KAT’s PDF report engine

Keiko works by preprocessing a LateX template and then compiling it to a PDF. Preprocessing is done with Jinja2.

Report input data is currently strictly typed and enforced by Pydantic models. This is done for now to avoid dependencies with other parts of KAT, allowing decoupling and independent development of reports.

Installation requirements

Install Python requirements with pip in a venv:

python3 -m pip install -r requirements.txt

Make sure pdflatex is installed and added to $PATH, because Keiko invokes pdflatex -synctex=1 -interaction=nonstopmode {filename} as a shell subprocess to compile the report.

Recommended LateX distro for report developers: MikTex

File system permissions

Keiko needs to be able to write to the reports directory, which should be created before running Keiko. The location of this folder is configurable with the environment variable KEIKO_REPORTS_FOLDER. If you are running Keiko as a non-root user, make sure that this directory is writable by the user.

Keiko creates a temporary directory for each report, which is deleted after the report is compiled. This directory is created with Python’s tempfile module. This should by default work fine without any additional configuration.

Running the API

Start API with:


And browse to http://localhost:8000/docs to see the API documentation. Reports can be generated using the /reports endpoint. The examples of this endpoint are seeded with the sample.json files found in each template directory.

After submitting the report request, the report id is immediately returned, while asynchronous processing of the LateX template is started. After the report is generated, the PDF is available at /reports/<id>.keiko.pdf.


Keiko logging can be configured through by supplying a logging.json, relative to the current working directory of the process. See logging.json for an example.

Building a new template

Create a new directory in the templates directory, with the following files:

  • template.tex: the template file

  • model.json: the python pydantic model, describing the shape of the report input data

  • sample.json: a sample input data file, this will be automatically shown in the API documentation

As an example, look into the templates/dns directory.

Generating a report

There are two ways to generate a report:

  • Using the API and the API examples

  • Using the command line

Testing a report with command line

To speed up debugging and development, a command line interface is provided to test a report without the API.

python3 -m keiko.cli templates/dns/sample.json