pypersistent 2.0.0a1

High-performance persistent (immutable) data structures for Python

PyPersistent

A high-performance collection of persistent (immutable) data structures for Python, written in C++.

Features

• Immutable: All operations return new maps, leaving originals unchanged
• Structural Sharing: New versions share most structure with old versions (O(log n) copies instead of O(n))
• Fast: 38% faster than pure Python implementation for insertions, ~10x slower than mutable dict
• Thread-Safe: Immutability makes concurrent access safe without locks
• Python 3.13+ Free-Threading Ready: Lock-free design with atomic reference counting for true parallelism
• Memory Efficient: Structural sharing minimizes memory overhead
• Dual Interface: Both functional (Clojure-style) and Pythonic APIs

Data Structures

PyPersistent provides five core persistent data structures:

PersistentMap

Unordered key-value map based on Hash Array Mapped Trie (HAMT).

• Use for: General-purpose dictionary needs with immutability
• Time complexity: O(log₃₂ n) ≈ 6 steps for 1M elements
• Features: Fast lookups, structural sharing, bulk merge operations
• Example:

  from pypersistent import PersistentMap
  
  m = PersistentMap.create(name='Alice', age=30)
  m2 = m.set('city', 'NYC')
  m3 = m | {'role': 'developer'}  # Merge
  

PersistentTreeMap

Sorted key-value map based on Left-Leaning Red-Black Tree.

• Use for: Ordered data, range queries, min/max operations
• Time complexity: O(log₂ n) for all operations
• Features: Sorted iteration, range queries (subseq/rsubseq), first/last
• Example:

  from pypersistent import PersistentTreeMap
  
  m = PersistentTreeMap.from_dict({3: 'c', 1: 'a', 2: 'b'})
  list(m.keys())  # [1, 2, 3] - always sorted
  
  # Range queries
  sub = m.subseq(start=1, end=2, start_inclusive=True, end_inclusive=False)
  list(sub.keys())  # [1]
  
  # Min/max
  m.first()  # (1, 'a')
  m.last()   # (3, 'c')
  

PersistentVector

Indexed sequence based on bit-partitioned vector trie (RRB-Tree variant).

• Use for: Ordered sequences with efficient random access and append
• Time complexity: O(log₃₂ n) for get/set, O(1) for append
• Features: Fast indexed access, efficient append, slicing
• Example:

  from pypersistent import PersistentVector
  
  v = PersistentVector.create(1, 2, 3)
  v2 = v.conj(4)  # Append
  v3 = v2.assoc(0, 10)  # Update index 0
  v[1]  # 2 - indexed access
  

PersistentHashSet

Unordered set based on HAMT (same as PersistentMap).

• Use for: Unique collection of items, set operations
• Time complexity: O(log₃₂ n) for add/remove/contains
• Features: Fast membership testing, set operations (union, intersection, difference)
• Example:

  from pypersistent import PersistentHashSet
  
  s = PersistentHashSet.create(1, 2, 3)
  s2 = s.conj(4)  # Add
  s3 = s.disj(2)  # Remove
  2 in s  # True
  

PersistentArrayMap

Small map optimization using array of key-value pairs.

• Use for: Maps with < 8 entries (automatic optimization)
• Time complexity: O(n) linear scan, but faster than HAMT for tiny maps
• Features: Lower memory overhead, faster for very small maps
• Note: Typically used internally; PersistentMap automatically uses this for small maps

Choosing the Right Data Structure

Need	Use	Why
General key-value storage	PersistentMap	Fastest for unordered data
Sorted keys / range queries	PersistentTreeMap	Maintains sort order, supports ranges
Indexed sequence	PersistentVector	Fast random access and append
Unique items / set operations	PersistentHashSet	Membership testing, set algebra
Very small maps (< 8 items)	PersistentArrayMap	Lower overhead for tiny maps

Performance

Quick Summary

pypersistent provides 6-8% faster bulk operations compared to baseline and is 3-150x faster than pure Python pyrsistent for most operations. The real value is structural sharing: creating variants is 3000x faster than copying dicts.

Benchmark Results

vs Python dict (1M elements)

Operation	pypersistent	dict	Notes
Construction	2.74s	299ms	9x slower (immutability cost)
Lookup (1K ops)	776ms	427ms	1.8x slower
Update (single)	147µs	~80ns	Comparable for single ops
Structural Sharing (100 variants)	0.48ms	1.57s	3000x FASTER

Key insight: For creating multiple versions, pypersistent is orders of magnitude faster.

vs pyrsistent (pure Python) - Apples-to-Apples

Size	Operation	pypersistent	pyrsistent	Speedup
1M	Merge (500K+500K)	11.5ms	1.60s	139x faster
1M	Construction	185ms	813ms	4.4x faster
1M	Lookup (1M ops)	726ms	796ms	1.1x faster
1M	Iteration	452ms	167ms	2.7x slower
1K	Merge (500+500)	8.7µs	1.22ms	141x faster
1K	Construction	76µs	192µs	2.5x faster
100	Merge (50+50)	17.8µs	120µs	6.7x faster
100	Construction	30µs	18.5µs	1.6x slower

Iteration note: Using items_list() instead of items() makes iteration 1.7-3x faster for maps < 100K

Performance by Map Size

Size	Construction	Lookup (1K ops)	Sharing (100 variants)
100	73µs	19µs	77µs
1K	765µs	205µs	95µs
10K	9.7ms	228µs	108µs
100K	110ms	255µs	133µs
1M	2.74s	776ms	158µs

Fast Iteration Methods

For complete iteration over small-medium maps, use materialized list methods:

m = PersistentMap.from_dict({...})

# Fast methods (1.7-3x faster for maps < 100K)
items = m.items_list()   # Returns list of (key, value) tuples
keys = m.keys_list()     # Returns list of keys
values = m.values_list() # Returns list of values

# Lazy iterators (better for very large maps or early exit)
for k, v in m.items():   # Generator, O(log n) memory
    ...

Performance:

• Maps ≤ 10K: 3x faster with items_list()
• Maps ≤ 100K: 1.7x faster with items_list()
• Maps > 100K: Use iterator (lazy, memory-efficient)

When to Use pypersistent

Use pypersistent when:

• Creating multiple versions of data (undo/redo, time-travel)
• Concurrent access across threads (lock-free reads)
• Functional programming patterns
• Merging large maps frequently

Use dict when:

• Single mutable map is sufficient
• Maximum raw construction speed is critical
• Memory per entry is constrained

Technical Details

Implementation: C++ HAMT (Hash Array Mapped Trie) with:

• Bottom-up bulk construction for from_dict()
• Arena allocator for fast node allocation
• Structural tree merging for merge()
• COW semantics for collision nodes
• Fast iteration with pre-allocated lists

Time Complexity: O(log₃₂ n) ≈ 6 steps for 1M elements Space Complexity: O(n) with structural sharing across versions

For detailed performance analysis and benchmarking methodology, see docs/.

Installation

pip install pypersistent

Or build from source:

git clone https://github.com/cmarschner/pypersistent.git
cd pypersistent
python setup.py install

Usage

PersistentMap - Hash Map

from pypersistent import PersistentMap

# Create
m = PersistentMap.create(name='Alice', age=30)
m = PersistentMap.from_dict({'name': 'Alice', 'age': 30})

# Add/update (functional style)
m2 = m.assoc('city', 'NYC')

# Add/update (Pythonic style)
m2 = m.set('city', 'NYC')
m3 = m | {'role': 'developer'}  # Merge

# Get
name = m['name']  # Raises KeyError if missing
name = m.get('name', 'default')

# Remove
m4 = m.dissoc('age')  # Functional
m4 = m.delete('age')  # Pythonic

# Check membership
'name' in m  # True

# Iterate
for key, value in m.items():
    print(key, value)

PersistentTreeMap - Sorted Map

from pypersistent import PersistentTreeMap

# Create (automatically sorted by keys)
m = PersistentTreeMap.from_dict({3: 'c', 1: 'a', 2: 'b'})
list(m.keys())  # [1, 2, 3]

# Same API as PersistentMap
m2 = m.assoc(4, 'd')
m3 = m.dissoc(1)

# Sorted-specific operations
first_entry = m.first()  # (1, 'a')
last_entry = m.last()    # (3, 'c')

# Range queries
sub = m.subseq(start=1, end=3)  # Keys 1-2 (end exclusive)
list(sub.keys())  # [1, 2]

rsub = m.rsubseq(start=3, end=1)  # Reverse order
list(rsub.keys())  # [3, 2]

PersistentVector - Indexed Sequence

from pypersistent import PersistentVector

# Create
v = PersistentVector.create(1, 2, 3)
v = PersistentVector.from_list([1, 2, 3])

# Append (functional)
v2 = v.conj(4)  # [1, 2, 3, 4]

# Update by index
v3 = v.assoc(0, 10)  # [10, 2, 3]

# Access by index
first = v[0]  # 1
last = v[-1]  # 3

# Slice
sub = v[1:3]  # PersistentVector([2, 3])

# Iterate
for item in v:
    print(item)

# Length
len(v)  # 3

PersistentHashSet - Set

from pypersistent import PersistentHashSet

# Create
s = PersistentHashSet.create(1, 2, 3)
s = PersistentHashSet.from_set({1, 2, 3})

# Add
s2 = s.conj(4)  # {1, 2, 3, 4}

# Remove
s3 = s.disj(2)  # {1, 3}

# Membership
2 in s  # True

# Set operations
s1 = PersistentHashSet.create(1, 2, 3)
s2 = PersistentHashSet.create(3, 4, 5)

union = s1 | s2  # {1, 2, 3, 4, 5}
intersection = s1 & s2  # {3}
difference = s1 - s2  # {1, 2}

# Iterate
for item in s:
    print(item)

API Summary

Common operations (all data structures):

• create(*args) - Create from elements
• from_X(...) - Create from Python collection
• Immutability - all operations return new instances
• Thread-safe reads - safe to share across threads

PersistentMap / PersistentTreeMap:

• assoc(k, v) / set(k, v) - Add/update
• dissoc(k) / delete(k) - Remove
• get(k, default=None) - Get value
• m[k] - Get (raises KeyError)
• k in m - Membership
• keys(), values(), items() - Iterators
• m1 | m2 - Merge

PersistentTreeMap only:

• first() - Min entry
• last() - Max entry
• subseq(start, end) - Range query
• rsubseq(start, end) - Reverse range

PersistentVector:

• conj(item) - Append
• assoc(idx, val) - Update by index
• v[idx] - Get by index
• v[start:end] - Slice
• len(v) - Length

PersistentHashSet:

• conj(item) - Add
• disj(item) - Remove
• item in s - Membership
• s1 | s2 - Union
• s1 & s2 - Intersection
• s1 - s2 - Difference

Use Cases

Use persistent data structures when:

• Creating multiple versions of data (undo/redo, time-travel, version history)
• Sharing data across threads without locks or defensive copying
• Functional programming patterns (immutability, pure functions)
• Creating modified copies is frequent (structural sharing makes this fast)
• Python 3.13+ free-threading for true parallelism

Use mutable collections when:

• Single mutable instance is sufficient
• Maximum raw construction speed is critical
• Memory per entry is highly constrained

Specific use cases:

• Config management: Share base config, each thread creates customized version
• Event sourcing: Maintain history of all states efficiently
• Reactive programming: Pass immutable state between components
• Concurrent caching: Multiple threads read/update cache without locks
• Functional data pipelines: Transform data through pipeline stages

How It Works

PyPersistent implements multiple classic persistent data structures:

PersistentMap - Hash Array Mapped Trie (HAMT)

Based on the HAMT data structure used by Clojure, Scala, and Haskell:

• 32-way branching tree indexed by hash bits
• Path copying for immutability
• Structural sharing between versions
• std::shared_ptr for entry sharing (44x fewer INCREF/DECREF)
• Inline storage with std::variant for cache-friendly access

PersistentTreeMap - Left-Leaning Red-Black Tree

Self-balancing binary search tree with:

• Path copying for immutability
• Sorted order maintenance (O(log₂ n))
• Atomic reference counting for node sharing
• Range query support via tree traversal

PersistentVector - Bit-Partitioned Trie

32-way branching tree for indexed access:

• Path copying for updates
• Tail optimization for fast append
• O(log₃₂ n) random access (~6 steps for 1M elements)
• Efficient slicing via structural sharing

PersistentHashSet - HAMT-based Set

Uses PersistentMap internally with:

• Keys as set elements
• Same O(log₃₂ n) complexity
• Set algebra operations

PersistentArrayMap - Simple Array

Linear array for tiny maps (< 8 entries):

• Lower overhead than HAMT for small sizes
• O(n) operations but faster than tree for n < 8
• Automatically used by PersistentMap when beneficial

Technical Details

Complexity:

• PersistentMap/HashSet: O(log₃₂ n) ≈ 6 steps for 1M elements
• PersistentTreeMap: O(log₂ n) for all operations
• PersistentVector: O(log₃₂ n) get/set, O(1) append
• PersistentArrayMap: O(n) but fast for n < 8

Implementation:

• C++ with pybind11 bindings
• Atomic reference counting (std::atomic<int>)
• Structural sharing for memory efficiency
• Thread-safe reads (fully immutable)

Python 3.13+ Free-Threading Support

PyPersistent is fully compatible with Python 3.13's experimental free-threading mode (nogil), making it ideal for parallel workloads:

Why It Works

1. Lock-Free Reads: Immutable data structure allows concurrent reads without synchronization
2. Atomic Reference Counting: Internal C++ reference counting uses std::atomic operations
3. Thread-Safe Entry Storage: Uses std::shared_ptr with built-in thread-safe reference counting
4. Independent Updates: Each thread can create new versions without blocking others

Usage with Free-Threading

# Python 3.13+ with --disable-gil or PYTHON_GIL=0
import threading
from pypersistent import PersistentMap

# Shared base map
base_config = PersistentMap.create(
    api_url='https://api.example.com',
    timeout=30,
    retries=3
)

def process_request(thread_id, data):
    # Each thread creates its own version - no locks needed!
    my_config = base_config.set('thread_id', thread_id)
    my_config = my_config.set('request_data', data)

    # Concurrent reads are completely lock-free
    url = my_config['api_url']
    timeout = my_config['timeout']

    # Do work...
    return my_config

# Spawn threads - true parallelism without GIL!
threads = []
for i in range(100):
    t = threading.Thread(target=process_request, args=(i, f'data_{i}'))
    threads.append(t)
    t.start()

for t in threads:
    t.join()

Performance Benefits

Without the GIL:

• Parallel reads: Multiple threads read simultaneously without contention
• Parallel updates: Each thread creates new versions independently
• No lock overhead: Zero synchronization cost for immutable operations
• Structural sharing shines: Creating 100 thread-local variants is 3000x faster than copying dicts

Enable Free-Threading

# Python 3.13+
python3.13 --disable-gil your_script.py

# Or set environment variable
export PYTHON_GIL=0
python3.13 your_script.py

Note: Free-threading is experimental in Python 3.13. Some packages may not be compatible yet.

Development

# Install development dependencies
pip install pytest pybind11

# Build extension
python setup.py build_ext --inplace

# Run all tests
pytest -v

# Run specific data structure tests
pytest test_persistent_map.py -v
pytest test_persistent_tree_map.py -v
pytest test_persistent_vector.py -v
pytest test_persistent_hash_set.py -v

# Run performance benchmarks
python performance_test.py
python performance_vector.py

License

MIT License - see LICENSE file for details

Credits

Inspired by Clojure's persistent data structures and the HAMT paper by Bagwell (2001).

Implementation by Clemens Marschner.