The parameters INPUT
selects which file is being read from, and similarly for
subcommands that write grids, OUTPUT
determines which file is being written
to. The CLI pineappl
does not care about file suffixes at all, but typically
PineAPPL grids have the file suffix .pineappl
, and if they are compressed
.pineappl.lz4
. If you write a grid and choose a file name with an .lz4
suffix, pineappl
will automatically compress the file using an LZ4
compression. If a compressed grid is read by pineappl
it will automatically
decompress it internally, so that this step does not have to be done by the
user.
A few convolutional subcommands accept a parameter ORDERS
which allows to
select a subset of perturbative orders to be taken into account during the
convolution. This must be a comma separated list of orders, which are of the
form:
asNaM
,aMasN
,asN
oraM
,
where N
must be the exponent of the strong coupling (denoted with as
) and
M
must be the exponent of the electroweak coupling (denoted with a
). For
example, in Drell–Yan lepton-pair production a2,a2as1
selects the leading
order (a2
) and the next-to-leading order QCD (a2as1
).
The parameter PDFSET
that appears for all convolutional-type subcommands
(channels
, convolve
, etc.) must be one of the following strings:
setname/member
: In this casesetname
must be a valid LHAPDF set name andmember
must be the member index. The index0
denotes the central PDF, and ifsetname
is a set that supports the computation of PDF uncertainties the indices1
throughn
denote the uncertainty members. The value ofn
can be read off fromlhapdf show setname
. For example, the stringCT18NNLO/1
selects the first uncertainty member of the NNLO PDF set of CT18.CT18NNLO/0
selects the central member.setname
: This is a special case of the previous specification, where the member with index0
is selected. For example, the stringsCT18NNLO
andCT18NNLO/0
select the same (central) PDF set.LHAID
: This allows to select PDFs using their LHAID, which is an integer. Non-central members are typically denoted by adding their index to the central LHAID. For example,14000
would select the same PDF set asCT18NNLO
and14001
corresponds toCT18NNLO/1
.
If an entire PDF set must be given for the calculation of PDF uncertainties,
that means for for pdfunc
, plot
or pull
, the member selection using /0
,
/1
can be used to show specific members instead of the central prediction.
For instance pineappl pdfunc ... NNPDF40_nnlo_as_01180
calculates the central
value using the average over all replicas, whereas pineappl pdfunc ... NNPDF40_nnlo_as_01180/0
uses the zeroth member. This is especially useful to
show different replicas in plots.
Finally, it is possible to re-label PDF sets by adding =label
to the
specification. This is particularly helpful when plotting predictions with
multiple PDF sets. For example, NNPDF31_nnlo_as_0118_luxqed=NNPDF31luxQED
instructs to use the the PDF set NNPDF31_nnlo_as_0118_luxqed
, but it would be
called NNPDF31luxQED
.
This section specifies the REMAPPING
parameter of pineappl remap
.
For performance/simplicity reasons most Monte Carlo generators neither support
- multi-dimensional distributions nor
- distributions whose bin sizes are not equally sized
during generation. To work around this problem a grid with a one-dimensional
distribution with equally-sized bins can be generated instead, and afterwards
the bins can be 'remapped' to an N-dimensional distribution using the limits
specified with the REMAPPING
string.
The remapping string uses the following special characters to achieve this (note that some of these characters must be escaped in certain shells):
-
,
: The comma constructs 1-dimensional bin limits (1DBL). For example, the 1DBL0,0.2,0.4,0.6,1
expects the grid to have 4 bins whose bin limits will be (0-0.2), (0.2-0.4), (0.4-0.6) and (0.6,1). -
;
: If higher-dimensional bins are needed, n-dimensional bin limits (NDBL) are constructed from a Cartesian product of 1DBL separated with a semicolon. For example,0,0.5,1;0,1,2
expects the grid to have 4 bins, whose 2DBL will be are: (0-0.5;0-1), (0-0.5;1-2), (0.5-1;0-1) and (0.5-1;1-2). -
|
: The previous operators are enough to construct NDBL with differently-sized bins, but they can not construct the following bin limits: (0-1;0-1), (0-1;1-2), (1-2;0-2), (1-2;2-4), (1-2;4-6); here the 1DBL for the second dimension depend on the first dimension and also have a different number of bins. For the first two bins the 1DBL is0,1,2
, but for the last three bins the 1DBL are0,2,4,6
. This can be achieved using the following remapping string:0,1,2;0,1,2|0,2,4,6
. Note that there have to be two 1DBL separated by|
, because the first dimension has two bins. If there are more dimensions and/or bins, the number of 1DBL separated by|
must match this number accordingly. An example of this is the following remapping string:0,1,2;-2,0,2;0,1,2|1,2,3|2,3,4|3,4,5|4,5,6|5,6,7
. Here the third dimension has 6 1DBL separated by|
because the first dimension has 2 bins and the second dimension has 3 bins.If the 1DBL is an empty string, the previous 1DBL is repeated, for example
0,1,2;0,1,2;0,1,2||0,2,4
is shorthand for0,1,2;0,1,2;0,1,2|0,1,2|0,2,4
. -
:
: The last feature of|
can combined with:
, which is used to 'cut' out bins from the left and/or right. For example, the remapping string0,1,2;0,1,2,3:2|:1||:1|2:
is a more succinct way of writing the following remapping string:0,1,2;0,1|0,1,2|0,1,2,3|0,1,2|2,3
.
Finally note that the differential cross sections are calculated using the bin
sizes (the product of bin widths of each dimension) given by the remapping
string. The option --ignore-obs-norm
can be used to remove certain dimensions
from the bin size determination, for example '0,10,20;0,2,4' --ignore-obs-norm 1
will normalize the bins with a size of 2
because the first dimension (with
index 1
) will be ignored