Skip to main content

Design Guide

This guides provides a high-level overview of the Frictionless Framework architecture. It will be useful for plugin authors and advanced users.

Reading Flow#

Frictionless uses modular approach for its architecture. During reading a data source goes through various subsystems which are selected depending on the data characteristics:


For example for a local CSV file it will use:

  • LocalLoader
  • CsvParser

System Object#

The most important undelaying object in the Frictionless Framework is system. It's an singletone object avaialble as frictionless.system. This object can be used to instantiate different kind of lower-level as though Check, Step, or Type. Here is a quick example of using the system object:

from frictionless import Resource, system
resource = Resource(path='')
check = system.create_check({'code': 'duplicate-row'})
control = system.create_control(resource, descriptor={'httpTimeout': 1000})
dialect = system.create_dialect(resource, descriptor={'delimiter': ';'})
loader = system.create_loader(resource)
parser = system.create_parser(resource)
server = system.create_server('api')
step = system.create_step({'code': 'table-validate'})
storage = system.create_storage('sql', 'postgresql://database')
type = system.create_type(resource.get_field('id'))

As an extension author you might use the system object in various cases. For example, take a look at this MultipartLoader excerpts:

def read_line_stream(self):
for number, path in enumerate(self.__path, start=1):
with system.create_loader(Resource(path=path)) as loader:
for line_number, line in enumerate(loader.byte_stream, start=1):
if not self.__headless and number > 1 and line_number == 1:
yield line

It's important to understand that creating low-level objects in general is more corect using the system object than just classes because it will include all the available plugins in the process.

Last updated on by roll