odb4py 1.3.4rc1

odb4py is a Python interface for ECMWF ODB1 databases, developed at the Royal Meteorological Institute of Belgium (RMI).

odb4py 1.3.4rc1 release

Description

odb4py (Observation DataBase FOR PYthon) is a C/Python interface to read and query ECMWF ODB1 databases.
It provides high-performance access to ODB data through a native C backend, with seamless integration into the Python scientific ecosystem.
The package embeds a customized version of the ECMWF ODB software odb_api_bundle-0.18.1-Source and is distributed as manylinux wheels, requiring no external ODB installation.

---

Features

• Native C backend based on ECMWF ODB1
• Support for IFS and ARPEGE ODB databases
• SQL-like query interface
• Fast data access with NumPy/C API and pandas integration
• No runtime dependency on system ODB libraries and binaries
• Conversion to other formats : ODB2 ,NetcDF and SQLITE

---

Installation

The odb4py package can be installed from PyPI using pip:

pip install odb4py  

Quick start

 
from odb4py import core      # The extension core
from odb4py import utils     # The helpers module


# Create a connction  
conn = core.odb_open( database  = "/path/to/ODB")   # ECMA or CCMA 

# Select lat/lon and obsvalue
sql_query="SELECT degrees(lat), degrees(lon) , obsvalue FROM hdr, body"

# Parse the query 
p=utils.SqlParser ("/path/to/ODB")

nf  =p.get_nfunc    ( sql_query )
sql =p.clean_string ( sql_query )


# Get the data as a Python dictionary 
data = conn.odb_dict( database  ="/path/to/ODB" ,   #  ODB path
                       sql_query= sql           ,   #  SQL query              
                       nfunc    = nf       )        #  Number of functions in the query
print(data)

For more documentation, please refer to the Official Documentation.

Requirements

Python ≥ 3.9
NumPy ≥ 2.0
Linux system (manylinux2014 compatible)

Optional

pyodc ≥ 1.6.0

Scientific context

ODB (Observation DataBase) is a column-oriented database format developed at ECMWF and widely used in numerical weather prediction systems such as IFS,ARPEGE and NWP limited are models

odb4py is primarily designed for:

• Meteorologists and atmospheric scientists (especially within the ACCORD consortium)
  
• Operational and research environments
• Post-processing and diagnostic workflows
• The current package version focuses on read-only access and data extraction for scientific analysis.

License

Apache License, Version 2.0. See LICENSE for details .

odb4py incorporates components derived from the ECMWF ODB software.

The original source code has been modified to:

• Expose functionality through a Python interface
• Reduce the runtime footprint
• Enable portable binary wheel distribution

All original copyrights remain with ECMWF.

Acknowledgements

• This project incorporates and is derived from the ECMWF ODB software.
  ODB was developed at the European Centre for Medium-Range Weather Forecasts (ECMWF) by S. Saarinen et al. All rights to the original ODB software remains with ECMWF and their respective owners.
• The author would like S. Martinez from Météo France for to provide ARPEGE global model ODB samples. The latters have been used to evaluate the performance of the package.
• The author would also like to thank 'M.O Ait Meziane' for testing odb4py and providing valuable feedback.