Perigon Python SDK

Python client for the Perigon API

A modern, fully‑typed Python SDK for the Perigon API, generated from the official OpenAPI specification.
Works in CPython 3.8+, PyPy, serverless runtimes, notebooks, and async frameworks.

Table of Contents

✨ Features

• Type‑hinted request/response models powered by Pydantic
• Async and sync support - choose the right approach for your application
• Ships with PEP 561 type hints for excellent IDE integration
• Generated directly from https://docs.perigon.io, so it's always in sync

📦 Installation

pip install perigon
# poetry add perigon
# pipx install perigon

🚀 Quick start

1. Instantiate the client

from perigon import V1Api, ApiClient

# Create client with API key
api = V1Api(ApiClient(api_key="YOUR_API_KEY"))

# Alternative: environment variable or callable
# api = V1Api(ApiClient(api_key=os.environ["PERIGON_API_KEY"]))
# api = V1Api(ApiClient(api_key=lambda: get_api_key_from_vault()))

2. Make calls

# 🔍 Search recent news articles (sync)
articles = api.search_articles(q="artificial intelligence", size=5)
print(articles.num_results, articles.articles[0].title)

# 👤 Look up a journalist by ID (sync)
journalist = api.get_journalist_by_id(id="123456")
print(journalist.name)

# 🔄 Use async variant for async applications
import asyncio

async def fetch_data():
    # Search articles asynchronously 
    articles = await api.search_articles_async(q="technology", size=5)
    
    # Look up journalist asynchronously
    journalist = await api.get_journalist_by_id_async(id="123456")
    
    return articles, journalist

# Run in async context
articles, journalist = asyncio.run(fetch_data())

All methods return typed objects with full IDE autocompletion support.

🧑‍💻 Endpoint examples

Articles – search and filter news (/v1/all)

Docs → https://docs.perigon.io/docs/overview

# Simple query
articles = api.search_articles(q="technology", size=5)

# With date range
articles = api.search_articles(
    q="business", 
    var_from="2025-04-01",  # Note: 'from' is a reserved keyword in Python
    to="2025-04-08"
)

# Restrict to specific sources
articles = api.search_articles(source=["nytimes.com"])

Companies – fetch structured company data (/v1/companies)

Docs → https://docs.perigon.io/docs/company-data

results = api.search_companies(name="Apple", size=5)

Journalists – search and detail look‑up (/v1/journalists)

Docs → https://docs.perigon.io/docs/journalist-data

# Search for journalists
results = api.search_journalists1(name="Kevin", size=1)

# Get detailed information
journalist = api.get_journalist_by_id(id=results.journalists[0].id)

Stories – discover related article clusters (/v1/stories)

Docs → https://docs.perigon.io/docs/stories-overview

stories = api.search_stories(q="climate change", size=5)

Vector search – semantic retrieval (/v1/vector)

Docs → https://docs.perigon.io/docs/vector-endpoint

from perigon.models.article_search_params import ArticleSearchParams

results = api.vector_search_articles(
    article_search_params=ArticleSearchParams(
        prompt="Latest advancements in artificial intelligence",
        size=5
    )
)

Summarizer – generate an instant summary (/v1/summarizer)

Docs → https://docs.perigon.io/docs/search-summarizer

from perigon.models.summary_body import SummaryBody

summary = api.search_summarizer(
    summary_body=SummaryBody(prompt="Key developments"),
    q="renewable energy", 
    size=10
).summary

print(summary)

Topics – explore taxonomy (/v1/topics)

Docs → https://docs.perigon.io/docs/topics

topics = api.search_topics(size=10)

Wikipedia – search and filter pages (/v1/wikipedia)

Docs → https://docs.perigon.io/docs/wikipedia

# Search Wikipedia pages
wikipedia_result = api.search_wikipedia(
    q="machine learning",
    size=3,
    sort_by="relevance"
)

# Filter by specific criteria
wikipedia_result = api.search_wikipedia(
    q="artificial intelligence",
    pageviews_from=100,  # Only popular pages
)

Wikipedia vector search – semantic retrieval (/v1/vector/wikipedia)

Docs → https://docs.perigon.io/docs/vector-wikipedia

from perigon.models.wikipedia_search_params import WikipediaSearchParams

results = api.vector_search_wikipedia(
    wikipedia_search_params=WikipediaSearchParams(
        prompt="artificial intelligence and neural networks in computing",
        size=3,
        pageviews_from=100
    )
)

🔄 Async Support

All methods have async counterparts with the _async suffix:

import asyncio
from perigon import V1Api, ApiClient

async def main():
    api = V1Api(ApiClient(api_key="YOUR_API_KEY"))
    
    # Concurrent API calls
    articles_task = api.search_articles_async(q="technology", size=5)
    journalist_task = api.get_journalist_by_id_async(id="123456")
    
    # Gather results
    articles, journalist = await asyncio.gather(articles_task, journalist_task)
    
    return articles, journalist

# Run the async function
articles, journalist = asyncio.run(main())

🪪 License

MIT © Perigon