# Detritus Wiki

### Site Tools

neuroimagen:neuro4.pm

# Modulos del pipeline

Esta pagina es solo de referencia de los modulos Perl que dan soporte al pipeline de neuroimagen.

# NEURO4

This is a set of functions for helping in the pipeline

• print_help

just print the help

this funtions reads the path of a TXT file and print it at STDOUT

usage: print_help(help_file);

• escape_name

This function takes a string and remove some especial characters in order to escape directory names with a lot of strange symbols.

It returns the escaped string

usage: escape_name(string);

• trim

This function takes a string and remove any trailing spaces after and before the text

usage: trim(string);

• check_or_make

This is mostly helpless, just takes a path, checks if exists and create it otherwise

usage: check_or_make(path);

• inplace

This function takes a path and a file name or two paths and returns a string with a single path as result of the concatenation of the first one plus the second one

usage: inplace(path, filename);

This function take the name of a project, reads the configuration file that is located at ~/.config/neuro/ and return every project configuration stored as a hash that can be used at the scripts

• check_subj

Here the fun begins

This function takes as input the name of the project and the subject ID Then it seeks along the BIDS structure for this subject and returns a hash, containing the MRI proper images.

It should return a single value, except for the T1w images, where an array is returned. This was though this way because mostly a single session is done. However, the skill to detect more than one MRI was introduced to allow the movement correction when ADNI images are analyzed

So, for T1w images the returned hash should be asked as

      @{$nifti{'T1w'}} but for other kind of image it should asked as $nifti{'T2w'}

usage: check_subj(project_path, bids_id);

• check_pet

This function takes as input the name of the project and the subject ID Then it seeks along the BIDS structure for this subject and returns a hash, containing the PET proper images.

If also a tracer is given as input, then the returned hash contains the PET-tau associated to this tracer. This was introduced as part of a project were the subjects were analyzed with different radiotracers.

If no tracer is given, it will seek for the FBB PETs. Those PETs are stored as

      - single: 4x5min
- combined: 20min

usage: check_pet(project_path, bids_id, $optional_radiotracer); • check_fs_subj This function checks if the Freesurfer directory of a given subjects exists usage: check_fs_subj(freesurfer_id) • get_lut I really don't even remenber what this shit does • run_dckey Get the content of a public tag from a DICOM file. usage: run_dckey(key, dicom) • dclokey Get the content of a private tag from a DICOM file. usage: dclokey(key, dicom) • centiloid_fbb Returns the proper centiloid value for a given SUVR. Only valid for FBB. usage: centiloid_fbb(suvr); • populate Takes a pattern and a filename and stores the content of the file into a HASH according to the given pattern usage: populate(pattern, filename); • get_subjects Parse a project database taking only the subjects and storing them into an array. The databse is expected to be build as,  0000;name usage: get_subjects(filename); • get_list Parse a project database taking only the subjects and storing them into an array. The databse is expected to be build with a four digits number at the beginning of line. Is similar to get_subjects() function but less restrictive usage: get_list(filename); • get_pair A single file is loaded as input and parse into a HASH. The file should be written in the format:  key;value usage: get_pair(filename); • shit_done this function is intended to be used after a script ends and then an email is send to the user with the name of the script, the name of the project and th results attached usage: shit_done(script_name, project_name, attached_file) • cut_shit This function takes a project database and a file with a list, then returns the elements that are common to both. It is intended to be used to restrict the scripts action over a few elements. It returns a single array. If it is correctly used, first the db is identified with load_project() function and then passed through this function to get the array of subjects to be analyzed. If the file with the cutting list do not exist, an array with all the subjects is returned. usage: cut_shit(db, list); • getLoggingTime This function returns a timestamp based string intended to be used to make unique filenames Stolen from Stackoverflow usage: getLoggingTime(); # FSMetrics Bunch of helpers for storing ROI structure and relative data • fs_file_metrics This function does not read any input. It sole purpose is to returns a HASH containing the templates of order for converting Freesurfer (FS) results into tables. Any hash element is composed by the template ('order'), a boolean ('active') to decide if the FS stats will be processed and the name of the FS stat file ('file'). The order template has two wildcards (<list> and <fs_output>) that should be parsed and changed by the FS subject id and the output directory where the data tables will be stored for each subject The function could be invoked as,  my %stats = fs_file_metrics(); in any script where this information would be needed. The boolean element could be used to choose the stats that should be processed and can be added or modified even at run time if needed. The stored booleans only provided a decent default • fs_fbb_rois deprecated This function exports a HASH that contains the Freesurfer composition of the usual segmentations used for building the SUVR ROI • tau_rois This function takes a string as input and returns an ARRAY containing the list of ROIs that should be build and where the SUVR should be calculated It is intended to be used for PET-Tau but could be used anywhere By default a list of Braak areas are returned. If the input string is alt a grouping of those Braak areas is returned. If the purpose is to build a meta_temporal ROI the string meta should be passed as input The main idea here is read the corresponding file for each ROI, stored at PIPEDIR/lib/tau/ and build each ROI with the FS LUTs store there • pet_rois This function takes a string as input and returns an ARRAY containing the list of ROIs that should be build and where the SUVR should be calculated Input values are parietal, frontal, pieces or global (default) The main idea here is read the corresponding file for each ROI, stored at PIPEDIR/lib/pet/ and build each ROI with the FS LUTs stored there # SLURM This module contains just a function to send the jobs to SLURM from the Perl scripts • send2slurm The function takes a HASH as input where all the information relative to the job should be stored. No data is mandatory inside the input HASH, since the minimal values are automagicaly asigned by default as a constructor (no really, but anyway). Take into account that this subroutine only pass the parameters to SLURM. So, the logic behind your actions should correspond to what you want to do in any case, exactly as if you were writing sbatch scripts. The managed options for SLURM jobs are:  - filename: File where the sbatch script will be stored - job_name: Job name for SLURM (-J) - cpus: Number of CPUs to be used by each job (-c) - mem_per_cpu: Amount of memory to be used for each CPU (--mem-per-cpu) - time: Maximum time that the job will be allowed to run (--time) - output: File where the sbatch script output will be stored (-o) - partition: SLURM partition to be used (-p) - gres: GPUs to be used (--gres) - command: Command to be executed at sbatch script - mailtype: Type of warning to be emailed (--mail-type) - dependency: Full dependency string to be used at sbatch execution (--dependency), see more below The function returns the jobid of the queued job, so it can be used to build complex workflows. usage: my$job_id = send2slurm(\%job_properties);

Warning email: By default, if an empty HASH is passed to the function, a no command sbatch script is launched with --mail-type=END option. The intention is that this could be used to warn at the end of any launched swarm. Also, by changing mailtype but ommiting the command value you can force the function to execute an empty sbatch job with whatever warning behavior that you choose.

Dependencies: If dependencies are going to be used, you need to pass to the function the full string that SLURM expects. That is, you can pass something like singleton or after:000000 or even afterok:000000,000001,000002. This last can be build, by example, storing every previous jobid into an ARRAY and passing then as,

      ...
my $jobid = send2slurm(\%previous); push @jobids,$jobid;
...
...
send2slurm(\%task);

Of course, if dependencies are not going to be used, the dependency option could be safely ignored. But notice that, if you are reusing a HASH then this key should be deleted from it.

# XNATACE

• xconf

Get the XNAT connection data into a HASH

usage: %xnat_data = xconf(configuration_file)

• xget_pet

Get the XNAT PET experiment ID

usage: xget_pet(host, jsession, project, subject)

• xget_mri

Get the XNAT MRI experiment ID

usage: xget_mri(host, jsession, project, subject)

• xget_session

Create a new JSESSIONID on XNAT

usage: xget_session(\%xconf);

• xput_report

Upload a pdf report to XNAT

usage: xput_report(host, jsession, subject, experiment, pdf_file);

• xput_rvr

Upload a JSON file with VR data

usage: xput_rvr(host, jsession, experiment, json_file);

• xget_rvr

Get VR results into a HASH. Output is a hash with filenames and URI of each element stored at RVR

usage: xget_rvr(host, jsession, project, experiment);

• xget_rvr_data

Get RVR JSON data into a hash

usage: xget_rvr_data(host, jsession, URI);

• xget_subjects

Get the list of subjects of a project into a HASH. El HASH de input, %sbjs, se construye como { XNAT_ID => Label }

usage: %sbjs = xget_subjects(host, jsession, project);

• xget_pet_reg

usage: xget_pet_reg(host, jsession, experiment, nifti_output);

# Dependencias

Data::Dump
File::Slurp
File::Basename
File::Temp
File::Copy::Recursive
File::Copy
File::Find::Rule
File::Remove
Cwd
JSON