Skip to main content

Cliente Python para mnemonexus — memória pra agentes de IA, open source, em português.

Project description

mnemonexus

License: MIT Python: 3.10+

Cliente Python oficial para o mnemonexus — memória pra agentes de IA, open source, em português.

Paridade com mem0 nas operações CRUD, mais o diferencial: compilação Karpathy, twice-rule pra lessons, grafo de entidades, staleness/supersede automáticos.

Instalação

pip install mnemonexus

Uso

from mnemonexus import Nexus

nexus = Nexus(api_url="http://localhost:8080", token="seu-token-sanctum")

# Paridade mem0
nexus.add("Reunião terça 14h", domain="memory")
result = nexus.search("quando é a reunião?")
print(result.answer)

# Twice rule — lesson aprendida
nexus.lesson(
    title="Sempre validar embedding dim antes de inserir em pgvector",
    body="Dim 1536 ≠ 1024 → INSERT explode silenciosamente",
    severity="critical",
)

# Decision com rationale
nexus.decision(
    title="Adotar pgvector em vez de Weaviate",
    description="Reduz containers; busca híbrida fica dentro de Eloquent",
    category="architecture",
)

# Grafo de entidades
graph = nexus.entity("foxpulse").graph()
for edge in graph.edges:
    print(edge.from_entity_slug, "→", edge.kind, "→", edge.to_entity_slug)

# Async também (recomendado em apps async)
import asyncio
from mnemonexus import AsyncNexus

async def main():
    async with AsyncNexus(token="...") as nexus:
        await nexus.add("…")
        r = await nexus.search("…")
asyncio.run(main())

API resumida

Paridade mem0

Método Endpoint O que faz
add(content, **) POST /ingest Adiciona memória
search(query, **) POST /query Recall híbrido (FTS + vetor)
get(slug) GET /pages/{slug} Lê página
update(slug, content) PUT /pages/{slug} Supersede
delete(slug) DELETE /pages/{slug} Hard delete
history(slug) GET /pages/{slug} Cadeia de versões

Diferencial mnemonexus

Método Endpoint O que faz
compile(slug) POST /pages/{slug}/compile Recompila wiki page
extract_facts(text) POST /facts/extract Facts + entidades + relações
lesson(title, body) POST /lessons Lesson com twice-rule
decision(title, description) POST /decisions Decisão com rationale
compact_session(id) PUT /sessions/{id} Fecha + sumariza sessão
entity(slug).graph() GET /entities/{slug}/graph Grafo BFS da entidade
search_entity(q) GET /entities Busca grafo de entidades

Errors

NexusError é a base. Subclasses:

  • AuthError — 401/403
  • NotFoundError — 404
  • ValidationError — 422

Cookbook

Veja examples/:

  1. 01_hello_world.py — add + search
  2. 02_with_karpathy.py — wiki compilada
  3. 03_with_lessons.py — twice rule
  4. 04_multi_tenant.py — isolamento por agent_id/tenant
  5. 05_rag_over_wiki.py — RAG com histórico

Dev

git clone https://github.com/lrdeoliveira/mnemonexus-python
cd mnemonexus-python
pip install -e ".[dev]"
pytest

Licença

MIT.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

mnemonexus-0.1.0.tar.gz (10.3 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

mnemonexus-0.1.0-py3-none-any.whl (9.0 kB view details)

Uploaded Python 3

File details

Details for the file mnemonexus-0.1.0.tar.gz.

File metadata

  • Download URL: mnemonexus-0.1.0.tar.gz
  • Upload date:
  • Size: 10.3 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.9.6

File hashes

Hashes for mnemonexus-0.1.0.tar.gz
Algorithm Hash digest
SHA256 60be41406e42f2bc9ffe89f6525018a33137611a2a0e9d60322318e3985c675f
MD5 f898afd4fb2039cbc39ce4dadb98ed01
BLAKE2b-256 d6b6de195fc5ea1a70034689c28c5292c8044d1c25c833dca0d243f01bb41a1f

See more details on using hashes here.

File details

Details for the file mnemonexus-0.1.0-py3-none-any.whl.

File metadata

  • Download URL: mnemonexus-0.1.0-py3-none-any.whl
  • Upload date:
  • Size: 9.0 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.9.6

File hashes

Hashes for mnemonexus-0.1.0-py3-none-any.whl
Algorithm Hash digest
SHA256 ee7a72e993705b1376bb863055f7dca7d3b3202496956fd4db70eb44546fb0c1
MD5 99be5a1b7048144531e81cb15156260c
BLAKE2b-256 9e3374be058a8b51e5a9a9bb082eb9e281ab036408e8693ad4d117a592f293be

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page