How do I install the PiSharp Python SDK for PI Web API?

Run pip install pisharp-piwebapi. Requires Python 3.9 or later. For Kerberos support, also install requests-kerberos.

Does the PiSharp SDK support Kerberos authentication?

Yes. Pass auth_method='kerberos' when creating a client. The SDK uses requests-kerberos under the hood and supports mutual authentication and delegation.

SDK

Python SDK for PI Web API

An open-source Python library that handles PI Web API authentication, sessions, pagination, error handling, and pandas conversion so you can focus on your integration logic.

The problem the SDK solves

Every PI Web API integration starts with the same 100 lines of boilerplate: session setup, SSL configuration, retry logic, WebID lookups, pagination handling, digital state detection, and JSON-to-DataFrame conversion. Teams copy-paste this code between projects, each copy drifting slightly and introducing its own bugs.

The PiSharp Python SDK packages all of that into a tested, versioned library. When we fix a bug or add a feature, every project that depends on the SDK gets the improvement via a version bump.

5 lines to first data

Go from zero to reading PI data in under 5 lines. Authentication, sessions, SSL, and connection pooling are handled automatically.

Production-ready defaults

Retry logic with exponential backoff, certificate handling, proper error types, and connection timeout management built in.

pandas-native output

Methods return DataFrames directly with proper timestamp indexing, quality flag filtering, and digital state handling. No manual JSON-to-DataFrame wrangling.

Open source, MIT licensed

Read the source, audit the code, contribute fixes, or fork for your own needs. No vendor lock-in.

Installation

Install from PyPIbash

pip install pisharp-piwebapi

Requirements

Dependency	Version
Python	3.9+
requests	2.28+
pandas (optional)	1.5+ (for DataFrame output)

Kerberos authentication

For Kerberos support, also install pip install requests-kerberos. The SDK detects it automatically and enables Kerberos auth methods.

Quick start

Connect and read your first valuepython

from pisharp_piwebapi import PIWebAPI

# Connect to your PI Web API server
pi = PIWebAPI(
    base_url="https://myserver/piwebapi",
    username="your-username",
    password="your-password",
    # ca_bundle="/path/to/cert.pem",  # Optional: custom CA cert
)

# Read the current value of a PI point
value = pi.points.get_value("sinusoid", server="MY-PI-SERVER")
print(f"Value: {value.value}")
print(f"Timestamp: {value.timestamp}")
print(f"Good: {value.good}")

# Pull recorded history into a pandas DataFrame
df = pi.points.get_recorded(
    "sinusoid",
    server="MY-PI-SERVER",
    start_time="*-24h",
    end_time="*",
)
print(df.describe())
print(f"Rows: {len(df)}, Quality: {df['good'].mean():.1%} good")

Common operationspython

# Find a point by path
point = pi.points.get("sinusoid", server="MY-PI-SERVER")
print(f"WebId: {point.web_id}, Type: {point.point_type}")

# Read interpolated values at fixed intervals
df = pi.points.get_interpolated(
    "sinusoid", server="MY-PI-SERVER",
    start_time="*-7d", end_time="*", interval="1h",
)

# Read summary statistics
summary = pi.points.get_summary(
    "sinusoid", server="MY-PI-SERVER",
    start_time="*-24h",
    summary_types=["Average", "Minimum", "Maximum", "StdDev"],
)

# Write a value
pi.points.write_value("my-output-tag", server="MY-PI-SERVER", value=42.0)

# Batch read current values for many points
values = pi.batch.read_current(web_ids)  # Automatic chunking

What the SDK covers

Capability	What the SDK handles
Authentication	Basic, Kerberos, and NTLM with automatic session management and connection pooling
SSL / Certificates	Custom CA bundle, Windows cert store integration, certificate validation
Point lookup	Find by path, name filter, or search pattern. WebIDs resolved and cached automatically
Reading values	Current, recorded, interpolated, and summary reads with DataFrame output. Digital state handling and quality filtering included
Writing values	Single and bulk writes with updateOption and bufferOption support
Batch operations	Automatic chunking, partial failure detection, and retry for transient errors
Pagination	Transparent iteration over paginated responses with truncation warnings
Error handling	Typed exceptions (`PIAuthError`, `PINotFoundError`,`PIServerError`), retry with backoff, and clear error messages

When to use the SDK vs raw requests

Use the SDK when:

You are building a production integration, working on a team project, need reliable error handling, or want DataFrames without manual conversion.

Use raw requests when:

You need to access an endpoint the SDK does not cover, are learning how the API works, or prefer full control over every HTTP detail. The SDK provides a pi.raw_request() escape hatch for this.