Introduction¶
This is the documentation for the command-line tool sct_dbtool. This tool will apply custom checks into each item of the MRI database.
Installation¶
If you want, you can create a python environment with Anaconda:
conda create -n dbtoolenv python=3.6
source activate dbtoolenv
To install it, clone the repository and install with:
cd sct_dbtool
pip install .
Note
If you are developing, install it in development mode with pip install -e ., so changed mande in the source code will reflect automatically in your environment.
Configuration¶
Before running the tool, it must be configured with the database credentials, in order to do that, just execute the following command below:
sct_dbtool setup
And then answer the questions about the credentials for the Web Management system.
How to use it¶
To use it, just run the following command:
sct_dbtool sanity [PATH_DUKE]/sct_testing/large
Where the last parameter is the path to where the database files are. After running it, the tool will generate a output report called report.html that can be opened on any browser.
Implementing more checks¶
All sanity checks are implemented in sct_dbtool.sanity
module. All
sanity checks should inherit from the class
sanity.SanityCheck
. An example is the checking
sanity.CheckEmptyDemographics
that will check
if the demographics field is present in each item of the database,
the code is shown below:
class CheckEmptyDemographics(SanityCheck):
def __init__(self, db_path: str):
super().__init__("warning", db_path)
def check(self, item: Dict):
if "demographic" not in item:
self.add_diagnostic("Demographic field not present.", item)
return
if item["demographic"] is None:
self.add_diagnostic("Demographic field is empty.", item)
As you can see, the class inherits from sanity.SanityCheck
and it implements two methods:
__init__()
: this is the class constructor, you should just specify the category of the check, which can be warning, error or info. And call the base class constructor with this argument and the db_path argument;check()
: this is the method that implements the check itself. The item argument is a Python dictionary with the database item data. In the example above, we’re just checking if the “demographic” field is present in this dictionary, and if not, it will call the methodSanityCheck.add_diagnostic()
that will add a new diagnostic message for this item with a custom message.After adding a new class, the tool will automatically use this new class to check every database item in the next run of the sct_dbtool sanity checking. It will also add this class name as filter in the report as well as all diagnostic messages generated by this checking.
If you need to refer any file in the file system, as used by other checks to check if the file mentioned in the database is present in the file system, you can just use the
SanityCheck.db_path
object attribute, that contains the path of the file system that was passed as argument to the sct_dbtool.
Please check other examples in the module sct_dbtool.sanity
for more information.