WebPath

A Python HTTP client library that makes interacting with APIs with lesser boilerplate. Built on httpx and jmespath.

• Client Interface: Manage sessions, headers, and retries centrally
• JSON Traversal: Use jmespath expressions to find whatever you need
• Async: async and await support
• Caching & Retries: Features with simple controls
• Path-like URL Building: Construct URLs with a clean, intuitive / operator.

---

The webpath Workflow

webpath tries to integrate the lifecycle of an API request. Eg, from building the URL to processing the response.

Without webpath

You manually wire together multiple libraries and steps as shown in the example below:

import httpx
import jmespath

# step 1 manual configuration
transport = httpx.HTTPTransport(retries=3)
client = httpx.Client(
    base_url="[https://api.github.com](https://api.github.com)",
    headers={"Accept": "application/vnd.github.v3+json"},
    transport=transport
)

# step 2 build the url
url = "/repos/python/cpython"

# step 3 make the request
response = client.get(url)
response.raise_for_status()
data = response.json()

description = jmespath.search("description", data)
print(description)

With webpath

The entire process is a single chain managed by Client:

from webpath.core import Client

# step 1.. create a reusable client for the API
with Client(
    "[https://api.github.com](https://api.github.com)",
    headers={"Accept": "application/vnd.github.v3+json"},
    retries=3
) as api:
    # step 2.. find the data in one line
    description = api.get("repos", "python", "cpython").find("description")
    print(description)

Installation

Core features (includes httpx and jmespath)

pip install webpath

For progress bar on downloads

pip install "webpath[progress]"

Quick Start

Client Usage

from webpath.core import Client
import asyncio

# step 1. Create the client with a base URL, shared headers, and retries
#    The client can be used for the lifetime of your application.
api = Client(
    "[https://api.github.com](https://api.github.com)",
    headers={"Accept": "application/vnd.github.v3+json"},
    retries=3,
    timeout=10.0
)

# step 2. make sync requests
repo_info = api.get("repos", "python", "cpython")
print(f"Repo Description: {repo_info.find('description')}")

# step 3. make async requests
async def get_user_name():
    user_info = await api.aget("users", "torvalds")
    print(f"User's Name: {user_info.find('name')}")

asyncio.run(get_user_name())


# step 4. use the context manager to ensure connections are closed
try:
    with Client("[https://httpbin.org](https://httpbin.org)", retries=2) as http:
        http.get("status/503")
except Exception as e:
    print(f"\nRequest failed as expected: {e}")
finally:
    ## **Always close the client when you're done**
    api.close()

Retries with a Pluggable Policy

import httpx

## create your custom backoff logic
def retry_backoff(response):
    if "Retry-After" in response.headers:
        wait_time = float(response.headers["Retry-After"])
        print(f"rate limited, sleeping {wait_time}s")
        return wait_time

## pass your custom logic into client
with Client("https://httpbin.org", retries=retry_backoff) as client:
    try:
        client.get("status/429")
    except httpx.HTTPStatusError as e:
        print(f"still failed: {e}")

Core Features

• Client-based Sessions: Manages connection pooling, default headers, retries, caching, and rate limiting for the API
• URL Building: Chain path segments with the / operator (e.g. api/"users"/123).
• JSON Traversal: A jmespath-powered find() method lets you query complex JSON (e.g., resp.find("users[?age > 18].name")) . -- In future may change to jonq when the library is more stable
• Async: Full async/await support for all HTTP verbs (aget, apost, etc.)
• Retries: Client will auto handle transient network or server errors
• Response Caching: Add .with_cache(ttl=120) to any request to cache the response
• File Downloads: .download() method with progress bars and checksum validation
• Rate Limiting: Automatically throttle requests with .with_rate_limit(requests_per_second=1).

Full Tutorial: Using the CoinGecko API

This tutorial will show you how to fetch cryptocurrency data using webpath

1. Set Up the API Client

First, create a Client (this is reusable) for the CoinGecko API. We'll set a 60-second TTL for caching

from webpath.core import Client

api = Client("[https://api.coingecko.com/api/v3](https://api.coingecko.com/api/v3)", cache_ttl=60)

2. Fetch Market Data

Now use the client to hit the endpoint. We can build the path and add query parameters in one call

coins = ["bitcoin", "ethereum", "dogecoin"]

# build your path + query params
market_data = api.get(
    "coins", "markets",
    params={
        "vs_currency": "usd",
        "ids": ",".join(coins),
        "order": "market_cap_desc"
    }
)

3. Extract Data with find() The response is a JSON array of objects. We can use find() with a jmespath expression to extract the names and prices from all coins

# jmes expression to get name and price
expression = "[*].{name: name, price: current_price}"

extracted_data = market_data.find(expression)

for coin in extracted_data:
    print(f"{coin['name']}: ${coin['price']:,}")

## output example:
## bitcoin: $120,000.00
## ethereum: $5,432.00
## dogecoin: $0.123

CLI Usage

• Join path segments

webpath join [https://api.example.com/v1](https://api.example.com/v1) users 42
# [https://api.example.com/v1/users/42](https://api.example.com/v1/users/42)

• Simple-Get

webpath get [https://api.github.com/repos/python/cpython](https://api.github.com/repos/python/cpython) -p | jq '.stargazers_count'

• Get with retries

webpath get [https://httpbin.org/status/503](https://httpbin.org/status/503) --retries 3

• Download

webpath download [https://speed.hetzner.de/100MB.bin](https://speed.hetzner.de/100MB.bin) 100MB.bin --checksum "5be551ef..."