WIP on README

This commit is contained in:
Andrew Brookins 2021-10-22 08:31:08 -07:00
parent 09c91fb756
commit 7c0bea751b
7 changed files with 319 additions and 49 deletions

156
README.md
View file

@ -44,9 +44,118 @@ This *preview release* includes our first major component: a **declarative model
### Object Mapping
With Redis OM, you get powerful data modeling, validation, and query expressions with a small amount of code.
With Redis OM, you get powerful data modeling, extensible data validation with [Pydantic](pydantic-url), and rich query expressions with a small amount of code.
Check out this example of data modeling and validation. First, we're going to create a `Customer` model that we can use to save data to Redis.
```python
import datetime
from typing import Optional
from pydantic import EmailStr
from redis_om.model import (
HashModel,
)
class Customer(HashModel):
first_name: str
last_name: str
email: EmailStr
join_date: datetime.date
age: int
bio: Optional[str]
```
Here, we've defined a `Customer` model with the `HashModel` class from redis-om. This model will save data in Redis as a [Redis Hash](https://redis.io/topics/data-types).
Next, let's see how Redis OM makes it easy to save and retrieve `Customer` data in Redis.
```python
# We can create a new Customer object:
andrew = Customer(
first_name="Andrew",
last_name="Brookins",
email="andrew.brookins@example.com",
join_date=datetime.date.today(),
age=38,
bio="Python developer, works at Redis, Inc."
)
# The model generates a globally unique primary key automatically without
# needing to talk to Redis.
print(andrew.pk)
# '01FJM6PH661HCNNRC884H6K30C'
# We can save the model to Redis.
andrew.save()
# Now, we can retrieve this customer with its primary key:
other_andrew = Customer.get('01FJM6PH661HCNNRC884H6K30C')
# The original model and this one pass an equality check.
assert other_andrew == andrew
```
Now, let's talk about **validation**. Did you notice the type annotation for the `email` field was `EmailStr`?
`EmailStr` is a [Pydantic field validator](https://pydantic-docs.helpmanual.io/usage/types/). Because every Redis OM model is also a Pydantic model, you can use Pydantic validators like `EmailStr`, `Pattern`, and many more!
Let's see what happens if we try to instantiate our `Customer` class with an invalid email address.
```python
# We'll get a validation error if we try to use an invalid email address!
Customer(
first_name="Andrew",
last_name="Brookins",
email="Not an email address!",
join_date=datetime.date.today(),
age=38,
bio="Python developer, works at Redis, Inc."
)
# Traceback:
# pydantic.error_wrappers.ValidationError: 1 validation error for Customer
# email
# value is not a valid email address (type=value_error.email)
# We'll also get a validation error if we try to save a model
# instance with an invalid email.
andrew = Customer(
first_name="Andrew",
last_name="Brookins",
email="andrew.brookins@example.com",
join_date=datetime.date.today(),
age=38,
bio="Python developer, works at Redis, Inc."
)
# Sometime later...
andrew.email = "Not valid"
andrew.save()
# Traceback:
# pydantic.error_wrappers.ValidationError: 1 validation error for Customer
# email
# value is not a valid email address (type=value_error.email)
```
Data modeling, validation, and persistent to Redis all work regardless of where you run Redis. But can we do more?
Yes, we can! Next, we'll talk about the **rich query expressions** and **embedded models** that Redis OM gives you when you're using the RediSearch and RedisJSON Redis modules.
### Querying
Querying uses a rich expression syntax inspired by the Django ORM, SQLAlchemy, and Peewee.
The example code defines `Address` and `Customer` models for use with a Redis database with the [RedisJSON](redis-json-url) module installed.
With these two classes defined, you can now:
* Validate data based on the model's type annotations using Pydantic
* Persist model instances to Redis as JSON
* Instantiate model instances from Redis by primary key (a client-generated [ULID](ulid-url))
* Query on any indexed fields in the models
Check out this example:
```python
import datetime
@ -80,18 +189,6 @@ class Customer(JsonModel):
address: Address
```
The example code defines `Address` and `Customer` models for use with a Redis database with the [RedisJSON](redis-json-url) module installed.
With these two classes defined, you can now:
* Validate data based on the model's type annotations using [Pydantic](pydantic-url)
* Persist model instances to Redis as JSON
* Instantiate model instances from Redis by primary key (a client-generated [ULID](ulid-url))
* Query on any indexed fields in the models
### Querying
Querying uses a rich expression syntax inspired by the Django ORM, SQLAlchemy, and Peewee.
Here are a few example queries that use the models we defined earlier:
```python
@ -139,34 +236,9 @@ hit us up on the [Redis Discord Server](http://discord.gg/redis).
## ✨ RediSearch and RedisJSON
Redis OM relies on core features from two source available Redis modules: **RediSearch** and **RedisJSON**.
Some advanced features of Redis OM rely on core features from two source available Redis modules: **RediSearch** and **RedisJSON**.
These modules are the "magic" behind the scenes:
* RediSearch adds querying, indexing, and full-text search to Redis
* RedisJSON adds the JSON data type to Redis
### Why this is important
Without RediSearch or RedisJSON installed, you can still use Redis OM to create declarative models backed by Redis.
We'll store your model data in Redis as Hashes, and you can retrieve models using their primary keys. You'll also get all the validation features from Pydantic.
So, what won't work without these modules?
1. Without RedisJSON, you won't be able to nest models inside each other, like we did with the example model of a `Customer` model that has an `Address` embedded inside it.
2. Without RediSearch, you won't be able to use our expressive queries to find models -- just primary keys.
### So how do you get RediSearch and RedisJSON?
You can use RediSearch and RedisJSON with your self-hosted Redis deployment. Just follow the instructions on installing the binary versions of the modules in their Quick Start Guides:
- [RedisJSON Quick Start - Running Binaries](https://oss.redis.com/redisjson/#download-and-running-binaries)
- [RediSearch Quick Start - Running Binaries](https://oss.redis.com/redisearch/Quick_Start/#download_and_running_binaries)
**NOTE**: Both Quick Start Guides also have instructions on how to run these modules in Redis with Docker.
Don't want to run Redis yourself? RediSearch and RedisJSON are also available on Redis Cloud. [Get started here.](https://redis.com/try-free/)
To learn more, read [our documentation](docs/redis_modules.md).
## ❤️ Contributing
@ -178,7 +250,7 @@ You can also **contribute documentation** -- or just let us know if something ne
## License
Redis OM is [MIT licensed][license-url].
Redis OM uses the [BSD 3-Clause license][license-url].
<!-- Badges -->

72
demo.py Normal file
View file

@ -0,0 +1,72 @@
import datetime
from typing import Optional
from pydantic import EmailStr
from redis_om.model import (
HashModel
)
class Customer(HashModel):
first_name: str
last_name: str
email: EmailStr
join_date: datetime.date
age: int
bio: Optional[str]
# Now we can create new Customer objects:
andrew = Customer(
first_name="Andrew",
last_name="Brookins",
email="andrew.brookins@example.com",
join_date=datetime.date.today(),
age=38,
bio="Python developer, works at Redis, Inc."
)
# The model generates a globally unique primary key automatically.
print(andrew.pk)
# '01FJM6PH661HCNNRC884H6K30C'
# You can save the model to Redis.
andrew.save()
# Later, you can retrieve this customer with its primary key:
other_andrew = Customer.get('01FJM6PH661HCNNRC884H6K30C')
# The original model and this one pass an equality check.
assert other_andrew == andrew
# We'll get a validation error if we try to use an invalid email address!
Customer(
first_name="Andrew",
last_name="Brookins",
email="Not an email address!",
join_date=datetime.date.today(),
age=38,
bio="Python developer, works at Redis, Inc."
)
# We'll also get a validation error if we try to save a model
# instance with an invalid email.
andrew = Customer(
first_name="Andrew",
last_name="Brookins",
email="andrew.brookins@example.com",
join_date=datetime.date.today(),
age=38,
bio="Python developer, works at Redis, Inc."
)
# Sometime later...
andrew.email = "Not valid"
andrew.save()
# Traceback:
# pydantic.error_wrappers.ValidationError: 1 validation error for Customer
# email
# value is not a valid email address (type=value_error.email)

30
docs/redis_modules.md Normal file
View file

@ -0,0 +1,30 @@
# Redis Modules
Some advanced features of Redis OM, like rich query expressions and saving data as JSON, rely on core features from two source available Redis modules: **RediSearch** and **RedisJSON**.
These modules are the "magic" behind the scenes:
* RediSearch adds querying, indexing, and full-text search to Redis
* RedisJSON adds the JSON data type to Redis
## Why this is important
Without RediSearch or RedisJSON installed, you can still use Redis OM to create declarative models backed by Redis.
We'll store your model data in Redis as Hashes, and you can retrieve models using their primary keys. You'll also get all the validation features from Pydantic.
So, what won't work without these modules?
1. Without RedisJSON, you won't be able to nest models inside each other, like we did with the example model of a `Customer` model that has an `Address` embedded inside it.
2. Without RediSearch, you won't be able to use our expressive queries to find models -- just primary keys.
## So how do you get RediSearch and RedisJSON?
You can use RediSearch and RedisJSON with your self-hosted Redis deployment. Just follow the instructions on installing the binary versions of the modules in their Quick Start Guides:
- [RedisJSON Quick Start - Running Binaries](https://oss.redis.com/redisjson/#download-and-running-binaries)
- [RediSearch Quick Start - Running Binaries](https://oss.redis.com/redisearch/Quick_Start/#download_and_running_binaries)
**NOTE**: Both Quick Start Guides also have instructions on how to run these modules in Redis with Docker.
Don't want to run Redis yourself? RediSearch and RedisJSON are also available on Redis Cloud. [Get started here.](https://redis.com/try-free/)

37
poetry.lock generated
View file

@ -217,6 +217,21 @@ category = "dev"
optional = false
python-versions = ">=3.5"
[[package]]
name = "dnspython"
version = "2.1.0"
description = "DNS toolkit"
category = "dev"
optional = false
python-versions = ">=3.6"
[package.extras]
dnssec = ["cryptography (>=2.6)"]
doh = ["requests", "requests-toolbelt"]
idna = ["idna (>=2.1)"]
curio = ["curio (>=1.2)", "sniffio (>=1.1)"]
trio = ["trio (>=0.14.0)", "sniffio (>=1.1)"]
[[package]]
name = "docutils"
version = "0.17.1"
@ -225,6 +240,18 @@ category = "dev"
optional = false
python-versions = ">=2.7, !=3.0.*, !=3.1.*, !=3.2.*, !=3.3.*, !=3.4.*"
[[package]]
name = "email-validator"
version = "1.1.3"
description = "A robust email syntax and deliverability validation library for Python 2.x/3.x."
category = "dev"
optional = false
python-versions = "!=3.0.*,!=3.1.*,!=3.2.*,!=3.3.*,!=3.4.*,>=2.7"
[package.dependencies]
dnspython = ">=1.15.0"
idna = ">=2.0.0"
[[package]]
name = "execnet"
version = "1.9.0"
@ -1013,7 +1040,7 @@ testing = ["pytest (>=4.6)", "pytest-checkdocs (>=2.4)", "pytest-flake8", "pytes
[metadata]
lock-version = "1.1"
python-versions = "^3.8"
content-hash = "6b31dc9e814263cba72d68f76b04bc3af635c685c6281a29e962d53cb05b2d0f"
content-hash = "de30b2382aaeb2fe0675658bce5a3e5bc21a14e85c66d94a90054bc73a7831cd"
[metadata.files]
aioredis = [
@ -1185,10 +1212,18 @@ decorator = [
{file = "decorator-5.1.0-py3-none-any.whl", hash = "sha256:7b12e7c3c6ab203a29e157335e9122cb03de9ab7264b137594103fd4a683b374"},
{file = "decorator-5.1.0.tar.gz", hash = "sha256:e59913af105b9860aa2c8d3272d9de5a56a4e608db9a2f167a8480b323d529a7"},
]
dnspython = [
{file = "dnspython-2.1.0-py3-none-any.whl", hash = "sha256:95d12f6ef0317118d2a1a6fc49aac65ffec7eb8087474158f42f26a639135216"},
{file = "dnspython-2.1.0.zip", hash = "sha256:e4a87f0b573201a0f3727fa18a516b055fd1107e0e5477cded4a2de497df1dd4"},
]
docutils = [
{file = "docutils-0.17.1-py2.py3-none-any.whl", hash = "sha256:cf316c8370a737a022b72b56874f6602acf974a37a9fba42ec2876387549fc61"},
{file = "docutils-0.17.1.tar.gz", hash = "sha256:686577d2e4c32380bb50cbb22f575ed742d58168cee37e99117a854bcd88f125"},
]
email-validator = [
{file = "email_validator-1.1.3-py2.py3-none-any.whl", hash = "sha256:5675c8ceb7106a37e40e2698a57c056756bf3f272cfa8682a4f87ebd95d8440b"},
{file = "email_validator-1.1.3.tar.gz", hash = "sha256:aa237a65f6f4da067119b7df3f13e89c25c051327b2b5b66dc075f33d62480d7"},
]
execnet = [
{file = "execnet-1.9.0-py2.py3-none-any.whl", hash = "sha256:a295f7cc774947aac58dde7fdc85f4aa00c42adf5d8f5468fc630c1acf30a142"},
{file = "execnet-1.9.0.tar.gz", hash = "sha256:8f694f3ba9cc92cab508b152dcfe322153975c29bda272e2fd7f3f00f36e47c5"},

View file

@ -32,6 +32,7 @@ coverage = "^6.0.2"
pytest-cov = "^3.0.0"
pytest-xdist = "^2.4.0"
twine = "^3.4.2"
email-validator = "^1.1.3"
[tool.poetry.scripts]
migrate = "redis_om.model.cli.migrate:migrate"

View file

@ -31,7 +31,7 @@ import redis
from pydantic import BaseModel, validator
from pydantic.fields import FieldInfo as PydanticFieldInfo
from pydantic.fields import ModelField, Undefined, UndefinedType
from pydantic.main import ModelMetaclass
from pydantic.main import ModelMetaclass, validate_model
from pydantic.typing import NoArgAnyCallable
from pydantic.utils import Representation
from redis.client import Pipeline
@ -1162,15 +1162,16 @@ class RedisModel(BaseModel, abc.ABC, metaclass=ModelMeta):
# TODO: Add transaction support
return [model.save() for model in models]
@classmethod
def values(cls):
"""Return raw values from Redis instead of model instances."""
raise NotImplementedError
@classmethod
def redisearch_schema(cls):
raise NotImplementedError
def check(self):
"""Run all validations."""
*_, validation_error = validate_model(self.__class__, self.__dict__)
if validation_error:
raise validation_error
class HashModel(RedisModel, abc.ABC):
def __init_subclass__(cls, **kwargs):
@ -1190,6 +1191,7 @@ class HashModel(RedisModel, abc.ABC):
)
def save(self, pipeline: Optional[Pipeline] = None) -> "HashModel":
self.check()
if pipeline is None:
db = self.db()
else:
@ -1226,6 +1228,12 @@ class HashModel(RedisModel, abc.ABC):
schema_parts = [schema_prefix] + cls.schema_for_fields()
return " ".join(schema_parts)
def update(self, **field_values):
validate_model_fields(self.__class__, field_values)
for field, value in field_values.items():
setattr(self, field, value)
self.save()
@classmethod
def schema_for_fields(cls):
schema_parts = []
@ -1312,6 +1320,7 @@ class JsonModel(RedisModel, abc.ABC):
cls.redisearch_schema()
def save(self, pipeline: Optional[Pipeline] = None) -> "JsonModel":
self.check()
if pipeline is None:
db = self.db()
else:
@ -1320,6 +1329,7 @@ class JsonModel(RedisModel, abc.ABC):
return self
def update(self, **field_values):
# TODO: Better support for embedded field models.
validate_model_fields(self.__class__, field_values)
for field, value in field_values.items():
setattr(self, field, value)

View file

@ -0,0 +1,50 @@
import abc
import datetime
from collections import namedtuple
import pytest
from pydantic import EmailStr, ValidationError
from redis_om.model import HashModel, Field
from redis_om.model.migrations.migrator import Migrator
today = datetime.date.today()
@pytest.fixture
def m(key_prefix):
class BaseHashModel(HashModel, abc.ABC):
class Meta:
global_key_prefix = key_prefix
class Member(BaseHashModel):
first_name: str
last_name: str
email: EmailStr = Field(index=True)
join_date: datetime.date
age: int
Migrator().run()
return namedtuple("Models", ["Member"])(Member)
def test_email_str(m):
with pytest.raises(ValidationError):
m.Member(
first_name="Andrew",
last_name="Brookins",
email="not an email!",
age=38,
join_date=today,
)
with pytest.raises(ValidationError):
m.Member(
first_name="Andrew",
last_name="Brookins",
email="andrew@bad-domain",
age=38,
join_date=today,
)