path: root/doc/man/crystfel_geometry.5

.\"
.\" Geometry man page
.\"
.\" Copyright © 2012-2021 Deutsches Elektronen-Synchrotron DESY,
.\"                       a research centre of the Helmholtz Association.
.\"
.\" Part of CrystFEL - crystallography with a FEL
.\"

.TH CRYSTFEL\_GEOMETRY 5
.SH NAME
CRYSTFEL DETECTOR GEOMETRY DESCRIPTION FILES

.SH CRYSTFEL DETECTOR GEOMETRY FILES
The detector geometry is taken from a text file rather than hardcoded into the
program.  Programs which care about the geometry (particularly
\fBindexamajig\fR and \fBpattern_sim\fR) take an argument
\fB--geometry=\fR\fIfilename\fR, where \fIfilename\fR contains the geometry.
.PP
A flexible (and pedantic) representation of the detector has been developed to
avoid all possible sources of ambiguity.  CrystFEL's representation of a
detector is broken down into one or more "panels", each of which has its own
camera length, geometry, resolution and so on.  Each panel fits into an overall
data block taken from the HDF5 file, defined by minimum and maximum coordinates in
the "fast scan" and "slow scan" directions.  "Fast scan" refers to the direction
whose coordinate changes most quickly as the bytes in the HDF5 file are moved
through.  The coordinates are specified inclusively, meaning that a minimum of 0
and a maximum of 9 results in a width of ten pixels.  Counting begins from zero.
All pixels in the image must be assigned to a panel - gaps are not permitted.
.PP
The job of the geometry file is to establish a relationship an array
of pixel values in the HDF5 file, defined in terms only of the "fast scan" and
"slow scan" directions, and the laboratory coordinate system defined as follows:

.IP
+z is the beam direction, and points along the beam (i.e. away from the source)

.IP
+y points towards the zenith (ceiling).

.IP
+x completes the right-handed coordinate system.

.PP
Naively speaking, this means that CrystFEL looks at the images from the "into the
beam" perspective, but please avoid thinking of things in this way.  It's much
better to consider the precise way in which the coordinates are mapped.

The geometry file should contain lines of the following form:

.IP
\fIpanel\fR/clen = 0.560

.PP
\fIpanel\fR can be any name of your choosing.  You can make up names for your panels however you please, as long as the first three letters are not "\fBbad\fR" or "rigid_group" (in lower case), because these are used for special purposes (see below).

.PP
You can also specify values without a panel name, for example:

.IP
clen = 0.560

.PP
In this case, the value will be used for all \fBsubsequent\fR panels appearing in the file which do not have their own specific values for the property, or until you specify another default value further down the file.  Panel-specific values always have priority over default values, and changing the default value has no effect for panels which had already be mentioned at the point in the file where the default value was specified.

.PP
Comments start with a semicolon.  They will be ignored.

.SH TOP-LEVEL ONLY PARAMETERS

The parameters in this section can only appear without a panel name.

.PD 0
.IP "\fBwavelength \fInnn\fR \fB[m|A]"
.IP "\fBphoton_energy \fInnn\fR \fB[eV|keV]"
.IP "\fBelectron_voltage \fInnn\fR \fB[V|kV]"
.PD
These statements specify the incident radiation wavelength.  You must include one (not more) of these statements.  \fBwavelength\fR specifies the wavelength directly, \fBphoton_energy\fR specifies the energy per photon for electromagnetic radiation (e.g. X-rays), and \fBelectron_voltage\fR specifies the accelerating voltage for an electron beam.
.IP
\fInnn\fR can be a literal number, or it can be a header location in the image data file.  In the latter case, the program will do what you expect in the case of multi-frame data files, e.g. a scalar value in the metadata will be applied to all frames, or an array of values can be used to provide a separate wavelength for each frame.
.IP
Units should be specified after the value (or location).  These can be \fBm\fR or \fBA\fR for \fBwavelength\fR, \fBeV\fR or \fBkeV\fR for \fBphoton_energy\fR, and \fBV\fR or \fBkV\fR for \fBelectron_voltage\fR.  For \fBphoton_energy\fR, if no units are given then the value will be taken to be in eV.

.PD 0
.IP "\fBdetector_shift_x = \fInnn \fB[m|mm]"
.IP "\fBdetector_shift_y = \fInnn \fB[m|mm]"
.PD
These specify that the entire detector should be shifted by this amount in the x and y directions.  The units should be specified as \fBm\fR or \fBmm\fR.  If units are not specified, the value will be taken as metres.  \fInnn\fR can be a file metadata location (e.g. an HDF5 path).

.PD 0
.IP "\fBpeak_list = \fIloc"
.PD
This gives the location of the peak list in the data files, for peak detection methods \fBhdf5\fR and \fBcxi\fR (see \fBman indexamajig\fR).

.PD 0
.IP "\fBpeak_list_type = \fIlayout"
.PD
Specify the layout of the peak list.  Allowed values are \fBcxi\fR, \fBlist3\fR and \fBauto\fR.
.IP
\fBlist3\fR expects the peak list to be a two dimensional array whose size in the first dimension equals the number of peaks and whose size in the second dimension is exactly three.  The first two columns contain the fast scan and slow scan coordinates, the third contains the intensities.  This is the correct option for "single-frame" HDF5 files as written by older versions of Cheetah.
.IP
\fBcxi\fR expects the peak list to be a group containing four separate HDF5 datasets: \fBnPeaks\fR, \fBpeakXPosRaw\fR, \fBpeakYPosRaw\fR and \fBpeakTotalIntensity\fR.  See the specification for the CXI file format at http://www.cxidb.org/ for more details.  This is the correct option for "multi-event" HDF5 files as output by recent versions of Cheetah.
.IP
\fBauto\fR tells CrystFEL to decide between the above options based on the file extension.
.IP
Note that CrystFEL considers all peak locations to be distances from the corner of the detector panel, in pixel units, consistent with its description of detector geometry (see 'man crystfel_geometry').  The software which generates the HDF5 or CXI files, including Cheetah, may instead consider the peak locations to be pixel indices in the data array.  To compensate for this discrepancy, CrystFEL will, by default, add 0.5 to all peak coordinates. Use \fB--no-half-pixel-shift\fR if this isn't what you want.

.SH PER-PANEL VALUES

The following parameters can be set for each panel individually.  Don't forget that they can also be used at the "top level" to set default values.

.PD 0
.IP \fBdata\fR
.PD
The location in the HDF5 file of the data block that contains the panel's data. The default value is \fB/data/data\fR.  If the HDF5 file contains multiple events, and each event is stored in a different data block, the variable part of the path can be represented using the % character placeholder.

Example:
.IP
data = /data/%/rawdata

The CrystFEL programs will look for the first event at /data/\fIevent1_name\fR/rawdata, for the second at /data/\fIevent2_name\fR/rawdata, etc., where \fIevent_name\fR and \fIevent2_name\fR are simply whatever the program could find in the HDF5 file which matched the pattern you gave.

.PD 0
.IP \fBdim\fIn\fR\fR
.PD
Information about the layout of the data block identified by the 'data' property. \fIn\fR is an integer number identifying an axis in a multidimensional HDF5 data block. The property value defines the kind of information encoded by the axis. Possible values are:
.RS
.IP %
.PD
event placeholder,the axis encodes events
.IP ss
.PD
the axis encoding the slow scan index
.IP fs
.PD
the axis encodes the fast scan index
.IP \fInumber\fR
.PD
the index in this dimension should be fixed at \fInumber\fR.
.RE
.IP
CrystFEL assumes that the data block defined by the 'data' property has a dimensionality corresponding to the axis with the highest value of \fIn\fR defined by the 'dim' property.  That is, if the geometry file specifies \fBdim0\fR, \fBdim1\fR and \fBdim2\fR, then the data block is expected to be three-dimensional.  The size of the data block along each of those axes comes from the image metadata (e.g. the array sizes in the HDF5 file).
.IP
The lowest number of \fIn\fR corresponds to the \fImost slowly-changing\fR array index as the data block is traversed.  The default values are \fBdim0=ss\fR and \fBdim1=fs\fR.  The value of \fIn\fR corresponding to \fBfs\fR must not be lower than the value assigned to \fBss\fR, i.e. "fast scan is always fast scan".

Example:
.RS
.PD 0
.IP
dim0 = %
.IP
dim1 = 4
.IP
dim2 = ss
.IP
dim3 = fs
.RE
.IP
The above snippet specifies that the data block is 3-dimensional. The first axis represents the event number, the index in the second axis is always 4, and the remaining two axes are the image coordinates.

.PD 0
.IP \fBmin_fs\fR
.IP \fBmin_ss\fR
.IP \fBmax_fs\fR
.IP \fBmax_ss\fR
.PD
The range of pixels in the data block specified by the 'data' property that corresponds to the panel, in fast scan/slow scan coordinates, specified \fBinclusively\fR.

.PD 0
.IP \fBadu_per_eV\fR
.IP \fBadu_per_photon\fR
.PD
The number of detector intensity units (ADU) which will arise from either one electron-Volt of photon energy, or one photon.  This is used to estimate Poisson errors.  Note that setting different values for this parameter for different panels does \fBnot\fR result in the intensities being scaled accordingly when integrating data, but it does affect the intensities calculated by \fBpattern_sim\fR.  You should only specify one out of \fBadu_per_eV\fR and \fBadu_per_photon\fR.

.PD 0
.IP \fBbadrow_direction\fR
.PD
This is the readout direction of a CCD panel, and can be \fBf\fR, \fBs\fR or \fB-\fR.
If more than three peaks are found in the same readout region with similar coordinates perpendicular to the readout direction, they will all be discarded.  This helps to avoid problems due to streaks appearing along the readout direction, which has happened in the past with pnCCD detectors.
If the badrow direction is '-', then the culling will not be performed for this panel.

.PD 0
.IP \fBres\fR
The resolution (in pixels per metre) for this panel.  This is one over the pixel size in metres.

.PD 0
.IP \fBclen\fR
.PD
The camera length (in metres) for this panel. You can also specify the HDF5 path to a floating point data block containing the camera length in millimetres.  For example: "panel0/clen = /LCLS/detectorPosition".  If the HDF5 file contains more than one event, and the data block is scalar, the camera length value
it contains will be used for all events. If, however, the data block is multidimensional and the second dimension is bigger than one, the CrystFEL programs will try to match the content of the data block with the events in the file, assigning the first value in the data block to the first event in the file,
the second value in the data block to the second event in the file, etc. See \fBcoffset\fR as well.

.PD 0
.IP \fBcoffset\fR
.PD
The camera length offset (in metres) for this panel.  This number will be added to the camera length (\fBclen\fR).  This can be useful if the camera length is taken from the HDF5 file and you need to make an adjustment, such as that from a calibration experiment.

.PD 0
.IP \fBfs\fR
.IP \fBss\fR
.PD
For this panel, the fast and slow scan directions correspond to the given directions in the lab coordinate system described above, measured in pixels.  Example: "panel0/fs = 0.5x+0.5y-0.0001z".  Including a component in the z direction means that the panel is not perpendicular to the X-ray beam.

.PD 0
.IP \fBcorner_x\fR
.IP \fBcorner_y\fR
.PD
The corner of this panel, defined as the first point in the panel to appear in the HDF5 file, is now given a position in the lab coordinate system. The units are pixel widths of the current panel.  Note that "first point in the panel" is a conceptual simplification.  We refer to that corner, and to the very corner of the pixel - not, for example, to the centre of the first pixel to appear.

.PD 0
.IP \fBmax_adu\fR
The saturation value for the panel.  You can use this to exclude saturated peaks from the peak search or to avoid integrating saturated reflections.  However, usually it's best to include saturated peaks, and exclude saturated reflections with the --max-adu option of process_hkl and partialator.  Therefore you should avoid setting this parameter - a warning will be displayed if you do.

.PD 0
.IP \fBflag_lessthan
.IP \fBflag_morethan
.IP \fBflag_equal
Mark pixels as "bad" if their values are respectively less than, more than or equal to the given value.  Note carefully that the inequalities are strict, not inclusive: "less than", not "less than or equal to".

.PD 0
.IP \fBmaskN_data\fR
.IP \fBmaskN_file\fR
.IP \fBmaskN_goodbits\fR
.IP \fBmaskN_badbits\fR
.PD
These specify the parameters for bad pixel mask number \fIN\fR.  You can have up to 8 bad pixel masks, numbered from 0 to 7 inclusive.  Placeholders ('%') in the location (\fBmaskN_data\fR) will be substituted with the same values as used for the placeholders in the image data, although there may be fewer of them for the masks than from the image data.
.IP
You can optionally give a filename for each mask with \fBmaskN_file\fR.  The filename may be specified as an absolute filename, or relative to the working directory.  If you don't specify a filename, the mask will be read from the same file as the image data.
.IP
A pixel will be considered bad unless \fIall\fR of the bits which are set in \fBgoodbits\fR are set.  A pixel will also be considered bad if \fIany\fR of the bits which are set in \fBbadbits\fR are set.  Note that pixels can additionally be marked as bad via other mechanisms as well (e.g. \fBno_index\fR or \fBbad\fR).
.IP
Example:
.RS
.PD 0
.IP
mask2_data = /data/bad_pixel_map
.IP
mask2_file = /home/myself/mybadpixels.h5
.IP
mask2_goodbits = 0x00
.IP
mask2_badbits = 0xff
.RE
.PD
.IP
There are some older mask directives which are still understood by this version of CrystFEL.  They are synonyms of the new directives as follows:
.RS
.PD 0
.IP
mask       ----->   mask0_data
.IP
mask_file  ----->   mask0_file
.IP
mask_good  ----->   mask0_goodbits
.IP
mask_bad   ----->   mask0_badbits
.RE

.PD 0
.IP \fBsaturation_map\fR
This specifies the location of the per-pixel saturation map in the HDF5 file.  This works just like \fBmask\fR in that it can come from the current file or a separate one (see \fBsaturation_map_file\fR).  Reflections will be rejected if they contain any pixel above the per-pixel values, \fBin addition\fR to the other checks (see \fBmax_adu\fR).

.PD
.IP \fBsaturation_map_file\fR
Specifies that the saturation map should come from the HDF5 file named here, instead of the HDF5 file being processed.  It can be an absolute filename or relative to the working directory.

.PD 0
.IP \fBno_index\fR
Set this to 1 or "true" to ignore this panel completely.

.PD 0
.IP \fBrail_direction\fR
.IP \fBclen_for_centering\fR
.PD
Specify the direction in which the panel should move when the camera length is increased.  \fBclen_for_centering\fR is the camera length at which the central beam intersects the centre of the detector.  If you've only calibrated the detector at one camera length, perhaps using prior known values for the rail direction, then this should be the camera length at which you calibrated the detector.  \fBclen_for_centering\fR is the camera length \fBbefore\fR applying the \fBcoffset\fR, i.e. for CSPAD/CXI/LCLS data this value should be an "encoder value" in metres. If you specify the rail direction, you must also specify clen_for_centering.  The default is for the panel to move in the +z direction, so rail_direction = z and clen_for_centering is irrelevant.  The modulus of the "rail vector" should normally be equal to one.  Otherwise, the camera length changes from the centering value will be correspondingly scaled (which might sometimes be what you want!).

.SH BAD REGIONS

Bad regions will be completely ignored by CrystFEL.  You can specify the pixels to exclude in pixel units, either in the lab coordinate system (see above) or in fast scan/slow scan coordinates (mixtures are not allowed).   In the latter case, the range of pixels is specified \fIinclusively\fR.  Bad regions are distinguished from normal panels by the fact that they begin with the three letters "bad".
.PP
You can specify a panel name for the bad region, in which case the pixels will only be considered bad if they are within the range you specify \fIand\fR in the panel you specify.  This might be necessary if your HDF5 file layout has overlapping ranges of fs/ss coordinates for different panels (e.g. if the data blocks for the panels are in different HDF5 datasets).

Examples:
.br
.br
badregionA/min_x = -20.0
.br
badregionA/max_x = +20.0
.br
badregionA/min_y = -100.0
.br
badregionA/max_y = +100.0

.br
badregionB/min_fs = 128
.br
badregionB/max_fs = 160
.br
badregionB/min_ss = 256
.br
badregionB/max_ss = 512
.br
badregionB/panel = q0a1


.SH RIGID GROUPS AND RIGID GROUP COLLECTIONS

Some operations in CrystFEL, such as refining the detector geometry, need a group of panels to be treated as a single rigid body.  Such "rigid groups" might describe the fact that certain panels are physically connected to one another, for example, a pair of adjacent ASICs in the CSPAD detector.  Rigid groups can be defined in the geometry file by listing the panels belonging to each group and assigning the group a name, like this:
.RS
.IP "\fBrigid_group_\fIname\fR = \fIpanel1\fR,\fIpanel2\fR"
.RE
.PP
This creates a rigid group called \fIname\fR, containing panels \fIpanel1\fR and \fIpanel2\fR.
.PP
You can specify multiple sets of rigid groups.  For example, as well as specifying the relationships between pairs of ASICs mentioned above, you may also want to specify that certain groups of panels belong to an independently-movable quadrant of the detector.  You can declare and name such "rigid group collections" as follows:
.RS
.IP "\fBrigid_group_collection_\fIname\fR = \fIrigidgroup1\fR,\fIrigidgroup2\fR"
.RE
.PP
This creates a rigid group collection called \fIname\fR, containing rigid groups \fIrigidgroup1\fR and \fIrigidgroup2\fR.
.PP
Definitions of rigid groups and rigid group collections can appear at any place in the geometry file and can be declared using the following global properties.  They are not panel properties, and therefore don't follow the usual panel/property syntax.  You can assign any number of panels to a rigid group, and any number of rigid groups to a rigid group collection.  A panel can be a member of any number of rigid groups.

.PP
See the "examples" folder for some examples (look at the ones ending in .geom).

.SH AUTHOR
This page was written by Thomas White and Valerio Mariani.

.SH REPORTING BUGS
Report bugs to <taw@physics.org>, or visit <http://www.desy.de/~twhite/crystfel>.

.SH COPYRIGHT AND DISCLAIMER
Copyright © 2012-2021 Deutsches Elektronen-Synchrotron DESY, a research centre of the Helmholtz Association.
.P
CrystFEL is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.
.P
CrystFEL is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for more details.
.P
You should have received a copy of the GNU General Public License along with CrystFEL.  If not, see <http://www.gnu.org/licenses/>.

.SH SEE ALSO
.BR crystfel (7),
.BR pattern_sim (1),
.BR indexamajig (1)