Facts Overview

Facts enable pyinfra to collect remote server state, which is used to diff against the desired state and determine which operations need to run.

What are Facts?

Facts are Python classes that gather information about a host by executing commands and processing the output. They provide a consistent interface for reading system state.

from pyinfra import host
from pyinfra.facts.server import Hostname, LinuxDistribution
from pyinfra.facts.files import File

# Get simple fact
hostname = host.get_fact(Hostname)
print(f"Hostname: {hostname}")

# Get fact with arguments
file_info = host.get_fact(File, path="/etc/hosts")
print(f"Size: {file_info['size']} bytes")

# Get complex fact
distro = host.get_fact(LinuxDistribution)
print(f"OS: {distro['name']} {distro['version']}")

Fact Categories

Pyinfra includes facts organized by category:

System Facts

Server - System information (hostname, OS, architecture, users, groups)
Hardware - CPU, memory, disk information

File System Facts

Files - File and directory information
File - Single file metadata
Directory - Directory contents
FindInFile - Search within files

Package Manager Facts

Packages - Installed packages across package managers
Apt - Debian/Ubuntu packages
Yum/Dnf - Red Hat/CentOS packages
Apk - Alpine packages
Brew - macOS packages
Pip - Python packages
Npm - Node.js packages
Gem - Ruby packages

Service Facts

Systemd - Systemd services
Upstart - Upstart services
Launchd - macOS services
Sysvinit - SysV init services

Application Facts

Git - Git repositories
Docker - Docker containers and images
MySQL - MySQL databases
PostgreSQL - PostgreSQL databases

Using Facts

In Operations

Use facts to make operations conditional:

from pyinfra import host
from pyinfra.facts.server import LinuxDistribution
from pyinfra.operations import apt, yum

# Get distribution info
distro = host.get_fact(LinuxDistribution)

# Conditional operations based on OS
if distro["name"] in ["Ubuntu", "Debian"]:
    apt.packages(
        name="Install nginx",
        packages=["nginx"],
    )
elif distro["name"] in ["CentOS", "RedHat"]:
    yum.packages(
        name="Install nginx",
        packages=["nginx"],
    )

Check File Exists

from pyinfra import host
from pyinfra.facts.files import File
from pyinfra.operations import files

config_file = host.get_fact(File, path="/etc/app/config.ini")

if not config_file:
    files.put(
        name="Upload config",
        src="config.ini",
        dest="/etc/app/config.ini",
    )
else:
    host.noop("Config already exists")

Check Package Installed

from pyinfra import host
from pyinfra.facts.server import Which
from pyinfra.operations import apt

nginx_path = host.get_fact(Which, command="nginx")

if not nginx_path:
    apt.packages(
        name="Install nginx",
        packages=["nginx"],
    )

Fact API

FactBase Class

All facts inherit from FactBase:

from pyinfra.api import FactBase

class MyFact(FactBase):
    def command(self, arg1, arg2):
        """Return the command to execute."""
        return f"my-command {arg1} {arg2}"
    
    def process(self, output: list[str]):
        """Process command output into structured data."""
        # Parse output and return data
        return {"result": output[0]}

Creating Custom Facts

Create custom facts for your specific needs:

# facts/custom.py
from pyinfra.api import FactBase

class AppVersion(FactBase):
    """Get the installed version of our app."""
    
    def command(self):
        return "/opt/myapp/bin/myapp --version"
    
    def process(self, output):
        # Output: "MyApp version 1.2.3"
        if output and output[0].startswith("MyApp version"):
            return output[0].split(" ")[-1]
        return None

class AppConfig(FactBase):
    """Read app configuration value."""
    
    def command(self, key):
        return f"grep '^{key}=' /etc/myapp/config.ini | cut -d= -f2"
    
    def process(self, output):
        return output[0] if output else None

Usage:

from facts.custom import AppVersion, AppConfig
from pyinfra import host

# Get app version
version = host.get_fact(AppVersion)
print(f"App version: {version}")

# Get config value
port = host.get_fact(AppConfig, key="port")
print(f"App port: {port}")

Fact Caching

Facts are cached per host to avoid redundant execution:

from pyinfra import host
from pyinfra.facts.server import Hostname

# First call executes command
hostname1 = host.get_fact(Hostname)

# Second call returns cached value
hostname2 = host.get_fact(Hostname)

assert hostname1 == hostname2

Facts with arguments are cached separately:

from pyinfra.facts.files import File

# Each path is cached separately
file1 = host.get_fact(File, path="/etc/hosts")
file2 = host.get_fact(File, path="/etc/passwd")

Fact Execution

Facts can use global arguments:

from pyinfra import host
from pyinfra.facts.files import File

# Execute fact with sudo
file_info = host.get_fact(
    File,
    path="/root/.ssh/authorized_keys",
    _sudo=True,
)

CLI Usage

Query facts from the command line:

# Get fact from all hosts
pyinfra inventory.py --facts server.Hostname

# Get fact with arguments
pyinfra inventory.py --facts files.File path=/etc/hosts

# Multiple facts
pyinfra inventory.py --facts server.Hostname server.Arch

# JSON output
pyinfra inventory.py --facts server.LinuxDistribution --json

Best Practices

1. Cache Facts in Variables

# Good - fact called once
from pyinfra import host
from pyinfra.facts.server import LinuxDistribution

distro = host.get_fact(LinuxDistribution)
if distro["name"] == "Ubuntu" and distro["major"] >= 20:
    # Use distro
    pass

# Bad - fact called multiple times (though cached)
if host.get_fact(LinuxDistribution)["name"] == "Ubuntu" and \
   host.get_fact(LinuxDistribution)["major"] >= 20:
    pass

2. Handle Missing Facts

from pyinfra.facts.files import File

file_info = host.get_fact(File, path="/opt/app/version.txt")

if file_info:
    print(f"File size: {file_info['size']}")
else:
    print("File does not exist")

3. Use Specific Facts

# Good - specific fact
from pyinfra.facts.server import Which
nginx_path = host.get_fact(Which, command="nginx")

# Bad - using shell command
from pyinfra.facts.server import Command
nginx_path = host.get_fact(Command, command="which nginx")

4. Reuse Facts

Create reusable fact-based checks:

from pyinfra import host
from pyinfra.facts.server import Which

def package_installed(command):
    """Check if a package is installed."""
    return host.get_fact(Which, command=command) is not None

if not package_installed("nginx"):
    # Install nginx
    pass

if not package_installed("git"):
    # Install git
    pass

Fact Types

Simple Facts

Return a single value:

from pyinfra.facts.server import Hostname

# Returns: "web1.example.com"
hostname = host.get_fact(Hostname)

Dict Facts

Return structured data:

from pyinfra.facts.server import LinuxDistribution

# Returns: {"name": "Ubuntu", "version": "22.04", "major": 22, ...}
distro = host.get_fact(LinuxDistribution)

List Facts

Return multiple items:

from pyinfra.facts.server import Groups

# Returns: ["root", "sudo", "users", ...]
groups = host.get_fact(Groups)

Parameterized Facts

Accept arguments:

from pyinfra.facts.files import File
from pyinfra.facts.server import User

# File with path argument
file_info = host.get_fact(File, path="/etc/hosts")

# User with username argument
user_info = host.get_fact(User, user="nginx")

Error Handling

Facts can fail if commands error:

try:
    file_info = host.get_fact(File, path="/etc/hosts", _sudo=True)
except Exception as e:
    print(f"Failed to get file info: {e}")

Some facts return None for missing items:

from pyinfra.facts.files import File
from pyinfra.facts.server import Which

# Returns None if file doesn't exist
file_info = host.get_fact(File, path="/nonexistent")

# Returns None if command not found
nginx_path = host.get_fact(Which, command="nonexistent")

Source Reference

Location: src/pyinfra/api/facts.py

Key Classes

FactBase - Base class for all facts (line 53)
ShortFactBase - Shorthand fact wrapper (line 98)

Key Functions

get_fact() - Get fact value (line in facts.py)

Overview

Core API

Commands

Connectors

Facts

What are Facts?

Fact Categories

System Facts

File System Facts

Package Manager Facts

Service Facts

Application Facts

Using Facts

In Operations

Check File Exists

Check Package Installed

Fact API

FactBase Class

Creating Custom Facts

Fact Caching

Fact Execution

CLI Usage

Best Practices

1. Cache Facts in Variables

2. Handle Missing Facts

3. Use Specific Facts

4. Reuse Facts

Fact Types

Simple Facts

Dict Facts

List Facts

Parameterized Facts

Error Handling

Source Reference

Key Classes

Key Functions

Build docs developers (and LLMs) love

Overview

Core API

Commands

Connectors

Facts

Documentation Index

​What are Facts?

​Fact Categories

​System Facts

​File System Facts

​Package Manager Facts

​Service Facts

​Application Facts

​Using Facts

​In Operations

​Check File Exists

​Check Package Installed

​Fact API

​FactBase Class

​Creating Custom Facts

​Fact Caching

​Fact Execution

​CLI Usage

​Best Practices

​1. Cache Facts in Variables

​2. Handle Missing Facts

​3. Use Specific Facts

​4. Reuse Facts

​Fact Types

​Simple Facts

​Dict Facts

​List Facts

​Parameterized Facts

​Error Handling

​Source Reference

​Key Classes

​Key Functions

Build docs developers (and LLMs) love

What are Facts?

Fact Categories

System Facts

File System Facts

Package Manager Facts

Service Facts

Application Facts

Using Facts

In Operations

Check File Exists

Check Package Installed

Fact API

FactBase Class

Creating Custom Facts

Fact Caching

Fact Execution

CLI Usage

Best Practices

1. Cache Facts in Variables

2. Handle Missing Facts

3. Use Specific Facts

4. Reuse Facts

Fact Types

Simple Facts

Dict Facts

List Facts

Parameterized Facts

Error Handling

Source Reference

Key Classes

Key Functions