<p align="center">
<img src="https://raw.githubusercontent.com/guilatrova/gracy/main/img/logo.png">
</p>
<h2 align="center">Python's most graceful API Client Framework</h2>
<p align="center">
<!-- CI --><a href="https://github.com/guilatrova/gracy/actions"><img alt="Actions Status" src="https://github.com/guilatrova/gracy/workflows/CI/badge.svg"></a>
<!-- PyPI --><a href="https://pypi.org/project/gracy/"><img alt="PyPI" src="https://img.shields.io/pypi/v/gracy"/></a>
<!-- Supported Python versions --><img src="https://badgen.net/pypi/python/gracy" />
<!-- Alternative Python versioning: <img alt="python version" src="https://img.shields.io/badge/python-3.9%20%7C%203.10-blue"> -->
<!-- PyPI downloads --><a href="https://pepy.tech/project/gracy/"><img alt="Downloads" src="https://static.pepy.tech/badge/gracy/week"/></a>
<!-- LICENSE --><a href="https://github.com/guilatrova/gracy/blob/main/LICENSE"><img alt="GitHub" src="https://img.shields.io/github/license/guilatrova/gracy"/></a>
<!-- Formatting --><a href="https://github.com/psf/black"><img alt="Code style: black" src="https://img.shields.io/badge/code%20style-black-000000.svg"/></a>
<!-- Tryceratops --><a href="https://github.com/guilatrova/tryceratops"><img alt="try/except style: tryceratops" src="https://img.shields.io/badge/try%2Fexcept%20style-tryceratops%20%F0%9F%A6%96%E2%9C%A8-black" /></a>
<!-- Typing --><a href="https://github.com/microsoft/pyright"><img alt="Types: pyright" src="https://img.shields.io/badge/types-pyright-blue.svg"/></a>
<!-- Follow handle --><a href="https://twitter.com/intent/user?screen_name=guilatrova"><img alt="Follow guilatrova" src="https://img.shields.io/twitter/follow/guilatrova?style=social"/></a>
<!-- Sponsor --><a href="https://github.com/sponsors/guilatrova"><img alt="Sponsor guilatrova" src="https://img.shields.io/github/sponsors/guilatrova?logo=GitHub%20Sponsors&style=social"/></a>
</p>
Gracy handles failures, logging, retries, throttling, parsing, and reporting for all your HTTP interactions. Gracy uses [httpx](https://github.com/encode/httpx) under the hood.
> "Let Gracy do the boring stuff while you focus on your application"
---
**Summary**
- [🧑💻 Get started](#-get-started)
- [Installation](#installation)
- [Usage](#usage)
- [Simple example](#simple-example)
- [More examples](#more-examples)
- [Settings](#settings)
- [Strict/Allowed status code](#strictallowed-status-code)
- [Custom Validators](#custom-validators)
- [Parsing](#parsing)
- [Parsing Typing](#parsing-typing)
- [Retry](#retry)
- [Throttling](#throttling)
- [Concurrent Requests](#concurrent-requests)
- [Logging](#logging)
- [Custom Exceptions](#custom-exceptions)
- [Reports](#reports)
- [Logger](#logger)
- [List](#list)
- [Table](#table)
- [Plotly](#plotly)
- [Replay requests](#replay-requests)
- [Recording](#recording)
- [Replay](#replay)
- [Resource Namespacing](#resource-namespacing)
- [Pagination](#pagination)
- [Advanced Usage](#advanced-usage)
- [Customizing/Overriding configs per method](#customizingoverriding-configs-per-method)
- [Customizing HTTPx client](#customizing-httpx-client)
- [Overriding default request timeout](#overriding-default-request-timeout)
- [Creating a custom Replay data source](#creating-a-custom-replay-data-source)
- [Hooks before/after request](#hooks-beforeafter-request)
- [Common Hooks](#common-hooks)
- [`HttpHeaderRetryAfterBackOffHook`](#httpheaderretryafterbackoffhook)
- [`RateLimitBackOffHook`](#ratelimitbackoffhook)
- [📚 Extra Resources](#-extra-resources)
- [Change log](#change-log)
- [License](#license)
- [Credits](#credits)
## 🧑💻 Get started
### Installation
```
pip install gracy
```
OR
```
poetry add gracy
```
### Usage
Examples will be shown using the [PokeAPI](https://pokeapi.co).
#### Simple example
```py
# 0. Import
import asyncio
import typing as t
from gracy import BaseEndpoint, Gracy, GracyConfig, LogEvent, LogLevel
# 1. Define your endpoints
class PokeApiEndpoint(BaseEndpoint):
GET_POKEMON = "/pokemon/{NAME}" # 👈 Put placeholders as needed
# 2. Define your Graceful API
class GracefulPokeAPI(Gracy[str]):
class Config:
BASE_URL = "https://pokeapi.co/api/v2/" # 👈 Optional BASE_URL
# 👇 Define settings to apply for every request
SETTINGS = GracyConfig(
log_request=LogEvent(LogLevel.DEBUG),
log_response=LogEvent(LogLevel.INFO, "{URL} took {ELAPSED}"),
parser={
"default": lambda r: r.json()
}
)
async def get_pokemon(self, name: str) -> t.Awaitable[dict]:
return await self.get(PokeApiEndpoint.GET_POKEMON, {"NAME": name})
pokeapi = GracefulPokeAPI()
async def main():
try:
pokemon = await pokeapi.get_pokemon("pikachu")
print(pokemon)
finally:
pokeapi.report_status("rich")
asyncio.run(main())
```
#### More examples
- [PokeAPI with retries, parsers, logs](./examples/pokeapi.py)
- [PokeAPI with throttling](./examples/pokeapi_throttle.py)
- [PokeAPI with SQLite replay](./examples/pokeapi_replay.py)
- [PokeAPI with Mongo replay](./examples/pokeapi_replay_mongo.py)
## Settings
### Strict/Allowed status code
By default Gracy considers any successful status code (200-299) as successful.
**Strict**
You can modify this behavior by defining a strict status code or increase the range of allowed status codes:
```py
from http import HTTPStatus
GracyConfig(
strict_status_code=HTTPStatus.CREATED
)
```
or a list of values:
```py
from http import HTTPStatus
GracyConfig(
strict_status_code={HTTPStatus.OK, HTTPStatus.CREATED}
)
```
Using `strict_status_code` means that any other code not specified will raise an error regardless of being successful or not.
**Allowed**
You can also keep the behavior, but extend the range of allowed codes.
```py
from http import HTTPStatus
GracyConfig(
allowed_status_code=HTTPStatus.NOT_FOUND
)
```
or a list of values
```py
from http import HTTPStatus
GracyConfig(
allowed_status_code={HTTPStatus.NOT_FOUND, HTTPStatus.FORBIDDEN}
)
```
Using `allowed_status_code` means that all successful codes plus your defined codes will be considered successful.
This is quite useful for parsing as you'll see soon.
⚠️ Note that `strict_status_code` takes precedence over `allowed_status_code`, probably you don't want to combine those. Prefer one or the other.
### Custom Validators
You can implement your own custom validator to do further checks on the response and decide whether to consider the request failed (and as consequence trigger retries if they're set).
```py
from gracy import GracefulValidator
class MyException(Exception):
pass
class MyCustomValidator(GracefulValidator):
def check(self, response: httpx.Response) -> None:
jsonified = response.json()
if jsonified.get('error', None):
raise MyException("Error is not expected")
return None
...
class Config:
SETTINGS = GracyConfig(
...,
retry=GracefulRetry(retry_on=MyException, ...), # Set up retry to work whenever our validator fails
validators=MyCustomValidator(), # Set up validator
)
```
### Parsing
Parsing allows you to handle the request based on the status code returned.
The basic example is parsing `json`:
```py
GracyConfig(
parser={
"default": lambda r: r.json()
}
)
```
In this example all successful requests will automatically return the `json()` result.
You can also narrow it down to handle specific status codes.
```py
class Config:
SETTINGS = GracyConfig(
...,
allowed_status_code=HTTPStatusCode.NOT_FOUND,
parser={
"default": lambda r: r.json()
HTTPStatusCode.NOT_FOUND: None
}
)
async def get_pokemon(self, name: str) -> dict| None:
# 👇 Returns either dict or None
return await self.get(PokeApiEndpoint.GET_POKEMON, {"NAME": name})
```
Or even customize [exceptions to improve your code readability](https://guicommits.com/handling-exceptions-in-python-like-a-pro/):
```py
class PokemonNotFound(GracyUserDefinedException):
... # More on exceptions below
class Config:
GracyConfig(
...,
allowed_status_code=HTTPStatusCode.NOT_FOUND,
parser={
"default": lambda r: r.json()
HTTPStatusCode.NOT_FOUND: PokemonNotFound
}
)
async def get_pokemon(self, name: str) -> Awaitable[dict]:
# 👇 Returns either dict or raises PokemonNotFound
return await self.get(PokeApiEndpoint.GET_POKEMON, {"NAME": name})
```
### Parsing Typing
Because parsers allow you to dynamically parse a payload based on the status code your IDE will not identify the return type by itself.
To avoid boring `typing.cast` for every method, Gracy provides typed http methods, so you can define a specific return type:
```py
async def list(self, offset: int = 0, limit: int = 20):
params = dict(offset=offset, limit=limit)
return await self.get[ResourceList]( # Specifies this method return a `ResourceList`
PokeApiEndpoint.BERRY_LIST, params=params
)
async def get_one(self, name_or_id: str | int):
return await self.get[models.Berry | None](
PokeApiEndpoint.BERRY_GET, params=dict(KEY=str(name_or_id))
)
```
### Retry
Who doesn't hate flaky APIs? 🙋
Yet there're many of them.
Using tenacity, backoff, retry, aiohttp_retry, and any other retry libs is **NOT easy enough**. 🙅
You still would need to code the implementation for each request which is annoying.
Here's how Gracy allows you to implement your retry logic:
```py
class Config:
GracyConfig(
retry=GracefulRetry(
delay=1,
max_attempts=3,
delay_modifier=1.5,
retry_on=None,
log_before=None,
log_after=LogEvent(LogLevel.WARNING),
log_exhausted=LogEvent(LogLevel.CRITICAL),
behavior="break",
)
)
```
| Parameter | Description | Example |
| ---------------- | --------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------ |
| `delay` | How many seconds to wait between retries | `2` would wait 2 seconds, `1.5` would wait 1.5 seconds, and so on |
| `max_attempts` | How many times should Gracy retry the request? | `10` means 1 regular request with additional 10 retries in case they keep failing. `1` should be the minimum |
| `delay_modifier` | Allows you to specify increasing delay times by multiplying this value to `delay` | Setting `1` means no delay change. Setting `2` means delay will be doubled every retry |
| `retry_on` | Should we retry for which status codes/exceptions? `None` means for any non successful status code or exception | `HTTPStatus.BAD_REQUEST`, or `{HTTPStatus.BAD_REQUEST, HTTPStatus.FORBIDDEN}`, or `Exception` or `{Exception, HTTPStatus.NOT_FOUND}` |
| `log_before` | Specify log level. `None` means don't log | More on logging later |
| `log_after` | Specify log level. `None` means don't log | More on logging later |
| `log_exhausted` | Specify log level. `None` means don't log | More on logging later |
| `behavior` | Allows you to define how to deal if the retry fails. `pass` will accept any retry failure | `pass` or `break` (default) |
| `overrides` | Allows to override `delay` based on last response status code | `{HTTPStatus.BAD_REQUEST: OverrideRetryOn(delay=0), HTTPStatus.INTERNAL_SERVER_ERROR: OverrideRetryOn(delay=10)}` |
### Throttling
Rate limiting issues? No more.
Gracy helps you proactively deal with it before any API throws 429 in your face.
**Creating rules**
You can define rules per endpoint using regex:
```py
SIMPLE_RULE = ThrottleRule(
url_pattern=r".*",
max_requests=2
)
print(SIMPLE_RULE)
# Output: "2 requests per second for URLs matching re.compile('.*')"
COMPLEX_RULE = ThrottleRule(
url_pattern=r".*\/pokemon\/.*",
max_requests=10,
per_time=timedelta(minutes=1, seconds=30),
)
print(COMPLEX_RULE)
# Output: 10 requests per 90 seconds for URLs matching re.compile('.*\\/pokemon\\/.*')
```
**Setting throttling**
You can set up logging and assign rules as:
```py
class Config:
GracyConfig(
throttling=GracefulThrottle(
rules=ThrottleRule(r".*", 2), # 2 reqs/s for any endpoint
log_limit_reached=LogEvent(LogLevel.ERROR),
log_wait_over=LogEvent(LogLevel.WARNING),
),
)
```
### Concurrent Requests
Maybe the API you're hitting have some slow endpoints and you want to ensure that no more than a custom number of requests are being made concurrently.
You can define a `ConcurrentRequestLimit` config.
The simplest usage is:
```py
from gracy import ConcurrentRequestLimit
class Config:
GracyConfig(
concurrent_requests=ConcurrentRequestLimit(
limit=1, # How many concurrent requests
log_limit_reached=LogEvent(LogLevel.WARNING),
log_limit_freed=LogEvent(LogLevel.INFO),
),
)
```
But you can also define it easily per method as:
```py
class MyApiClient(Gracy[Endpoint]):
@graceful(concurrent_requests=5)
async def get_concurrently_five(self, name: str):
...
```
### Logging
You can **define and customize logs** for events by using `LogEvent` and `LogLevel`:
```py
verbose_log = LogEvent(LogLevel.CRITICAL)
custom_warn_log = LogEvent(LogLevel.WARNING, custom_message="{METHOD} {URL} is quite slow and flaky")
custom_error_log = LogEvent(LogLevel.INFO, custom_message="{URL} returned a bad status code {STATUS}, but that's fine")
```
Note that placeholders are formatted and replaced later on by Gracy based on the event type, like:
**Placeholders per event**
| Placeholder | Description | Example | Supported Events |
| ----------------------- | ------------------------------------------------------------- | -------------------------------------------------------------- | -------------------- |
| `{URL}` | Full url being targetted | `https://pokeapi.co/api/v2/pokemon/pikachu` | *All* |
| `{UURL}` | Full **Unformatted** url being targetted | `https://pokeapi.co/api/v2/pokemon/{NAME}` | *All* |
| `{ENDPOINT}` | Endpoint being targetted | `/pokemon/pikachu` | *All* |
| `{UENDPOINT}` | **Unformatted** endpoint being targetted | `/pokemon/{NAME}` | *All* |
| `{METHOD}` | HTTP Request being used | `GET`, `POST` | *All* |
| `{STATUS}` | Status code returned by the response | `200`, `404`, `501` | *After Request* |
| `{ELAPSED}` | Amount of seconds taken for the request to complete | *Numeric* | *After Request* |
| `{REPLAY}` | A placeholder that is displayed only when request is replayed | `REPLAYED` when replay, otherwise it's a blank str (``) | *After Request* |
| `{IS_REPLAY}` | Boolean value to show whether it's replayed or not | String with `TRUE` when replayed or `FALSE` | *After Request* |
| `{RETRY_DELAY}` | How long Gracy will wait before repeating the request | *Numeric* | *Any Retry event* |
| `{RETRY_CAUSE}` | What caused the retry logic to trigger | `[Bad Status Code: 404]`, `[Request Error: ConnectionTimeout]` | *Any Retry event* |
| `{CUR_ATTEMPT}` | Current attempt count for the current request | *Numeric* | *Any Retry event* |
| `{MAX_ATTEMPT}` | Max attempt defined for the current request | *Numeric* | *Any Retry event* |
| `{THROTTLE_LIMIT}` | How many reqs/s is defined for the current request | *Numeric* | *Any Throttle event* |
| `{THROTTLE_TIME}` | How long Gracy will wait before calling the request | *Numeric* | *Any Throttle event* |
| `{THROTTLE_TIME_RANGE}` | Time range defined by the throttling rule | `second`, `90 seconds` | *Any Throttle event* |
and you can set up the log events as follows:
**Requests**
1. Before request
2. After response
3. Response has non successful errors
```py
GracyConfig(
log_request=LogEvent(),
log_response=LogEvent(),
log_errors=LogEvent(),
)
```
**Retry**
1. Before retry
2. After retry
3. When retry exhausted
```py
GracefulRetry(
...,
log_before=LogEvent(),
log_after=LogEvent(),
log_exhausted=LogEvent(),
)
```
**Throttling**
1. When reqs/s limit is reached
2. When limit decreases again
```py
GracefulThrottle(
...,
log_limit_reached=LogEvent()
log_wait_over=LogEvent()
)
```
**Dynamic Customization**
You can customize it even further by passing a lambda:
```py
LogEvent(
LogLevel.ERROR,
lambda r: "Request failed with {STATUS}" f" and it was {'redirected' if r.is_redirect else 'NOT redirected'}"
if r
else "",
)
```
Consider that:
- Not all log events have the response available, so you need to guard yourself against it
- Placeholders still works (e.g. `{STATUS}`)
- You need to watch out for some attrs that might break the formatting logic (e.g. `r.headers`)
### Custom Exceptions
You can define custom exceptions for more [fine grained control over your exception messages/types](https://guicommits.com/how-to-structure-exception-in-python-like-a-pro/).
The simplest you can do is:
```py
from gracy import Gracy, GracyConfig
from gracy.exceptions import GracyUserDefinedException
class MyCustomException(GracyUserDefinedException):
pass
class MyApi(Gracy[str]):
class Config:
SETTINGS = GracyConfig(
...,
parser={
HTTPStatus.BAD_REQUEST: MyCustomException
}
)
```
This will raise your custom exception under the conditions defined in your parser.
You can improve it even further by customizing your message:
```py
class PokemonNotFound(GracyUserDefinedException):
BASE_MESSAGE = "Unable to find a pokemon with the name [{NAME}] at {URL} due to {STATUS} status"
def _format_message(self, request_context: GracyRequestContext, response: httpx.Response) -> str:
format_args = self._build_default_args()
name = request_context.endpoint_args.get("NAME", "Unknown")
return self.BASE_MESSAGE.format(NAME=name, **format_args)
```
## Reports
### Logger
Recommended for production environments.
Gracy reports a short summary using `logger.info`.
```python
pokeapi = GracefulPokeAPI()
# do stuff with your API
pokeapi.report_status("logger")
# OUTPUT
❯ Gracy tracked that 'https://pokeapi.co/api/v2/pokemon/{NAME}' was hit 1 time(s) with a success rate of 100.00%, avg latency of 0.45s, and a rate of 1.0 reqs/s.
❯ Gracy tracked a total of 2 requests with a success rate of 100.00%, avg latency of 0.24s, and a rate of 1.0 reqs/s.
```
### List
Uses `print` to generate a short list with all attributes:
```python
pokeapi = GracefulPokeAPI()
# do stuff with your API
pokeapi.report_status("list")
# OUTPUT
____
/ ___|_ __ __ _ ___ _ _
| | _| '__/ _` |/ __| | | |
| |_| | | | (_| | (__| |_| |
\____|_| \__,_|\___|\__, |
|___/ Requests Summary Report
1. https://pokeapi.co/api/v2/pokemon/{NAME}
Total Reqs (#): 1
Success (%): 100.00%
Fail (%): 0.00%
Avg Latency (s): 0.39
Max Latency (s): 0.39
2xx Resps: 1
3xx Resps: 0
4xx Resps: 0
5xx Resps: 0
Avg Reqs/sec: 1.0 reqs/s
2. https://pokeapi.co/api/v2/generation/{ID}/
Total Reqs (#): 1
Success (%): 100.00%
Fail (%): 0.00%
Avg Latency (s): 0.04
Max Latency (s): 0.04
2xx Resps: 1
3xx Resps: 0
4xx Resps: 0
5xx Resps: 0
Avg Reqs/sec: 1.0 reqs/s
TOTAL
Total Reqs (#): 2
Success (%): 100.00%
Fail (%): 0.00%
Avg Latency (s): 0.21
Max Latency (s): 0.00
2xx Resps: 2
3xx Resps: 0
4xx Resps: 0
5xx Resps: 0
Avg Reqs/sec: 1.0 reqs/s
```
### Table
It requires you to install [Rich](https://github.com/Textualize/rich).
```py
pokeapi = GracefulPokeAPI()
# do stuff with your API
pokeapi.report_status("rich")
```
Here's an example of how it looks:
### Plotly
It requires you to install [plotly 📊](https://github.com/plotly/plotly.py) and [pandas 🐼](https://github.com/pandas-dev/pandas).
```py
pokeapi = GracefulPokeAPI()
# do stuff with your API
plotly_fig = pokeapi.report_status("plotly")
plotly_fig.show()
```
Here's an example of how it looks:
## Replay requests
Gracy allows you to replay requests and responses from previous interactions.
This is powerful because it allows you to test APIs without latency or consuming your rate limit. Now writing unit tests that relies on third-party APIs is doable.
It works in two steps:
| **Step** | **Description** | **Hits the API?** |
| ------------ | ------------------------------------------------------------------------------ | ----------------- |
| 1. Recording | Stores all requests/responses to be later replayed | **Yes** |
| 2. Replay | Returns all previously generated responses based on your request as a "replay" | No |
### Recording
The effort to record requests/responses is ZERO. You just need to pass a recording config to your Graceful API:
```py
from gracy import GracyReplay
from gracy.replays.storages.sqlite import SQLiteReplayStorage
record_mode = GracyReplay("record", SQLiteReplayStorage("pokeapi.sqlite3"))
pokeapi = GracefulPokeAPI(record_mode)
```
**Every request** will be recorded to the defined data source.
### Replay
Once you have recorded all your requests you can enable the replay mode:
```py
from gracy import GracyReplay
from gracy.replays.storages.sqlite import SQLiteReplayStorage
replay_mode = GracyReplay("replay", SQLiteReplayStorage("pokeapi.sqlite3"))
pokeapi = GracefulPokeAPI(replay_mode)
```
**Every request** will be routed to the defined data source resulting in faster responses.
**⚠️ Note that parsers, retries, throttling, and similar configs will work as usual**.
## Resource Namespacing
You can have multiple namespaces to organize your API endpoints as you wish.
To do so, you just have to inherit from `GracyNamespace` and instantiate it within the `GracyAPI`:
```py
from gracy import Gracy, GracyNamespace, GracyConfig
class PokemonNamespace(GracyNamespace[PokeApiEndpoint]):
async def get_one(self, name: str):
return await self.get(PokeApiEndpoint.GET_POKEMON, {"NAME": name})
class BerryNamespace(GracyNamespace[PokeApiEndpoint]):
async def get_one(self, name: str):
return await self.get(PokeApiEndpoint.GET_BERRY, {"NAME": name})
class GracefulPokeAPI(Gracy[PokeApiEndpoint]):
class Config:
BASE_URL = "https://pokeapi.co/api/v2/"
SETTINGS = GracyConfig(
retry=RETRY,
allowed_status_code={HTTPStatus.NOT_FOUND},
parser={HTTPStatus.NOT_FOUND: None},
)
# These will be automatically assigned on init
berry: BerryNamespace
pokemon: PokemonNamespace
```
And the usage will work as:
```py
await pokeapi.pokemon.get_one("pikachu")
await pokeapi.berry.get_one("cheri")
```
Note all configs are propagated to namespaces, but namespaces can still have their own which would cause merges when instantiatedg.
## Pagination
There're endpoints that may require pagination. For that you can use `GracyPaginator`.
For a simple case where you pass `offset` and `limit`, you can use `GracyOffsetPaginator`:
```py
from gracy import GracyOffsetPaginator
class BerryNamespace(GracyNamespace[PokeApiEndpoint]):
@parsed_response(ResourceList)
async def list(self, offset: int = 0, limit: int = 20):
params = dict(offset=offset, limit=limit)
return await self.get(PokeApiEndpoint.BERRY_LIST, params=params)
def paginate(self, limit: int = 20) -> GracyOffsetPaginator[ResourceList]:
return GracyOffsetPaginator[ResourceList](
gracy_func=self.list,
has_next=lambda r: bool(r["next"]) if r else True,
page_size=limit,
)
```
and then use it as:
```py
async def main():
api = PokeApi()
paginator = api.berry.paginate(2)
# Just grabs the next page
first = await paginator.next_page()
print(first)
# Resets current page to 0
paginator.set_page(0)
# Loop throught it all
async for page in paginator:
print(page)
```
## Advanced Usage
### Customizing/Overriding configs per method
APIs may return different responses/conditions/payloads based on the endpoint.
You can override any `GracyConfig` on a per method basis by using the `@graceful` decorator.
NOTE: Use `@graceful_generator` if your function uses `yield`.
```python
from gracy import Gracy, GracyConfig, GracefulRetry, graceful, graceful_generator
retry = GracefulRetry(...)
class GracefulPokeAPI(Gracy[PokeApiEndpoint]):
class Config:
BASE_URL = "https://pokeapi.co/api/v2/"
SETTINGS = GracyConfig(
retry=retry,
log_errors=LogEvent(
LogLevel.ERROR, "How can I become a pokemon master if {URL} keeps failing with {STATUS}"
),
)
@graceful(
retry=None, # 👈 Disables retry set in Config
log_errors=None, # 👈 Disables log_errors set in Config
allowed_status_code=HTTPStatus.NOT_FOUND,
parser={
"default": lambda r: r.json()["order"],
HTTPStatus.NOT_FOUND: None,
},
)
async def maybe_get_pokemon_order(self, name: str):
val: str | None = await self.get(PokeApiEndpoint.GET_POKEMON, {"NAME": name})
return val
@graceful( # 👈 Retry and log_errors are still set for this one
strict_status_code=HTTPStatus.OK,
parser={"default": lambda r: r.json()["order"]},
)
async def get_pokemon_order(self, name: str):
val: str = await self.get(PokeApiEndpoint.GET_POKEMON, {"NAME": name})
return val
@graceful_generator( # 👈 Retry and log_errors are still set for this one
parser={"default": lambda r: r.json()["order"]},
)
async def get_2_pokemons(self):
names = ["charmander", "pikachu"]
for name in names:
r = await self.get(PokeApiEndpoint.GET_POKEMON, {"NAME": name})
yield r
```
### Customizing HTTPx client
You might want to modify the HTTPx client settings, do so by:
```py
class YourAPIClient(Gracy[str]):
class Config:
...
def __init__(self, token: token) -> None:
self._token = token
super().__init__()
# 👇 Implement your logic here
def _create_client(self) -> httpx.AsyncClient:
client = super()._create_client()
client.headers = {"Authorization": f"token {self._token}"} # type: ignore
return client
```
### Overriding default request timeout
As default Gracy won't enforce a request timeout.
You can define your own by setting it on Config as:
```py
class GracefulAPI(GracyApi[str]):
class Config:
BASE_URL = "https://example.com"
REQUEST_TIMEOUT = 10.2 # 👈 Here
```
### Creating a custom Replay data source
Gracy was built with extensibility in mind.
You can create your own storage to store/load anywhere (e.g. SQL Database), here's an example:
```py
import httpx
from gracy import GracyReplayStorage
class MyCustomStorage(GracyReplayStorage):
def prepare(self) -> None: # (Optional) Executed upon API instance creation.
...
async def record(self, response: httpx.Response) -> None:
... # REQUIRED. Your logic to store the response object. Note the httpx.Response has request data.
async def _load(self, request: httpx.Request) -> httpx.Response:
... # REQUIRED. Your logic to load a response object based on the request.
# Usage
record_mode = GracyReplay("record", MyCustomStorage())
replay_mode = GracyReplay("replay", MyCustomStorage())
pokeapi = GracefulPokeAPI(record_mode)
```
### Hooks before/after request
You can set up hooks simply by defining `async def before` and `async def after` methods.
⚠️ NOTE: Gracy configs are disabled within these methods which means that retries/parsers/throttling won't take effect inside it.
```py
class GracefulPokeAPI(Gracy[PokeApiEndpoint]):
class Config:
BASE_URL = "https://pokeapi.co/api/v2/"
SETTINGS = GracyConfig(
retry=RETRY,
allowed_status_code={HTTPStatus.NOT_FOUND},
parser={HTTPStatus.NOT_FOUND: None},
)
def __init__(self, *args: t.Any, **kwargs: t.Any) -> None:
self.before_count = 0
self.after_status_counter = defaultdict[HTTPStatus, int](int)
self.after_aborts = 0
self.after_retries_counter = 0
super().__init__(*args, **kwargs)
async def before(self, context: GracyRequestContext):
self.before_count += 1
async def after(
self,
context: GracyRequestContext, # Current request context
response_or_exc: httpx.Response | Exception, # Either the request or an error
retry_state: GracefulRetryState | None, # Set when this is generated from a retry
):
if retry_state:
self.after_retries_counter += 1
if isinstance(response_or_exc, httpx.Response):
self.after_status_counter[HTTPStatus(response_or_exc.status_code)] += 1
else:
self.after_aborts += 1
async def get_pokemon(self, name: str):
return await self.get(PokeApiEndpoint.GET_POKEMON, {"NAME": name})
```
In the example above invoking `get_pokemon()` will trigger `before()`/`after()` hooks in sequence.
#### Common Hooks
##### `HttpHeaderRetryAfterBackOffHook`
This hook checks for 429 (TOO MANY REQUESTS), and then reads the
[`retry-after` header](https://developer.mozilla.org/en-US/docs/Web/HTTP/Headers/Retry-After).
If the value is set, then Gracy pauses **ALL** client requests until the time is over. This behavior can be modified to happen on a per-endpoint basis if `lock_per_endpoint` is True.
Example Usage:
```py
from gracy.common_hooks import HttpHeaderRetryAfterBackOffHook
class GracefulAPI(GracyAPI[Endpoint]):
def __init__(self):
self._retry_after_hook = HttpHeaderRetryAfterBackOffHook(
self._reporter,
lock_per_endpoint=True,
log_event=LogEvent(
LogLevel.WARNING,
custom_message=(
"{ENDPOINT} produced {STATUS} and requested to wait {RETRY_AFTER}s "
"- waiting {RETRY_AFTER_ACTUAL_WAIT}s"
),
),
# Wait +10s to avoid this from happening again too soon
seconds_processor=lambda secs_requested: secs_requested + 10,
)
super().__init__()
async def before(self, context: GracyRequestContext):
await self._retry_after_hook.before(context)
async def after(
self,
context: GracyRequestContext,
response_or_exc: httpx.Response | Exception,
retry_state: GracefulRetryState | None,
):
retry_after_result = await self._retry_after_hook.after(context, response_or_exc)
```
##### `RateLimitBackOffHook`
This hook checks for 429 (TOO MANY REQUESTS) and locks requests for an arbitrary amount of time defined by you.
If the value is set, then Gracy pauses **ALL** client requests until the time is over.
This behavior can be modified to happen on a per-endpoint basis if `lock_per_endpoint` is True.
```py
from gracy.common_hooks import RateLimitBackOffHook
class GracefulAPI(GracyAPI[Endpoint]):
def __init__(self):
self._ratelimit_backoff_hook = RateLimitBackOffHook(
30,
self._reporter,
lock_per_endpoint=True,
log_event=LogEvent(
LogLevel.INFO,
custom_message="{UENDPOINT} got rate limited, waiting for {WAIT_TIME}s",
),
)
super().__init__()
async def before(self, context: GracyRequestContext):
await self._ratelimit_backoff_hook.before(context)
async def after(
self,
context: GracyRequestContext,
response_or_exc: httpx.Response | Exception,
retry_state: GracefulRetryState | None,
):
backoff_result = await self._ratelimit_backoff_hook.after(context, response_or_exc)
```
```py
from gracy.common_hooks import HttpHeaderRetryAfterBackOffHook, RateLimitBackOffHook
```
## 📚 Extra Resources
Some good practices I learned over the past years guided Gracy's philosophy, you might benefit by reading:
- [How to log](https://guicommits.com/how-to-log-in-python-like-a-pro/)
- [How to handle exceptions](https://guicommits.com/handling-exceptions-in-python-like-a-pro/)
- [How to structure exceptions](https://guicommits.com/how-to-structure-exception-in-python-like-a-pro/)
- [How to use Async correctly](https://guicommits.com/effective-python-async-like-a-pro/)
- [Book: Python like a PRO](https://guilatrova.gumroad.com/l/python-like-a-pro)
- [Book: Effective Python](https://amzn.to/3bEVHpG)
<!-- ## Contributing -->
<!-- Thank you for considering making Gracy better for everyone! -->
<!-- Refer to [Contributing docs](docs/CONTRIBUTING.md).-->
## Change log
See [CHANGELOG](CHANGELOG.md).
## License
MIT
## Credits
Thanks to the last three startups I worked which forced me to do the same things and resolve the same problems over and over again. I got sick of it and built this lib.
Most importantly: **Thanks to God**, who allowed me (a random 🇧🇷 guy) to work for many different 🇺🇸 startups. This is ironic since due to God's grace, I was able to build Gracy. 🙌
Also, thanks to the [httpx](https://github.com/encode/httpx) and [rich](https://github.com/Textualize/rich) projects for the beautiful and simple APIs that powers Gracy.
Raw data
{
"_id": null,
"home_page": "https://github.com/guilatrova/gracy",
"name": "gracy",
"maintainer": null,
"docs_url": null,
"requires_python": "<4.0,>=3.8.1",
"maintainer_email": null,
"keywords": "api, throttling, http, https, async, retry",
"author": "Guilherme Latrova",
"author_email": "hello@guilatrova.dev",
"download_url": "https://files.pythonhosted.org/packages/34/b9/5f48bbeaf8ae9a8ae40c6cce5d52291eebd7fd6b698684cfb5e7a46f210d/gracy-1.33.1.tar.gz",
"platform": null,
"description": "<p align=\"center\">\n <img src=\"https://raw.githubusercontent.com/guilatrova/gracy/main/img/logo.png\">\n</p>\n\n<h2 align=\"center\">Python's most graceful API Client Framework</h2>\n\n<p align=\"center\">\n <!-- CI --><a href=\"https://github.com/guilatrova/gracy/actions\"><img alt=\"Actions Status\" src=\"https://github.com/guilatrova/gracy/workflows/CI/badge.svg\"></a>\n <!-- PyPI --><a href=\"https://pypi.org/project/gracy/\"><img alt=\"PyPI\" src=\"https://img.shields.io/pypi/v/gracy\"/></a>\n <!-- Supported Python versions --><img src=\"https://badgen.net/pypi/python/gracy\" />\n <!-- Alternative Python versioning: <img alt=\"python version\" src=\"https://img.shields.io/badge/python-3.9%20%7C%203.10-blue\"> -->\n <!-- PyPI downloads --><a href=\"https://pepy.tech/project/gracy/\"><img alt=\"Downloads\" src=\"https://static.pepy.tech/badge/gracy/week\"/></a>\n <!-- LICENSE --><a href=\"https://github.com/guilatrova/gracy/blob/main/LICENSE\"><img alt=\"GitHub\" src=\"https://img.shields.io/github/license/guilatrova/gracy\"/></a>\n <!-- Formatting --><a href=\"https://github.com/psf/black\"><img alt=\"Code style: black\" src=\"https://img.shields.io/badge/code%20style-black-000000.svg\"/></a>\n <!-- Tryceratops --><a href=\"https://github.com/guilatrova/tryceratops\"><img alt=\"try/except style: tryceratops\" src=\"https://img.shields.io/badge/try%2Fexcept%20style-tryceratops%20%F0%9F%A6%96%E2%9C%A8-black\" /></a>\n <!-- Typing --><a href=\"https://github.com/microsoft/pyright\"><img alt=\"Types: pyright\" src=\"https://img.shields.io/badge/types-pyright-blue.svg\"/></a>\n <!-- Follow handle --><a href=\"https://twitter.com/intent/user?screen_name=guilatrova\"><img alt=\"Follow guilatrova\" src=\"https://img.shields.io/twitter/follow/guilatrova?style=social\"/></a>\n <!-- Sponsor --><a href=\"https://github.com/sponsors/guilatrova\"><img alt=\"Sponsor guilatrova\" src=\"https://img.shields.io/github/sponsors/guilatrova?logo=GitHub%20Sponsors&style=social\"/></a>\n</p>\n\nGracy handles failures, logging, retries, throttling, parsing, and reporting for all your HTTP interactions. Gracy uses [httpx](https://github.com/encode/httpx) under the hood.\n\n> \"Let Gracy do the boring stuff while you focus on your application\"\n\n---\n\n**Summary**\n\n- [\ud83e\uddd1\u200d\ud83d\udcbb Get started](#-get-started)\n - [Installation](#installation)\n - [Usage](#usage)\n - [Simple example](#simple-example)\n - [More examples](#more-examples)\n- [Settings](#settings)\n - [Strict/Allowed status code](#strictallowed-status-code)\n - [Custom Validators](#custom-validators)\n - [Parsing](#parsing)\n - [Parsing Typing](#parsing-typing)\n - [Retry](#retry)\n - [Throttling](#throttling)\n - [Concurrent Requests](#concurrent-requests)\n - [Logging](#logging)\n - [Custom Exceptions](#custom-exceptions)\n- [Reports](#reports)\n - [Logger](#logger)\n - [List](#list)\n - [Table](#table)\n - [Plotly](#plotly)\n- [Replay requests](#replay-requests)\n - [Recording](#recording)\n - [Replay](#replay)\n- [Resource Namespacing](#resource-namespacing)\n- [Pagination](#pagination)\n- [Advanced Usage](#advanced-usage)\n - [Customizing/Overriding configs per method](#customizingoverriding-configs-per-method)\n - [Customizing HTTPx client](#customizing-httpx-client)\n - [Overriding default request timeout](#overriding-default-request-timeout)\n - [Creating a custom Replay data source](#creating-a-custom-replay-data-source)\n - [Hooks before/after request](#hooks-beforeafter-request)\n - [Common Hooks](#common-hooks)\n - [`HttpHeaderRetryAfterBackOffHook`](#httpheaderretryafterbackoffhook)\n - [`RateLimitBackOffHook`](#ratelimitbackoffhook)\n- [\ud83d\udcda Extra Resources](#-extra-resources)\n- [Change log](#change-log)\n- [License](#license)\n- [Credits](#credits)\n\n\n## \ud83e\uddd1\u200d\ud83d\udcbb Get started\n\n### Installation\n\n```\npip install gracy\n```\n\nOR\n\n```\npoetry add gracy\n```\n\n### Usage\n\nExamples will be shown using the [PokeAPI](https://pokeapi.co).\n\n#### Simple example\n\n```py\n# 0. Import\nimport asyncio\nimport typing as t\nfrom gracy import BaseEndpoint, Gracy, GracyConfig, LogEvent, LogLevel\n\n# 1. Define your endpoints\nclass PokeApiEndpoint(BaseEndpoint):\n GET_POKEMON = \"/pokemon/{NAME}\" # \ud83d\udc48 Put placeholders as needed\n\n# 2. Define your Graceful API\nclass GracefulPokeAPI(Gracy[str]):\n class Config:\n BASE_URL = \"https://pokeapi.co/api/v2/\" # \ud83d\udc48 Optional BASE_URL\n # \ud83d\udc47 Define settings to apply for every request\n SETTINGS = GracyConfig(\n log_request=LogEvent(LogLevel.DEBUG),\n log_response=LogEvent(LogLevel.INFO, \"{URL} took {ELAPSED}\"),\n parser={\n \"default\": lambda r: r.json()\n }\n )\n\n async def get_pokemon(self, name: str) -> t.Awaitable[dict]:\n return await self.get(PokeApiEndpoint.GET_POKEMON, {\"NAME\": name})\n\npokeapi = GracefulPokeAPI()\n\nasync def main():\n try:\n pokemon = await pokeapi.get_pokemon(\"pikachu\")\n print(pokemon)\n\n finally:\n pokeapi.report_status(\"rich\")\n\n\nasyncio.run(main())\n```\n\n#### More examples\n\n- [PokeAPI with retries, parsers, logs](./examples/pokeapi.py)\n- [PokeAPI with throttling](./examples/pokeapi_throttle.py)\n- [PokeAPI with SQLite replay](./examples/pokeapi_replay.py)\n- [PokeAPI with Mongo replay](./examples/pokeapi_replay_mongo.py)\n\n## Settings\n\n### Strict/Allowed status code\n\nBy default Gracy considers any successful status code (200-299) as successful.\n\n**Strict**\n\nYou can modify this behavior by defining a strict status code or increase the range of allowed status codes:\n\n```py\nfrom http import HTTPStatus\n\nGracyConfig(\n strict_status_code=HTTPStatus.CREATED\n)\n```\n\nor a list of values:\n\n```py\nfrom http import HTTPStatus\n\nGracyConfig(\n strict_status_code={HTTPStatus.OK, HTTPStatus.CREATED}\n)\n```\n\nUsing `strict_status_code` means that any other code not specified will raise an error regardless of being successful or not.\n\n**Allowed**\n\nYou can also keep the behavior, but extend the range of allowed codes.\n\n```py\nfrom http import HTTPStatus\n\nGracyConfig(\n allowed_status_code=HTTPStatus.NOT_FOUND\n)\n```\n\nor a list of values\n\n\n```py\nfrom http import HTTPStatus\n\nGracyConfig(\n allowed_status_code={HTTPStatus.NOT_FOUND, HTTPStatus.FORBIDDEN}\n)\n```\n\nUsing `allowed_status_code` means that all successful codes plus your defined codes will be considered successful.\n\nThis is quite useful for parsing as you'll see soon.\n\n\u26a0\ufe0f Note that `strict_status_code` takes precedence over `allowed_status_code`, probably you don't want to combine those. Prefer one or the other.\n\n### Custom Validators\n\nYou can implement your own custom validator to do further checks on the response and decide whether to consider the request failed (and as consequence trigger retries if they're set).\n\n```py\nfrom gracy import GracefulValidator\n\nclass MyException(Exception):\n pass\n\nclass MyCustomValidator(GracefulValidator):\n def check(self, response: httpx.Response) -> None:\n jsonified = response.json()\n if jsonified.get('error', None):\n raise MyException(\"Error is not expected\")\n\n return None\n\n...\n\nclass Config:\n SETTINGS = GracyConfig(\n ...,\n retry=GracefulRetry(retry_on=MyException, ...), # Set up retry to work whenever our validator fails\n validators=MyCustomValidator(), # Set up validator\n )\n\n```\n\n### Parsing\n\nParsing allows you to handle the request based on the status code returned.\n\nThe basic example is parsing `json`:\n\n```py\nGracyConfig(\n parser={\n \"default\": lambda r: r.json()\n }\n)\n```\n\nIn this example all successful requests will automatically return the `json()` result.\n\nYou can also narrow it down to handle specific status codes.\n\n```py\nclass Config:\n SETTINGS = GracyConfig(\n ...,\n allowed_status_code=HTTPStatusCode.NOT_FOUND,\n parser={\n \"default\": lambda r: r.json()\n HTTPStatusCode.NOT_FOUND: None\n }\n )\n\nasync def get_pokemon(self, name: str) -> dict| None:\n # \ud83d\udc47 Returns either dict or None\n return await self.get(PokeApiEndpoint.GET_POKEMON, {\"NAME\": name})\n```\n\nOr even customize [exceptions to improve your code readability](https://guicommits.com/handling-exceptions-in-python-like-a-pro/):\n\n```py\nclass PokemonNotFound(GracyUserDefinedException):\n ... # More on exceptions below\n\nclass Config:\n GracyConfig(\n ...,\n allowed_status_code=HTTPStatusCode.NOT_FOUND,\n parser={\n \"default\": lambda r: r.json()\n HTTPStatusCode.NOT_FOUND: PokemonNotFound\n }\n )\n\nasync def get_pokemon(self, name: str) -> Awaitable[dict]:\n # \ud83d\udc47 Returns either dict or raises PokemonNotFound\n return await self.get(PokeApiEndpoint.GET_POKEMON, {\"NAME\": name})\n```\n\n### Parsing Typing\n\nBecause parsers allow you to dynamically parse a payload based on the status code your IDE will not identify the return type by itself.\n\nTo avoid boring `typing.cast` for every method, Gracy provides typed http methods, so you can define a specific return type:\n\n```py\nasync def list(self, offset: int = 0, limit: int = 20):\n params = dict(offset=offset, limit=limit)\n return await self.get[ResourceList]( # Specifies this method return a `ResourceList`\n PokeApiEndpoint.BERRY_LIST, params=params\n )\n\nasync def get_one(self, name_or_id: str | int):\n return await self.get[models.Berry | None](\n PokeApiEndpoint.BERRY_GET, params=dict(KEY=str(name_or_id))\n )\n```\n\n### Retry\n\nWho doesn't hate flaky APIs? \ud83d\ude4b\n\nYet there're many of them.\n\nUsing tenacity, backoff, retry, aiohttp_retry, and any other retry libs is **NOT easy enough**. \ud83d\ude45\n\nYou still would need to code the implementation for each request which is annoying.\n\nHere's how Gracy allows you to implement your retry logic:\n\n```py\nclass Config:\n GracyConfig(\n retry=GracefulRetry(\n delay=1,\n max_attempts=3,\n delay_modifier=1.5,\n retry_on=None,\n log_before=None,\n log_after=LogEvent(LogLevel.WARNING),\n log_exhausted=LogEvent(LogLevel.CRITICAL),\n behavior=\"break\",\n )\n )\n```\n\n| Parameter | Description | Example |\n| ---------------- | --------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------ |\n| `delay` | How many seconds to wait between retries | `2` would wait 2 seconds, `1.5` would wait 1.5 seconds, and so on |\n| `max_attempts` | How many times should Gracy retry the request? | `10` means 1 regular request with additional 10 retries in case they keep failing. `1` should be the minimum |\n| `delay_modifier` | Allows you to specify increasing delay times by multiplying this value to `delay` | Setting `1` means no delay change. Setting `2` means delay will be doubled every retry |\n| `retry_on` | Should we retry for which status codes/exceptions? `None` means for any non successful status code or exception | `HTTPStatus.BAD_REQUEST`, or `{HTTPStatus.BAD_REQUEST, HTTPStatus.FORBIDDEN}`, or `Exception` or `{Exception, HTTPStatus.NOT_FOUND}` |\n| `log_before` | Specify log level. `None` means don't log | More on logging later |\n| `log_after` | Specify log level. `None` means don't log | More on logging later |\n| `log_exhausted` | Specify log level. `None` means don't log | More on logging later |\n| `behavior` | Allows you to define how to deal if the retry fails. `pass` will accept any retry failure | `pass` or `break` (default) |\n| `overrides` | Allows to override `delay` based on last response status code | `{HTTPStatus.BAD_REQUEST: OverrideRetryOn(delay=0), HTTPStatus.INTERNAL_SERVER_ERROR: OverrideRetryOn(delay=10)}` |\n\n\n### Throttling\n\nRate limiting issues? No more.\n\nGracy helps you proactively deal with it before any API throws 429 in your face.\n\n**Creating rules**\n\nYou can define rules per endpoint using regex:\n\n```py\nSIMPLE_RULE = ThrottleRule(\n url_pattern=r\".*\",\n max_requests=2\n)\nprint(SIMPLE_RULE)\n# Output: \"2 requests per second for URLs matching re.compile('.*')\"\n\nCOMPLEX_RULE = ThrottleRule(\n url_pattern=r\".*\\/pokemon\\/.*\",\n max_requests=10,\n per_time=timedelta(minutes=1, seconds=30),\n)\nprint(COMPLEX_RULE)\n# Output: 10 requests per 90 seconds for URLs matching re.compile('.*\\\\/pokemon\\\\/.*')\n```\n\n**Setting throttling**\n\nYou can set up logging and assign rules as:\n\n```py\nclass Config:\n GracyConfig(\n throttling=GracefulThrottle(\n rules=ThrottleRule(r\".*\", 2), # 2 reqs/s for any endpoint\n log_limit_reached=LogEvent(LogLevel.ERROR),\n log_wait_over=LogEvent(LogLevel.WARNING),\n ),\n )\n```\n\n### Concurrent Requests\n\nMaybe the API you're hitting have some slow endpoints and you want to ensure that no more than a custom number of requests are being made concurrently.\n\nYou can define a `ConcurrentRequestLimit` config.\n\nThe simplest usage is:\n\n```py\nfrom gracy import ConcurrentRequestLimit\n\n\nclass Config:\n GracyConfig(\n concurrent_requests=ConcurrentRequestLimit(\n limit=1, # How many concurrent requests\n log_limit_reached=LogEvent(LogLevel.WARNING),\n log_limit_freed=LogEvent(LogLevel.INFO),\n ),\n )\n```\n\nBut you can also define it easily per method as:\n\n```py\nclass MyApiClient(Gracy[Endpoint]):\n\n @graceful(concurrent_requests=5)\n async def get_concurrently_five(self, name: str):\n ...\n```\n\n### Logging\n\nYou can **define and customize logs** for events by using `LogEvent` and `LogLevel`:\n\n```py\nverbose_log = LogEvent(LogLevel.CRITICAL)\ncustom_warn_log = LogEvent(LogLevel.WARNING, custom_message=\"{METHOD} {URL} is quite slow and flaky\")\ncustom_error_log = LogEvent(LogLevel.INFO, custom_message=\"{URL} returned a bad status code {STATUS}, but that's fine\")\n```\n\nNote that placeholders are formatted and replaced later on by Gracy based on the event type, like:\n\n**Placeholders per event**\n\n| Placeholder | Description | Example | Supported Events |\n| ----------------------- | ------------------------------------------------------------- | -------------------------------------------------------------- | -------------------- |\n| `{URL}` | Full url being targetted | `https://pokeapi.co/api/v2/pokemon/pikachu` | *All* |\n| `{UURL}` | Full **Unformatted** url being targetted | `https://pokeapi.co/api/v2/pokemon/{NAME}` | *All* |\n| `{ENDPOINT}` | Endpoint being targetted | `/pokemon/pikachu` | *All* |\n| `{UENDPOINT}` | **Unformatted** endpoint being targetted | `/pokemon/{NAME}` | *All* |\n| `{METHOD}` | HTTP Request being used | `GET`, `POST` | *All* |\n| `{STATUS}` | Status code returned by the response | `200`, `404`, `501` | *After Request* |\n| `{ELAPSED}` | Amount of seconds taken for the request to complete | *Numeric* | *After Request* |\n| `{REPLAY}` | A placeholder that is displayed only when request is replayed | `REPLAYED` when replay, otherwise it's a blank str (``) | *After Request* |\n| `{IS_REPLAY}` | Boolean value to show whether it's replayed or not | String with `TRUE` when replayed or `FALSE` | *After Request* |\n| `{RETRY_DELAY}` | How long Gracy will wait before repeating the request | *Numeric* | *Any Retry event* |\n| `{RETRY_CAUSE}` | What caused the retry logic to trigger | `[Bad Status Code: 404]`, `[Request Error: ConnectionTimeout]` | *Any Retry event* |\n| `{CUR_ATTEMPT}` | Current attempt count for the current request | *Numeric* | *Any Retry event* |\n| `{MAX_ATTEMPT}` | Max attempt defined for the current request | *Numeric* | *Any Retry event* |\n| `{THROTTLE_LIMIT}` | How many reqs/s is defined for the current request | *Numeric* | *Any Throttle event* |\n| `{THROTTLE_TIME}` | How long Gracy will wait before calling the request | *Numeric* | *Any Throttle event* |\n| `{THROTTLE_TIME_RANGE}` | Time range defined by the throttling rule | `second`, `90 seconds` | *Any Throttle event* |\n\nand you can set up the log events as follows:\n\n**Requests**\n\n1. Before request\n2. After response\n3. Response has non successful errors\n\n```py\nGracyConfig(\n log_request=LogEvent(),\n log_response=LogEvent(),\n log_errors=LogEvent(),\n)\n```\n\n**Retry**\n\n1. Before retry\n2. After retry\n3. When retry exhausted\n\n```py\nGracefulRetry(\n ...,\n log_before=LogEvent(),\n log_after=LogEvent(),\n log_exhausted=LogEvent(),\n)\n```\n\n**Throttling**\n\n1. When reqs/s limit is reached\n2. When limit decreases again\n\n```py\nGracefulThrottle(\n ...,\n log_limit_reached=LogEvent()\n log_wait_over=LogEvent()\n)\n```\n\n**Dynamic Customization**\n\nYou can customize it even further by passing a lambda:\n\n```py\nLogEvent(\n LogLevel.ERROR,\n lambda r: \"Request failed with {STATUS}\" f\" and it was {'redirected' if r.is_redirect else 'NOT redirected'}\"\n if r\n else \"\",\n)\n```\n\nConsider that:\n\n- Not all log events have the response available, so you need to guard yourself against it\n- Placeholders still works (e.g. `{STATUS}`)\n- You need to watch out for some attrs that might break the formatting logic (e.g. `r.headers`)\n\n### Custom Exceptions\n\nYou can define custom exceptions for more [fine grained control over your exception messages/types](https://guicommits.com/how-to-structure-exception-in-python-like-a-pro/).\n\nThe simplest you can do is:\n\n```py\nfrom gracy import Gracy, GracyConfig\nfrom gracy.exceptions import GracyUserDefinedException\n\nclass MyCustomException(GracyUserDefinedException):\n pass\n\nclass MyApi(Gracy[str]):\n class Config:\n SETTINGS = GracyConfig(\n ...,\n parser={\n HTTPStatus.BAD_REQUEST: MyCustomException\n }\n )\n```\n\nThis will raise your custom exception under the conditions defined in your parser.\n\nYou can improve it even further by customizing your message:\n\n```py\nclass PokemonNotFound(GracyUserDefinedException):\n BASE_MESSAGE = \"Unable to find a pokemon with the name [{NAME}] at {URL} due to {STATUS} status\"\n\n def _format_message(self, request_context: GracyRequestContext, response: httpx.Response) -> str:\n format_args = self._build_default_args()\n name = request_context.endpoint_args.get(\"NAME\", \"Unknown\")\n return self.BASE_MESSAGE.format(NAME=name, **format_args)\n```\n\n## Reports\n\n### Logger\n\nRecommended for production environments.\n\nGracy reports a short summary using `logger.info`.\n\n```python\npokeapi = GracefulPokeAPI()\n# do stuff with your API\npokeapi.report_status(\"logger\")\n\n# OUTPUT\n\u276f Gracy tracked that 'https://pokeapi.co/api/v2/pokemon/{NAME}' was hit 1 time(s) with a success rate of 100.00%, avg latency of 0.45s, and a rate of 1.0 reqs/s.\n\u276f Gracy tracked a total of 2 requests with a success rate of 100.00%, avg latency of 0.24s, and a rate of 1.0 reqs/s.\n```\n\n### List\n\nUses `print` to generate a short list with all attributes:\n\n```python\npokeapi = GracefulPokeAPI()\n# do stuff with your API\npokeapi.report_status(\"list\")\n\n# OUTPUT\n ____\n / ___|_ __ __ _ ___ _ _\n | | _| '__/ _` |/ __| | | |\n | |_| | | | (_| | (__| |_| |\n \\____|_| \\__,_|\\___|\\__, |\n |___/ Requests Summary Report\n\n\n1. https://pokeapi.co/api/v2/pokemon/{NAME}\n Total Reqs (#): 1\n Success (%): 100.00%\n Fail (%): 0.00%\n Avg Latency (s): 0.39\n Max Latency (s): 0.39\n 2xx Resps: 1\n 3xx Resps: 0\n 4xx Resps: 0\n 5xx Resps: 0\n Avg Reqs/sec: 1.0 reqs/s\n\n\n2. https://pokeapi.co/api/v2/generation/{ID}/\n Total Reqs (#): 1\n Success (%): 100.00%\n Fail (%): 0.00%\n Avg Latency (s): 0.04\n Max Latency (s): 0.04\n 2xx Resps: 1\n 3xx Resps: 0\n 4xx Resps: 0\n 5xx Resps: 0\n Avg Reqs/sec: 1.0 reqs/s\n\n\nTOTAL\n Total Reqs (#): 2\n Success (%): 100.00%\n Fail (%): 0.00%\n Avg Latency (s): 0.21\n Max Latency (s): 0.00\n 2xx Resps: 2\n 3xx Resps: 0\n 4xx Resps: 0\n 5xx Resps: 0\n Avg Reqs/sec: 1.0 reqs/s\n```\n\n### Table\n\nIt requires you to install [Rich](https://github.com/Textualize/rich).\n\n```py\npokeapi = GracefulPokeAPI()\n# do stuff with your API\npokeapi.report_status(\"rich\")\n```\n\nHere's an example of how it looks:\n\n\n\n\n### Plotly\n\nIt requires you to install [plotly \ud83d\udcca](https://github.com/plotly/plotly.py) and [pandas \ud83d\udc3c](https://github.com/pandas-dev/pandas).\n\n```py\npokeapi = GracefulPokeAPI()\n# do stuff with your API\nplotly_fig = pokeapi.report_status(\"plotly\")\nplotly_fig.show()\n```\n\nHere's an example of how it looks:\n\n\n\n## Replay requests\n\nGracy allows you to replay requests and responses from previous interactions.\n\nThis is powerful because it allows you to test APIs without latency or consuming your rate limit. Now writing unit tests that relies on third-party APIs is doable.\n\nIt works in two steps:\n\n| **Step** | **Description** | **Hits the API?** |\n| ------------ | ------------------------------------------------------------------------------ | ----------------- |\n| 1. Recording | Stores all requests/responses to be later replayed | **Yes** |\n| 2. Replay | Returns all previously generated responses based on your request as a \"replay\" | No |\n\n### Recording\n\nThe effort to record requests/responses is ZERO. You just need to pass a recording config to your Graceful API:\n\n```py\nfrom gracy import GracyReplay\nfrom gracy.replays.storages.sqlite import SQLiteReplayStorage\n\nrecord_mode = GracyReplay(\"record\", SQLiteReplayStorage(\"pokeapi.sqlite3\"))\npokeapi = GracefulPokeAPI(record_mode)\n```\n\n**Every request** will be recorded to the defined data source.\n\n### Replay\n\nOnce you have recorded all your requests you can enable the replay mode:\n\n```py\nfrom gracy import GracyReplay\nfrom gracy.replays.storages.sqlite import SQLiteReplayStorage\n\nreplay_mode = GracyReplay(\"replay\", SQLiteReplayStorage(\"pokeapi.sqlite3\"))\npokeapi = GracefulPokeAPI(replay_mode)\n```\n\n**Every request** will be routed to the defined data source resulting in faster responses.\n\n**\u26a0\ufe0f Note that parsers, retries, throttling, and similar configs will work as usual**.\n\n\n## Resource Namespacing\n\nYou can have multiple namespaces to organize your API endpoints as you wish.\n\nTo do so, you just have to inherit from `GracyNamespace` and instantiate it within the `GracyAPI`:\n\n```py\nfrom gracy import Gracy, GracyNamespace, GracyConfig\n\nclass PokemonNamespace(GracyNamespace[PokeApiEndpoint]):\n async def get_one(self, name: str):\n return await self.get(PokeApiEndpoint.GET_POKEMON, {\"NAME\": name})\n\n\nclass BerryNamespace(GracyNamespace[PokeApiEndpoint]):\n async def get_one(self, name: str):\n return await self.get(PokeApiEndpoint.GET_BERRY, {\"NAME\": name})\n\n\nclass GracefulPokeAPI(Gracy[PokeApiEndpoint]):\n class Config:\n BASE_URL = \"https://pokeapi.co/api/v2/\"\n SETTINGS = GracyConfig(\n retry=RETRY,\n allowed_status_code={HTTPStatus.NOT_FOUND},\n parser={HTTPStatus.NOT_FOUND: None},\n )\n\n # These will be automatically assigned on init\n berry: BerryNamespace\n pokemon: PokemonNamespace\n```\n\nAnd the usage will work as:\n\n```py\nawait pokeapi.pokemon.get_one(\"pikachu\")\nawait pokeapi.berry.get_one(\"cheri\")\n```\n\nNote all configs are propagated to namespaces, but namespaces can still have their own which would cause merges when instantiatedg.\n\n\n## Pagination\n\nThere're endpoints that may require pagination. For that you can use `GracyPaginator`.\n\nFor a simple case where you pass `offset` and `limit`, you can use `GracyOffsetPaginator`:\n\n```py\nfrom gracy import GracyOffsetPaginator\n\nclass BerryNamespace(GracyNamespace[PokeApiEndpoint]):\n @parsed_response(ResourceList)\n async def list(self, offset: int = 0, limit: int = 20):\n params = dict(offset=offset, limit=limit)\n return await self.get(PokeApiEndpoint.BERRY_LIST, params=params)\n\n def paginate(self, limit: int = 20) -> GracyOffsetPaginator[ResourceList]:\n return GracyOffsetPaginator[ResourceList](\n gracy_func=self.list,\n has_next=lambda r: bool(r[\"next\"]) if r else True,\n page_size=limit,\n )\n\n```\n\nand then use it as:\n\n```py\nasync def main():\n api = PokeApi()\n paginator = api.berry.paginate(2)\n\n # Just grabs the next page\n first = await paginator.next_page()\n print(first)\n\n # Resets current page to 0\n paginator.set_page(0)\n\n # Loop throught it all\n async for page in paginator:\n print(page)\n```\n\n## Advanced Usage\n\n### Customizing/Overriding configs per method\n\nAPIs may return different responses/conditions/payloads based on the endpoint.\n\nYou can override any `GracyConfig` on a per method basis by using the `@graceful` decorator.\n\nNOTE: Use `@graceful_generator` if your function uses `yield`.\n\n```python\nfrom gracy import Gracy, GracyConfig, GracefulRetry, graceful, graceful_generator\n\nretry = GracefulRetry(...)\n\nclass GracefulPokeAPI(Gracy[PokeApiEndpoint]):\n class Config:\n BASE_URL = \"https://pokeapi.co/api/v2/\"\n SETTINGS = GracyConfig(\n retry=retry,\n log_errors=LogEvent(\n LogLevel.ERROR, \"How can I become a pokemon master if {URL} keeps failing with {STATUS}\"\n ),\n )\n\n @graceful(\n retry=None, # \ud83d\udc48 Disables retry set in Config\n log_errors=None, # \ud83d\udc48 Disables log_errors set in Config\n allowed_status_code=HTTPStatus.NOT_FOUND,\n parser={\n \"default\": lambda r: r.json()[\"order\"],\n HTTPStatus.NOT_FOUND: None,\n },\n )\n async def maybe_get_pokemon_order(self, name: str):\n val: str | None = await self.get(PokeApiEndpoint.GET_POKEMON, {\"NAME\": name})\n return val\n\n @graceful( # \ud83d\udc48 Retry and log_errors are still set for this one\n strict_status_code=HTTPStatus.OK,\n parser={\"default\": lambda r: r.json()[\"order\"]},\n )\n async def get_pokemon_order(self, name: str):\n val: str = await self.get(PokeApiEndpoint.GET_POKEMON, {\"NAME\": name})\n return val\n\n @graceful_generator( # \ud83d\udc48 Retry and log_errors are still set for this one\n parser={\"default\": lambda r: r.json()[\"order\"]},\n )\n async def get_2_pokemons(self):\n names = [\"charmander\", \"pikachu\"]\n\n for name in names:\n r = await self.get(PokeApiEndpoint.GET_POKEMON, {\"NAME\": name})\n yield r\n```\n\n### Customizing HTTPx client\n\nYou might want to modify the HTTPx client settings, do so by:\n\n```py\nclass YourAPIClient(Gracy[str]):\n class Config:\n ...\n\n def __init__(self, token: token) -> None:\n self._token = token\n super().__init__()\n\n # \ud83d\udc47 Implement your logic here\n def _create_client(self) -> httpx.AsyncClient:\n client = super()._create_client()\n client.headers = {\"Authorization\": f\"token {self._token}\"} # type: ignore\n return client\n```\n\n### Overriding default request timeout\n\nAs default Gracy won't enforce a request timeout.\n\nYou can define your own by setting it on Config as:\n\n```py\nclass GracefulAPI(GracyApi[str]):\n class Config:\n BASE_URL = \"https://example.com\"\n REQUEST_TIMEOUT = 10.2 # \ud83d\udc48 Here\n```\n\n### Creating a custom Replay data source\n\nGracy was built with extensibility in mind.\n\nYou can create your own storage to store/load anywhere (e.g. SQL Database), here's an example:\n\n```py\nimport httpx\nfrom gracy import GracyReplayStorage\n\nclass MyCustomStorage(GracyReplayStorage):\n def prepare(self) -> None: # (Optional) Executed upon API instance creation.\n ...\n\n async def record(self, response: httpx.Response) -> None:\n ... # REQUIRED. Your logic to store the response object. Note the httpx.Response has request data.\n\n async def _load(self, request: httpx.Request) -> httpx.Response:\n ... # REQUIRED. Your logic to load a response object based on the request.\n\n\n# Usage\nrecord_mode = GracyReplay(\"record\", MyCustomStorage())\nreplay_mode = GracyReplay(\"replay\", MyCustomStorage())\n\npokeapi = GracefulPokeAPI(record_mode)\n```\n\n### Hooks before/after request\n\nYou can set up hooks simply by defining `async def before` and `async def after` methods.\n\n\u26a0\ufe0f NOTE: Gracy configs are disabled within these methods which means that retries/parsers/throttling won't take effect inside it.\n\n```py\nclass GracefulPokeAPI(Gracy[PokeApiEndpoint]):\n class Config:\n BASE_URL = \"https://pokeapi.co/api/v2/\"\n SETTINGS = GracyConfig(\n retry=RETRY,\n allowed_status_code={HTTPStatus.NOT_FOUND},\n parser={HTTPStatus.NOT_FOUND: None},\n )\n\n def __init__(self, *args: t.Any, **kwargs: t.Any) -> None:\n self.before_count = 0\n\n self.after_status_counter = defaultdict[HTTPStatus, int](int)\n self.after_aborts = 0\n self.after_retries_counter = 0\n\n super().__init__(*args, **kwargs)\n\n async def before(self, context: GracyRequestContext):\n self.before_count += 1\n\n async def after(\n self,\n context: GracyRequestContext, # Current request context\n response_or_exc: httpx.Response | Exception, # Either the request or an error\n retry_state: GracefulRetryState | None, # Set when this is generated from a retry\n ):\n if retry_state:\n self.after_retries_counter += 1\n\n if isinstance(response_or_exc, httpx.Response):\n self.after_status_counter[HTTPStatus(response_or_exc.status_code)] += 1\n else:\n self.after_aborts += 1\n\n async def get_pokemon(self, name: str):\n return await self.get(PokeApiEndpoint.GET_POKEMON, {\"NAME\": name})\n```\n\nIn the example above invoking `get_pokemon()` will trigger `before()`/`after()` hooks in sequence.\n\n#### Common Hooks\n\n##### `HttpHeaderRetryAfterBackOffHook`\n\nThis hook checks for 429 (TOO MANY REQUESTS), and then reads the\n[`retry-after` header](https://developer.mozilla.org/en-US/docs/Web/HTTP/Headers/Retry-After).\n\nIf the value is set, then Gracy pauses **ALL** client requests until the time is over. This behavior can be modified to happen on a per-endpoint basis if `lock_per_endpoint` is True.\n\nExample Usage:\n\n```py\nfrom gracy.common_hooks import HttpHeaderRetryAfterBackOffHook\n\nclass GracefulAPI(GracyAPI[Endpoint]):\n def __init__(self):\n self._retry_after_hook = HttpHeaderRetryAfterBackOffHook(\n self._reporter,\n lock_per_endpoint=True,\n log_event=LogEvent(\n LogLevel.WARNING,\n custom_message=(\n \"{ENDPOINT} produced {STATUS} and requested to wait {RETRY_AFTER}s \"\n \"- waiting {RETRY_AFTER_ACTUAL_WAIT}s\"\n ),\n ),\n # Wait +10s to avoid this from happening again too soon\n seconds_processor=lambda secs_requested: secs_requested + 10,\n )\n\n super().__init__()\n\n async def before(self, context: GracyRequestContext):\n await self._retry_after_hook.before(context)\n\n async def after(\n self,\n context: GracyRequestContext,\n response_or_exc: httpx.Response | Exception,\n retry_state: GracefulRetryState | None,\n ):\n retry_after_result = await self._retry_after_hook.after(context, response_or_exc)\n```\n\n##### `RateLimitBackOffHook`\n\nThis hook checks for 429 (TOO MANY REQUESTS) and locks requests for an arbitrary amount of time defined by you.\n\nIf the value is set, then Gracy pauses **ALL** client requests until the time is over.\nThis behavior can be modified to happen on a per-endpoint basis if `lock_per_endpoint` is True.\n\n\n```py\nfrom gracy.common_hooks import RateLimitBackOffHook\n\nclass GracefulAPI(GracyAPI[Endpoint]):\n def __init__(self):\n self._ratelimit_backoff_hook = RateLimitBackOffHook(\n 30,\n self._reporter,\n lock_per_endpoint=True,\n log_event=LogEvent(\n LogLevel.INFO,\n custom_message=\"{UENDPOINT} got rate limited, waiting for {WAIT_TIME}s\",\n ),\n )\n\n super().__init__()\n\n async def before(self, context: GracyRequestContext):\n await self._ratelimit_backoff_hook.before(context)\n\n async def after(\n self,\n context: GracyRequestContext,\n response_or_exc: httpx.Response | Exception,\n retry_state: GracefulRetryState | None,\n ):\n backoff_result = await self._ratelimit_backoff_hook.after(context, response_or_exc)\n```\n\n\n```py\nfrom gracy.common_hooks import HttpHeaderRetryAfterBackOffHook, RateLimitBackOffHook\n```\n\n## \ud83d\udcda Extra Resources\n\nSome good practices I learned over the past years guided Gracy's philosophy, you might benefit by reading:\n\n- [How to log](https://guicommits.com/how-to-log-in-python-like-a-pro/)\n- [How to handle exceptions](https://guicommits.com/handling-exceptions-in-python-like-a-pro/)\n - [How to structure exceptions](https://guicommits.com/how-to-structure-exception-in-python-like-a-pro/)\n- [How to use Async correctly](https://guicommits.com/effective-python-async-like-a-pro/)\n- [Book: Python like a PRO](https://guilatrova.gumroad.com/l/python-like-a-pro)\n- [Book: Effective Python](https://amzn.to/3bEVHpG)\n\n<!-- ## Contributing -->\n<!-- Thank you for considering making Gracy better for everyone! -->\n<!-- Refer to [Contributing docs](docs/CONTRIBUTING.md).-->\n\n## Change log\n\nSee [CHANGELOG](CHANGELOG.md).\n\n## License\n\nMIT\n\n## Credits\n\nThanks to the last three startups I worked which forced me to do the same things and resolve the same problems over and over again. I got sick of it and built this lib.\n\nMost importantly: **Thanks to God**, who allowed me (a random \ud83c\udde7\ud83c\uddf7 guy) to work for many different \ud83c\uddfa\ud83c\uddf8 startups. This is ironic since due to God's grace, I was able to build Gracy. \ud83d\ude4c\n\nAlso, thanks to the [httpx](https://github.com/encode/httpx) and [rich](https://github.com/Textualize/rich) projects for the beautiful and simple APIs that powers Gracy.\n\n",
"bugtrack_url": null,
"license": "MIT",
"summary": "Gracefully manage your API interactions",
"version": "1.33.1",
"project_urls": {
"Changelog": "https://github.com/guilatrova/gracy/blob/main/CHANGELOG.md",
"Homepage": "https://github.com/guilatrova/gracy",
"Repository": "https://github.com/guilatrova/gracy"
},
"split_keywords": [
"api",
" throttling",
" http",
" https",
" async",
" retry"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "b850fd0afc975579550ee06fd281a9f7645f35edbd6b74607dfa404fc5dc3b7e",
"md5": "4740b7f98cd397f8fe6a4e2b023988d8",
"sha256": "983feb06f9300fd1454493c6180baad53beef2fdf0f033ba2d6b6e3b1f563720"
},
"downloads": -1,
"filename": "gracy-1.33.1-py3-none-any.whl",
"has_sig": false,
"md5_digest": "4740b7f98cd397f8fe6a4e2b023988d8",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": "<4.0,>=3.8.1",
"size": 46167,
"upload_time": "2024-10-07T20:30:53",
"upload_time_iso_8601": "2024-10-07T20:30:53.004359Z",
"url": "https://files.pythonhosted.org/packages/b8/50/fd0afc975579550ee06fd281a9f7645f35edbd6b74607dfa404fc5dc3b7e/gracy-1.33.1-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "34b95f48bbeaf8ae9a8ae40c6cce5d52291eebd7fd6b698684cfb5e7a46f210d",
"md5": "7a7d97c26fee2643075675881b4042ec",
"sha256": "c0d4e469b396cc5eb6928cb8fd2147d5cae39eaf9e8adfd2717fb2e9672eda7e"
},
"downloads": -1,
"filename": "gracy-1.33.1.tar.gz",
"has_sig": false,
"md5_digest": "7a7d97c26fee2643075675881b4042ec",
"packagetype": "sdist",
"python_version": "source",
"requires_python": "<4.0,>=3.8.1",
"size": 48423,
"upload_time": "2024-10-07T20:30:54",
"upload_time_iso_8601": "2024-10-07T20:30:54.623214Z",
"url": "https://files.pythonhosted.org/packages/34/b9/5f48bbeaf8ae9a8ae40c6cce5d52291eebd7fd6b698684cfb5e7a46f210d/gracy-1.33.1.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-10-07 20:30:54",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "guilatrova",
"github_project": "gracy",
"travis_ci": false,
"coveralls": false,
"github_actions": true,
"lcname": "gracy"
}
JSON
