fdi 1.3.0

Flexible Data Integrator

Flexible Dataset Integrator (FDI) helps data producers and processors to build connections into isolated heterogeneous datasets. to assemble, organize, and integrate data into self-describing, modular, hierarchical, persistent, referenceable Products, whose component datasets keep their own characteristicss and are easily accessible.

FDI provides scripting-friendly APIs and tools to define custom Products and generating Python class files. An integrated Product takes care of inter-platform compatibility, string representation, serialisation to simple exchange format, persistence to disk or server, and carrying references of other Products, enabling context-building and lazy-loading.

FDI’s base data model is defined in sub-package dataset. Persistent data access using storage pools, referencing, and Universal Resource Names, and referencie-carrying product Context, are defined in sub-package pal. A reference RESTful API server designed to communicate with a data processing server/docker using the data model, and a reference HTTP pool server are in sub-package pns.

Install/Uninstall

<a href=”http://mercury.bao.ac.cn:9006/mh/fdi/-/commits/master”><img alt=”pipeline status” src=”http://mercury.bao.ac.cn:9006/mh/fdi/badges/master/pipeline.svg” /></a>

Tip

It is a good idea to add --user at the end or to use a virtualenv to avoid disturbing your system Python setup.

For Users

If you need to use fdi, install from PyPI:

python3 -m pip install fdi

or from git repo:

python3 -m pip install http://mercury.bao.ac.cn:9006/mh/fdi/-/archive/master/fdi-master.tar.gz

If you want to install the develop and server dependencies:

python3 -m pip install fdi[DEV,SERV]

or

python3 -m pip install http://mercury.bao.ac.cn:9006/mh/fdi/-/archive/develop/fdi-develop.tar.gz#egg=fdi[DEV,SERV]

To uninstall:

python3 -m pip uninstall fdi

For Developers and Admins (or Those who are Not Sure which to Choose)

To install

FDIINSTDIR=/tmp   # change this to your installation dir
cd $FDIINSTDIR
git clone http://mercury.bao.ac.cn:9006/mh/fdi.git
cd fdi
git checkout develop
make install I="[DEV,SERV]"

If you want to install the master branch, remove the git checkout develop line above.

To test your installation

make test

Tip

To pass command-line arguments to pytest do, for example,

make test T='-k Bas'

to test BaseProduct.

To Generate Python Product Source Code

Re-generate baseproduct.py and product.py from YAML schema files in fdi/dataset/resources:

make py

Modify/Generate Documents

If you plan to compile documents in the docs directory, generate diagrams, API files, or HTML pages, run (in that order, respectively):

make docs_plots
make docs_api
make docs_html

The generated HTML page is at docs/html/index.html.

Note

https://readthedocs.io makes web pages from sources in docs/sphinx in the repository. Locally generated HTML pages are not on RTD or in the repository. The API files and plots, however, are in api and _static sub-directpries, respectively.

Run Servers for Quick Tests

If you plan to run the pns and/or the http pool server locally, install the dependencies:

python3 -m pip install -e .[SERV]
make installpns

To test your pns servers installation, in one window, run:

make runserver

in another window run:

make testpns

To test your httppool servers installation, in one window, run:

make runpoolserver

in another window run:

make testhttppool

For More

For more examples see tests/test_*.py.

Read more on package introduction, description, quick start, and API documents on readthedocs.io.