Pydantic models for Nokia SR Linux

Important

This is an experimental project that may graduate to a proper set of Pydantic models for SR Linux in the near future.

Limitations:

• By default, models are generated without if-features flag set, this means all if-features are implicitly enabled. Note, that this will render models for features that your hardware might not have. Use the --platform flag to render models that respect feature flags.
• if you found additional limitations, please open an issue or reach out in Discord.

Installation

uv add git+https://github.com/srl-labs/pydantic-srlinux

or with pip:

pip install git+https://github.com/srl-labs/pydantic-srlinux

Usage

With the pydantic-srlinux package installed, you can get access to the generated models by importing them from the pydantic_srlinux module.

import pydantic_srlinux.models.interfaces as srl_if

e1_1 = srl_if.InterfaceListEntry(
    name="ethernet-1/1",
    admin_state=srl_if.EnumerationEnum.enable,
    vlan_tagging=True,
    subinterface=[
        srl_if.SubinterfaceListEntry(
            index=0,
            type="bridged",
            admin_state=srl_if.EnumerationEnum.enable,
            vlan=srl_if.VlanContainer(
                encap=srl_if.EncapContainer(
                    single_tagged=srl_if.SingleTaggedContainer(
                        vlan_id=srl_if.VlanIdType(100)
                    )
                )
            ),
        )
    ],
)

The following top level imports are available:

Model	Schema URL
acl	acl doc
bfd	bfd doc
interfaces	interfaces doc
network_instance	network_instance doc
platform	platform doc
qos	qos doc
routing_policy	routing_policy doc
system	system doc
tunnel_interface	tunnel_interface doc
tunnel	tunnel doc

The top level imports match the top level YANG objects that you will see in the SR Linux NOS. Consult with the schema doc or IDE hints to find the structure of the objects and their hierarchy.

With an object instance, you can produce the JSON serialized representation of the object:

print(
    e1_1.model_dump_json(indent=2, exclude_none=True, exclude_unset=True, by_alias=True)
)

And use the serialized JSON with gNMI or JSON-RPC interfaces of SR Linux. Have a look at the JSON-RPC-based example in the example directory.

Module schemas

To help you navigate the module schemas, the table above includes links to the JSON Schema documents, these are generated by using the generate_models_schema.py script.

Generate YANG mapping file

Clone the nokia/srlinux-yang-models repo.

git clone https://github.com/nokia/srlinux-yang-models.git

Generate the YANG map file:

./yang_map.py --dir {path to a dir where you cloned nokia/srlinux-yang-models} --version v24.10.1

The script will navigate to the directory with the yang modules identified by the --dir argument and checkout the tag specified by the --version argument.

Create an environment variable ${SRL_YANG_REPO_DIR} with the path to the cloned repo as this env var is used in the YANG map file to keep it generically applicable.

export SRL_YANG_REPO_DIR=~/path/to/srlinux-yang-models

Generate Pydantic models (automatic)

To manually regenerate the Pydantic models for a given module, run the following command:

# generate_models.py --module <model name>
./generate_models.py --module srl_nokia-acl

This can optionally be scoped to a specific platform from the yang map:

# generate_models.py --module <model name> --platform <platform name>
./generate_models.py --module srl_nokia-acl --platform 7250-IXR-X1B

The full list of the top level modules:

./generate_models.py --module srl_nokia-acl
./generate_models.py --module srl_nokia-bfd
./generate_models.py --module srl_nokia-interfaces
./generate_models.py --module srl_nokia-network-instance
./generate_models.py --module srl_nokia-platform
./generate_models.py --module srl_nokia-qos
./generate_models.py --module srl_nokia-routing-policy
./generate_models.py --module srl_nokia-system
./generate_models.py --module srl_nokia-tunnel-interfaces
./generate_models.py --module srl_nokia-tunnel

This will generate the xxx.py file in the pydantic_srlinux/models directory as well as create the CLI command used to generate the Pydantic models and store it in the ./temp directory.

Pyang and Yanglint

The yang map file contains pastable commands to generate YANG trees for all the modules. The commands are generated for pyang and yanglint. Before using these commands, make sure that the env var ${SRL_YANG_REPO_DIR} is set to the path of the cloned YANG repo.

Yanglint

Yanglint is easy to consume in a container flavor; paste this alias in your shell:

alias yanglint='docker run --rm -i -t -v ${SRL_YANG_REPO_DIR}:/${SRL_YANG_REPO_DIR} ghcr.io/hellt/yanglint:3.7.8 $@'

Then you will be able to paste the yanglint command from the map file and get the tree output.

By default yanglint will output warnings for all the implemented modules (these are the modules that your particular module relies on). You can suppress the warnings by adding -Q