drf-rust 0.1.1

RustModelSerializer: DRF ModelSerializer read accelerated by Rust (PyO3)

drf-rust

37× Faster Serialization for Django REST Framework

Drop-in Rust acceleration for Django REST Framework's ModelSerializer serialization process.

Just replace:

ModelSerializer → RustModelSerializer

No other changes required.

Works with:

• serializers
• views
• pagination
• filtering
• URL configs

Everything continues to work exactly the same.

---

Benchmark

Tested on a production PostgreSQL database with a 3-level nested serializer.

Rows	DRF	drf-rust	Speedup
100	0.042s	0.019s	2.2x
400	0.888s	0.035s	25.4x
800	1.775s	0.056s	31.7x
1,200	2.541s	0.079s	32.2x
2,400	5.090s	0.144s	35.3x
4,800	10.356s	0.278s	37.3x
9,600	20.442s	0.558s	36.6x
19,200	39.753s	1.208s	32.9x
38,400	79.135s	4.145s	19.1x

Key takeaway

• Up to 37x faster

---

Example

Before

from rest_framework import serializers

class ArticleSerializer(serializers.ModelSerializer):
    author_name = serializers.CharField(source="author.name")

    class Meta:
        model = Article
        fields = ["id", "title", "author_name"]

After

from drf_rust.serializers import RustModelSerializer

class ArticleSerializer(RustModelSerializer):
    author_name = serializers.CharField(source="author.name")

    class Meta:
        model = Article
        fields = ["id", "title", "author_name"]

That's it.

---

Why DRF Serialization Is Slow

DRF serialization suffers from several performance issues:

1. Python per-row processing

DRF serializes each object individually.

for obj in queryset:
    serializer.to_representation(obj)

For large querysets this becomes extremely expensive.

---

2. Nested serializer N+1 queries

Nested serializers often trigger extra database queries.

Example:

Order
 └── user
      └── organization
           └── name

DRF may execute many queries instead of one JOIN.

---

3. Python attribute resolution overhead

Every field requires:

• attribute lookup
• field conversion
• serializer processing

This overhead accumulates quickly.

---

How drf-rust Works

drf-rust moves serialization into compiled Rust code.

Pipeline:

Django View
     │
     ▼
RustModelSerializer
     │
     ▼
Field classification
     │
     ▼
SQL generation
     │
     ▼
Rust execution engine
     │
     ▼
Database
     │
     ▼
Python list[dict]

Key improvements:

• SQL JOIN flattening
• zero N+1 queries
• compiled row processing
• GIL released during I/O

---

Features

Accelerated fields

• CharField
• IntegerField
• DecimalField
• BooleanField
• DateTimeField
• ...

Supported relationships

• ForeignKey
• OneToOne
• ManyToMany
• deep join chains

Example:

source="author.department.company.name"

Nested serializers

Nested serializers are automatically optimized into SQL JOINs.

---

Python-only fields

Some fields are handled by DRF normally:

• SerializerMethodField
• source='*'
• callable sources

These are filled after Rust returns results.

---

Automatic Fallback

If Rust fails for any reason:

drf-rust → fallback to native DRF

Your API will still work.

---

Installation

Install from source

git clone https://github.com/john-tanner-dev/drf-rust
cd drf-rust

pip install maturin
maturin develop --release

---

Requirements

• Python ≥ 3.8
• Django ≥ 3.0
• djangorestframework ≥ 3.12
• Rust toolchain
• maturin ≥ 1.0

---

Supported Databases

Database	Status
PostgreSQL	✓
MySQL	✓
SQLite	✓

---

Roadmap

Planned improvements:

• ManyToMany optimization
• async database execution
• query planner
• serializer caching
• Running speed increased to 50X of DRF

---

License

MIT

---