Paramify

Paramify is a lightweight Python library for dynamic parameter management. Define, validate, and manage parameters from a YAML or JSON schema with optional CLI integration and a web-based configuration UI.

---

Key Features

• Dynamic Parameter Management — define parameters in YAML/JSON or as plain Python dicts
• Groups & Nesting — organise parameters hierarchically with type: group
• YAML Includes — split large configs across multiple files with include:
• Typed Lists — list[bool], list[int], list[float], list[double], list[str], list[string]
• C++-compatible type names — double (alias for float) and string (alias for str)
• value: keyword — preferred over default: for setting parameter values (both accepted)
• Scope-Based Control — scope: all | cli | runtime; runtime params are never exposed to CLI
• CLI Integration — auto-generate CLI arguments from the schema; bool params get --flag / --no-flag
• Custom Callbacks — on_<param>_set(self, value) triggered on every update
• Persistence — changes written back to the source YAML/JSON file automatically
• save_as(path) — export current state to a new YAML file
• Web Interface — optional Bottle-based UI for runtime configuration (pip install paramify[web])

---

Installation

# Core (parameter management + CLI)
pip install paramify

# With web interface
pip install paramify[web]

---

Quick Start

Basic usage

from paramify import Paramify

class MyApp(Paramify):
    def on_speed_set(self, value):
        print(f"speed updated to {value}")

params = {
    "parameters": [
        {"name": "enabled", "type": "bool",   "default": True},
        {"name": "speed",   "type": "int",    "default": 10},
        {"name": "label",   "type": "string", "default": "robot"},
    ]
}

app = MyApp(params, enable_cli=False)
print(app.parameters.enabled)   # True
print(app.parameters.speed)     # 10

app.set_speed(42)               # triggers on_speed_set
print(app.get_parameters())     # {'enabled': True, 'speed': 42, 'label': 'robot'}

Groups and nested parameters

Parameters can be organised into groups. Nested params are accessed with dot notation (params.camera.enabled) and their setters/callbacks use underscores (set_camera_enabled, on_camera_enabled_set).

# config.yaml
name: "My Robot App"
parameters:

  - name: "enabled"
    type: "bool"
    value: true
    scope: "all"

  - name: "camera"
    type: "group"
    parameters:
      - name: "enabled"
        type: "bool"
        value: true
      - name: "fps"
        type: "int"
        value: 30
      - name: "gain"
        type: "double"       # alias for float
        value: 1.0

from paramify import Paramify

class MyApp(Paramify):
    def on_camera_enabled_set(self, value):   # dots → underscores
        print(f"camera.enabled → {value}")

app = MyApp("config.yaml")

print(app.parameters.camera.enabled)   # True
print(app.parameters.camera.fps)       # 30

app.set_camera_enabled(False)          # triggers callback + saves file
app.set_camera_gain(2.5)

YAML includes

Large configs can be split across files. Use include: inside a group to load parameters from one or more external YAML files.

# main.yaml
parameters:
  - name: "camera"
    type: "group"
    include: "camera_params.yaml"      # single file

  - name: "audio"
    type: "group"
    include:                           # list of files
      - "audio_base.yaml"
      - "audio_advanced.yaml"

# camera_params.yaml
parameters:
  - name: "enabled"
    type: "bool"
    value: true
  - name: "resolution"
    type: "string"
    value: "1080p"

Included files are saved independently when their parameters change. Circular includes are detected and raise a ValueError.

Typed lists

parameters:
  - name: "channels"
    type: "list[int]"
    value: [1, 2]

  - name: "labels"
    type: "list[string]"
    value: ["cat", "dog"]

  - name: "weights"
    type: "list[double]"
    value: [0.1, 0.9]

Supported list types: list[bool], list[int], list[float], list[double], list[str], list[string], and untyped list.

value: vs default:

Both keywords are accepted. value: takes precedence if both are present (matches paramify-cpp behaviour).

- name: "gain"
  type: "double"
  value: 2.0      # preferred
  default: 1.0    # fallback if value: is absent

Scope

Scope	CLI	Runtime
all (default)	✅	✅
cli	✅	❌
runtime	❌	✅

- name: "debug"
  type: "bool"
  value: false
  scope: "runtime"   # never shown in --help, never parsed from argv

save_as

Export the current parameter state to a new YAML file:

app.save_as("/tmp/config_snapshot.yaml")

---

YAML Configuration Reference

name: "My App"                    # optional app name
description: "..."                # optional description shown in --help

parameters:

  # Scalar types
  - name: "flag"
    type: "bool"                  # bool | int | float | double | str | string
    value: true                   # 'value:' preferred; 'default:' also accepted
    scope: "all"                  # all | cli | runtime  (default: all)
    label: "Enable flag"          # display label (used in web UI)
    description: "..."            # shown in --help

  # Typed list
  - name: "ids"
    type: "list[int]"             # list[bool|int|float|double|str|string] or plain list
    value: [1, 2, 3]

  # Inline group
  - name: "camera"
    type: "group"
    parameters:
      - name: "enabled"
        type: "bool"
        value: true
      - name: "fps"
        type: "int"
        value: 30

  # Group loaded from external file(s)
  - name: "audio"
    type: "group"
    include: "audio_params.yaml"  # string or list of strings

---

JSON Configuration

JSON configs support all scalar types and typed lists but not type: group or include:. Use "default" as the value key in JSON.

{
    "name": "My App",
    "parameters": [
        {"name": "enabled", "type": "bool",   "default": true},
        {"name": "speed",   "type": "int",    "default": 10},
        {"name": "labels",  "type": "list[string]", "default": ["a", "b"]}
    ]
}

---

CLI Integration

Parameters with scope: all or scope: cli are automatically exposed as CLI arguments.

python my_app.py --speed 99 --camera-enabled --no-camera-mute

• Dot-notation keys become kebab-case flags: camera.enabled → --camera-enabled
• Bool params get both --flag and --no-flag
• List params accept multiple values: --channels 1 2 3
• CLI values override file defaults but are not persisted

---

Web Interface

pip install paramify[web]

from paramify.web import ParamifyWeb

class MyApp(ParamifyWeb):
    def on_speed_set(self, value):
        print(f"speed → {value}")

app = MyApp("config.yaml", host="0.0.0.0", port=5000)
input("Press Enter to stop...")

Open http://localhost:5000 to adjust parameters in the browser. All groups (including those loaded via include:) appear as a flat list in the UI.

---

Callbacks and Setters

For a parameter with key camera.enabled:

What	Name
Setter method	set_camera_enabled(value)
Callback method	on_camera_enabled_set(self, value)
Namespace access	app.parameters.camera.enabled
Dict access	app.get_parameters()["camera.enabled"]

Dots in parameter keys are replaced by underscores in method names.

---

Contributing

Contributions are welcome! Feel free to open issues or submit pull requests.

---

License

This project is licensed under the MIT License. See the LICENSE file for details.