in particular the Oracle libraries which are free to use. easyaccess is based heavily on the Oracle python client cx_Oracle, you can follow the installation instructions from here. For cx_Oracle to work, you will need the Oracle Instant Client packages which can be obtained from here.

Make sure you have these libraries installed before proceeding to the installation of easyaccess, you can try by opening a Python interpreter and type:

import cx_Oracle

If you have issues, please check the Troubleshooting page or our FAQ page.

Source Installation

You can clone this repository and install easyaccess with:

python setup.py install

Pip installation

easyaccess can also be installed using pip but it'd require the installation of the oracle instant client first as described above

pip install easyaccess==1.4.7

or directly from github:

pip install git+https://github.com/mgckind/easyaccess.git

Conda installation

For Collaborators, now easyaccess can be installed using conda out of the box!

conda install easyaccess==1.4.7 -c mgckind -c anaconda

Docker

For collaborators, We have a Docker image with easyaccess pre-installed which you can obtained from:

docker pull mgckind/easyaccess

FAQ

We have a running list of FAQ which we will constantly update, please check here.

Contributing

Please take a look st our Code of Conduct and or contribution guide.

Citation

If you use easyaccess in your work we would encourage to use this reference https://arxiv.org/abs/1810.02721 or copy/paste this BibTeX:

@ARTICLE{2018arXiv181002721C,
       author = {{Carrasco Kind}, M. and {Drlica-Wagner}, A. and {Koziol}, A.~M.~G. and
        {Petravick}, D.},
        title = "{easyaccess: Enhanced SQL command line interpreter for astronomical surveys}",
      journal = {arXiv e-prints},
     keywords = {Astrophysics - Instrumentation and Methods for Astrophysics},
         year = 2018,
        month = Oct,
          eid = {arXiv:1810.02721},
        pages = {arXiv:1810.02721},
archivePrefix = {arXiv},
       eprint = {1810.02721},
 primaryClass = {astro-ph.IM},
       adsurl = {https://ui.adsabs.harvard.edu/\#abs/2018arXiv181002721C},
      adsnote = {Provided by the SAO/NASA Astrophysics Data System}
}

Usage

For a short tutorial and documentation see here, note that not all the features are available for the public use, i.e., DR1 users.

Some great features

  • Nice output format (using pandas)
  • Very flexible configuration
  • Smart tab autocompletion for commands, table names, column names, and file paths
  • Write output results to CSV, TAB, FITS, or HDF5 files
  • Load tables from CSV, FITS or HDF5 files directly into DB (memory friendly by using number of rows or memory limit)
  • Intrinsic DB commands to describe tables, schema, quota, and more
  • Easyaccess can be imported as module from Python with a complete Python API
  • Run commands directly from command line
  • Load SQL queries from a file and/or from the editor
  • Show the execution plan of a query if needed
  • Python functions can be run in a inline query

Interactive interpreter

Assuming that easyaccess is in your path, you can enter the interactive interpreter by calling easyaccess without any command line arguments:

    easyaccess

Running SQL commands

Once inside the interpreter run SQL queries by adding a ";" at the end::

    DESDB ~> select ... from ... where ... ;

To save the results into a table add ">" after the end of the query (after ";") and namefile at the end of line

    DESDB ~> select ... from ... where ... ; > test.fits

The file types supported so far are: .csv, .tab, .fits, and .h5. Any other extension is ignored.

Load tables

To load a table it needs to be in a csv format with columns names in the first row the name of the table is taken from filename or with optional argument --tablename

    DESDB ~> load_table <filename> --tablename <mytable> --chunksize <number of rows to read/upload> --memsize <memory in MB to read at a time>

The --chunsize and --memsize are optional arguments to facilitate uploading big files.

Load SQL queries

To load SQL queries just run:

    DESDB ~> loadsql <filename.sql>

or

    DESDB ~> @filename.sql

The query format is the same as the interpreter, SQL statement must end with ";" and to write output files the query must be followed by " > "

Configuration

The configuration file is located at $HOME/.easyaccess/config.ini but everything can be configured from inside easyaccess type:

    DESDB ~> help config

to see the meanings of all the options, and:

    DESDB ~> config all show

to see the current values, to modify one value, e.g., the prefetch value

    DESDB ~> config prefetch set 50000

and to see any particular option (e.g., timeout):

    DESDB ~> config timeout show

Command line usage

Much of the functionality provided through the interpreter is also available directly from the command line. To see a list of command-line options, use the --help option

    easyaccess --help

Architecture

We have included a simplified UML diagram describing the architecture and dependencies of easyaccess which shows only the different methods for a given class and the name of the file hosting a given class. The main class, easy_or(), inherits all methods from all different subclasses, making this model flexible and extendable to other surveys or databases. These methods are then converted to command line commands and functions that can be called inside easyaccess. Given that there are some DES specific functions, we have moved DES methods into a separate class DesActions().

easyaccess architecture diagram