Nanga Parbat: a TMD fitting framework

Nanga Parbat is a fitting framework aimed at the determination of the non-perturbative component of TMD distributions.

Download

You can obtain NangaParbat directly from the github repository:

https://github.com/MapCollaboration/NangaParbat

For the last development branch you can clone the master code:

git clone git@github.com:MapCollaboration/NangaParbat.git

If you instead want to download a specific tag:

git tag -l
git checkout tags/tag_name

Dependencies

In order to install the code a number of external but relatively standard libraries are required. Here is the list:

• yaml-cpp,
• eigen3,
• ceres-solver,
• GSL,
• ROOT with Minuit2,
• LHAPDF6,
• APFEL++.

Most of these libraries can be installed through standard package managers such as Homebrew on MacOS and apt-get on Linux.

NangaParbat also has a lite version, that does not require all the dependencies listed above. The mandatory dependencies are yaml-cpp, eigen3, LHAPDF6, APFEL++.

During the installation, NangaParbat detects if one or more libraries among GSL, ROOT or ceres-solver are missing and installs and compiles only the part of the code that is possible to run.

Installation

The code can be compiled using the following procedure:

cd NangaParbat
mkdir build && cd build
cmake -DCMAKE_INSTALL_PREFIX=/your/installation/path/ ..
make && make install

By the default, if no prefix specification is given, the program will be installed in the /usr/local folder. If you want (or need) to use a different path, remember to export the NangaParbat /lib folder into the LD_LIBRARY_PATH. More configuration options can be accessed by typing:

ccmake .

Running the code

Despite the core of the code is written in C++, the main functionalities can be accessed through the command-line-interface utilities. If the installation was successful, it is possible to run a fit of TMD PDFs just by typing from the main folder:

python3 cli/fit.py

and following the instructions. Any such fit relies on the dataset, the interpolation tables, and the TMD parameterisations currently present in the code. It is however possible to generate new interpolation tables using the cli/tables.py utility that gives the possibility change the theory settings (perturbative order, collinear PDF set, etc.). Including more parameterisations and new experimental datasets is also possible but requires some additional work on the core of the code. Feel free to contact us should you want to extend the code in this respect.

Once the fit is complete (including a number of Monte Carlo replicas), it is possible to create a summary report of the fit by running cli/report.py. This will create a document (in markdown and html format) that collects the main statistical features of the fit, histograms, data-theory comparison plots, and TMD plots.

Documentation

Some further general documentation can be found at the following links:

• command line interface
• utilities
• raw data files
• preprocessed data files
• convolution prefactors

Code documentation generated with Doxygen can be found here.

Public results

The results obtained with NangaParbat for the fit of TMD PDFs to Drell-Yan data can be found at the following links:

• fit at N3LL,
• fit at NNLL',
• fit at NNLL,
• fit at NLL'.

The reports above have been generated using the CLI cli/report.py.

TMD grids

Below you can find a list of TMD and structure functions grids that are compatible with NangaParbat specifications and can be used with the Nanga Parbat interpolation routines. The grids will be made available also through the TMDlib library.

• PV17 fit [Reference: arXiv:1703.10157]: here https://drive.google.com/drive/folders/1cXHZKS0OmY7woUBR2zw_Nf6RmAQYimjl?usp=sharing you can find the grids of the TMD PDFs and FFs, and the SIDIS F_UUT structure functions

Use of TMD grids

In tools/ there are some test codes that can be run by the user in order to interpolate the TMD and structure function grids provided at the link above. This part of NangaParbat is compiled also in the lite version.

The interpolation of the grids can be done with TMDGridInterpolation.cc for TMDs and with StructGridInterpolation.cc for structure functions, while GridsConvolution.cc does the convolution between a TMD PDF grid and a TMD FF grid. The input files, where the user can choose the kinematical points for the interpolation and the convolution, are in tools/inputs/.

References

• Transverse-momentum-dependent parton distributions up to N3LL from Drell-Yan data; Alessandro Bacchetta, Valerio Bertone, Chiara Bissolotti, Giuseppe Bozzi, Filippo Delcarro, Fulvio Piacenza, Marco Radici [arXiv:1912.07550]

Contacts

• Valerio Bertone: valerio.bertone@cern.ch
• Alessandro Bacchetta: alessandro.bacchetta@unipv.it
• Chiara Bissolotti: chiara.bissolotti01@universitadipavia.it

Credits

NangaParbat has been developed with the support of the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation program (grant agreement No. 647981, 3DSPIN).