easyaccess: an enhanced command line SQL interpreter client for astronomical surveys.
Jump to:
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.
To access the DES public data releases, you first need an account, which you can register yourself here. Once you have login credentials for the public DES data server, you can start easyaccess with:
easyaccess -s desdr- 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
We have a running list of FAQ which we will constantly update, please check here.
Please take a look at our Code of Conduct and our contribution guide.
If you use easyaccess in your research, we encourage you 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}
}See also: Carrasco Kind et al., (2019). easyaccess: Enhanced SQL command line interpreter for astronomical surveys. Journal of Open Source Software, 4(33), 1022, https://doi.org/10.21105/joss.01022
Installing easyaccess can be a little bit tricky given the external libraries required, in particular the Oracle libraries which are free to use. If you are primarily interested in using the easyaccess client, we recommend running the Docker image as described below.
Running easyaccess in Docker is easy. Execute the command below to download and run our published image.
$ docker run -it --rm \
hub.ncsa.illinois.edu/des-public/easyaccess:latest \
easyaccess -s desdr
Enter username :
Enter password :
Connecting to DB ** desdr ** ...
Loading metadata into cache...
_______
\ \
// / . .\
// / . _\
// / . / //
\\ \ . / //
\\ \_____/ //
\\_______// DARK ENERGY SURVEY
`-------` DATA MANAGEMENT
easyaccess x.y.z. The DESDM Database shell.
_________
DESDR ~> SELECT RA, DEC, MAG_AUTO_G, TILENAME FROM DR2_MAIN sample(0.001) FETCH FIRST 5 ROWS ONLY ;
RA DEC MAG_AUTO_G TILENAME
1 8.236249 -24.021460 24.450422 DES0032-2415
2 8.084798 -25.715401 26.279263 DES0033-2541
3 8.142266 -35.854926 26.509785 DES0032-3540
4 8.197418 -48.274010 25.243387 DES0030-4831
5 8.107404 -26.313876 24.758778 DES0032-2623Alternatively, you may build and run the image yourself using the included Dockerfile with the commands:
docker build -t des-easyaccess .
docker run -it --rm des-easyaccess easyaccess -s desdreasyaccess 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_OracleIf you have issues, please check the Troubleshooting page or our FAQ page.
You can clone this repository and install easyaccess with:
python setup.py install- 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
- See the
setup.pyfile for additional requirements.
For a short tutorial and documentation see here. Note that not all the features are available for public database use.
Assuming that easyaccess is in your path, you can enter the interactive interpreter by calling easyaccess without any command line arguments:
easyaccessMuch 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 --helpOnce 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.fitsThe file types supported so far are: .csv, .tab, .fits, and .h5. Any other extension is ignored.
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.
To load SQL queries just run:
DESDB ~> loadsql <filename.sql>or
DESDB ~> @filename.sqlThe query format is the same as the interpreter, SQL statement must end with ; and to write output files the query must be followed by a standard output redirect command > outfile
The configuration file is located at $HOME/.easyaccess/config.ini but everything can be configured from inside easyaccess type:
DESDB ~> help configto see the meanings of all the options, and:
DESDB ~> config all showto see the current values, to modify one value, e.g., the prefetch value
DESDB ~> config prefetch set 50000and to see any particular option (e.g., timeout):
DESDB ~> config timeout showWe 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().
Create a git tag with the semantic version number and make a release in the GitHub repo based on that git tag.
Run the release.sh shell script to check that all steps have been executed.
-
Make sure version_tag in
./easyaccess/version.pymatches git repo tagversion_tag = (x, y, z) -
Make sure you have twine installed on system
pip install twine -
In repository directory:
- Create tar.gz source file:
python3 setup.py sdist- Create py2-py3 build wheel
python3 setup.py bdist_wheel --universal- Upload build and source files to PyPi
twine upload dist/*