Management library for LHAPDF

This is a pure-python (unofficial) LHAPDF management library ans scripts to download, list and load PDFs using the LHAPDF format. It include a number of utilities to inspect the PDF metadata and the grids, but it is not meant to be a substitute of the LHAPDF library, only of the lhapdf script. For more information about LHAPDF please see the official repository and webpage for LHAPDF and to see their citation policy.

Install

It can be installed from PyPI with a simple command

  pip install lhapdf-management

Features

This library offers a script which aims to be a drop-in replacement of the lhapdf python script (not the library!) In order not to clash with an existing installation of LHAPDF, the script name is lhapdf-management.

Update

Updates the local reference index

  lhapdf-management update

If no installation of LHAPDF is found, the program will fail to do anything, as creating the LHAPDF directory is not the business of the management module. In order to run lhapdf-management and make it believe that LHAPDF already exist a possibility is to populate the LHAPDF_DATA_PATH environment variable, i.e.,

LHAPDF_DATA_PATH=$(python -c 'from pathlib import Path ; from sys import prefix ; print(Path(prefix) / "share" / "LHAPDF")' ; lhapdf-management update

It is also possible to create the directory in the "best guess location" by doing lhapdf_management update --init

List

Lists all available PDFs

  lhapdf-management list [PATTERNS ...] [--installed] [--codes]

Install

Installs a given PDF

  lhapdf-management install <pdf_name> [--upgrade] [--keep]

Open a PDF

It can also be used to programatically get an object pointing to all the right parts of a PDF. No interpolation will be provided by this script at this point and when/if it is ever provided will be done by importing the normal python interface of LHAPDF (and so at that point LHAPDF will need to be installed)

  from lhapdf_management.pdfsets import PDF
  from lhapdf_management.configuration import get_lhapdf_datapath
  data_path = get_lhapdf_datapath()
  pdf = PDF(data_path / "NNPDF31_nnlo_as_0118")
  grids = pdf.get_member_grids(0)

Programatically use the interface

A very useful feature of this library is the possibility of using everything programatically. The three given interfaces (install, list and update) take as input the same arguments as the script interface.

  from lhapdf_management import pdf_install, pdf_list, pdf_update
  list_of_pdfs = pdf_list("--installed")
  pdf_install("NNPDF31_nnlo_as_0118")

Download all PDFs matching pattern

There are some arguments that have been modified for simplicity as their behaviour can also be obtained by other means. For instance, in order to download all PDFs matching a certain pattern one can write the following short bash script:

  for pdf in $(lhapdf-management list NNPDF31*)
  do
    lhapdf-management install ${pdf} --upgrade
  done

Related projects

• A similar project providing rust bindings to LHAPDF is https://github.com/cschwan/managed-lhapdf
• The PDFFlow library https://github.com/N3PDF/pdfflow is a drop-in replacement of the (interpolation) capabilities of LHAPDF using TensorFlow to offer hardware acceleration. It uses lhapdf-management to download the PDF files.
• The NeoPDF library https://github.com/Radonirinaunimi/neopdf provides a drop-in replacement for LHAPDF grid interpolation with extra features for polarized or nuclear PDFs.