Omnesight Python SDK
The official Python client library for Omnesight Sovereign Operational Intelligence & Autonomous Execution Infrastructure.
Overview
The Omnesight Python SDK provides typed access to Omnesight's sovereign Non-LLM scientific engines (Hackmalim v1.0, Sandalfon v1.0) alongside enterprise operational intelligence gateways:
- Hackmalim v1.0 (
client.hackmalim): 3-layer Non-LLM control combining Continuous-Time Liquid Neural ODEs, Kirchhoff Physics-Informed GNNs, and Pearl Do-Calculus + SAT Axiom Verification. - Sandalfon v1.0 (
client.sandalfon): Sparse differential equation discovery (SINDy/STLSQ), FFT spectral wave precursor extraction, and Popperian permutation falsification. - Ingest Multi-Source Telemetry: Stream telemetry from SCADA, IoT sensors, GPS fleets, and warehouse terminals in sub-second latency.
- Closed-Loop ERP Reconciliation: Automatically detect and resolve transactional variances across SAP S/4HANA, Oracle NetSuite, Microsoft Dynamics 365, and financial ledgers.
- Supervised Action Protocol (SAP): Execute automated operational interventions under deterministic governance policies and cryptographic audit trails.
Installation
pip install omnesight
Quickstart
1. Sovereign Scientific & Control Engines (Hackmalim & Sandalfon)
from omnesight import Omnesight
# Reads OMNESIGHT_API_KEY and OMNESIGHT_BASE_URL (defaults to http://127.0.0.1:8001)
client = Omnesight(
api_key="omne_live_your_secret_key",
base_url="https://mlops-api.omnesight.com",
)
# Sandalfon v1.0: Extract sparse ODEs + spectral FFT wave precursors
discovery = client.sandalfon.discover(
preset="thermal_runaway",
noise_level=0.02,
sparsity_threshold=0.10,
precursor_amp=0.18,
hypothesis_text="Cable thermal overload precursor detection",
)
print(discovery.sindy["discovered_eq_x"], discovery.hypothesis_test["z_score"])
dialogue = client.sandalfon.chat(
message="Analyse les précurseurs thermiques sur les câbles",
preset="thermal_runaway",
)
print(dialogue.spoken_reply)
# Hackmalim v1.0: Closed-loop Liquid ODE + PI-GNN + Causal SAT Verifier
res = client.hackmalim.infer(
shock_amplitude=1.45,
leak_node=2,
leak_magnitude=22.5,
p_max_kw=55.0,
t_max_c=72.0,
)
print(res.target_entity_id, res.enterprise_guardrail)
2. Synchronous Enterprise Telemetry Client
from omnesight import Omnesight
# Initialize the client (reads OMNESIGHT_API_KEY & OMNESIGHT_BASE_URL from environment if omitted)
client = Omnesight(
api_key="your_api_key_here",
base_url="https://omnesight.internal.company.com" # Your private VPC or dedicated sovereign gateway
)
# 1. Ingest asset telemetry
response = client.telemetry.ingest(
stream_id="freight-corridor-lagos-kano",
records=[
{
"device_id": "TRUCK-GPS-104",
"metric": "cargo_seal_integrity",
"value": 1.0,
"metadata": {"driver": "ID-992", "cargo": "Refined Petroleum"}
}
]
)
print(f"Batch {response.batch_id} ingested in {response.duration_ms}ms")
# 2. Scan ERP vs Ledger Discrepancies
discrepancies = client.reconciliation.scan_discrepancies(
erp_system="SAP",
entity_type="PURCHASE_ORDERS",
lookback_days=7
)
for d in discrepancies:
print(f"Discrepancy #{d.discrepancy_id}: Variance ${d.variance_amount_usd:,.2f} ({d.status})")
2. Asynchronous Client (AsyncIO / High Concurrency)
For data engineering pipelines or FastAPI backends handling thousands of events per second:
import asyncio
from omnesight import AsyncOmnesight
async def main():
async with AsyncOmnesight(api_key="your_api_key_here") as client:
response = await client.telemetry.ingest(
stream_id="scada-compressor-alpha",
records=[
{"device_id": "CMP-01", "metric": "vibration_hz", "value": 44.2}
]
)
print(f"Ingested batch: {response.batch_id}")
asyncio.run(main())
Enterprise VPC & Sovereign Deployments
For single-tenant institutional customers running Omnesight within a private VPC (AWS, Azure, GCP, Alibaba Cloud) or air-gapped on-premise Kubernetes cluster:
client = Omnesight(
api_key="your_sovereign_vpc_token",
base_url="https://omnesight.internal.company.com", # Private VPC endpoint
timeout=60.0
)
Supervised Action Protocol (SAP)
Automated write-backs to core ERPs and ledger reconciliations require policy validation:
from omnesight.models import ActionApprovalRequest
# Retrieve status of an automated action staged by the cognitive engine
action = client.actions.get_status("ACT-88492")
print(f"Action: {action.action_type} | Impact: ${action.financial_impact_usd} | State: {action.state}")
# Submit human-in-the-loop supervisor approval
client.actions.submit_decision(ActionApprovalRequest(
action_id="ACT-88492",
action_type="ERP_LEDGER_WRITE",
target_entity="PO-77189",
decision="APPROVE",
supervisor_id="SUP-MAV-01",
supervisor_notes="Validated with physical bill of lading #BL-9921"
))
Security & Data Sovereignty
- Zero Frontier Model Training: Omnesight never ingests customer transactional payloads or telemetry into public foundation model weights.
- Cryptographic Traceability: Every staged action is signed with a SHA-256 audit hash.
- Security Advisories: For vulnerability disclosure, see SECURITY.md or email
security@omnesight.com.
License
Apache License 2.0. See LICENSE for details.
Metadata
Release files for omnesight 1.1.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| omnesight-1.1.0.tar.gz | 13.9 kB | Without provenance | Details for omnesight-1.1.0.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| omnesight-1.1.0-py3-none-any.whl | Python 3 | none | any | Details for omnesight-1.1.0-py3-none-any.whl |
Total release size: 29.2 kB
Release files / omnesight-1.1.0.tar.gz
| Download URL | omnesight-1.1.0.tar.gz |
|---|---|
| Size | 13.9 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
fdf03f73aceb7a8e112cfd3fc3518f110c5822c2cc60996ccef8665a94a33fbf
|
|
BLAKE2b-256 checksum How to use checksums |
af10739454c8499d4be51f06db22c25d07cfb3473acf8c936d913e7fe4452b6c
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/7.0.0 CPython/3.14.6
|
Release files / omnesight-1.1.0-py3-none-any.whl
| Download URL | omnesight-1.1.0-py3-none-any.whl |
|---|---|
| Size | 15.3 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
9a66ec4047412f411ee52b04ea4419998858c6e74de17443b425593d811a2d24
|
|
BLAKE2b-256 checksum How to use checksums |
a4299d2f9475c255cea62fe8e67e2805084061db3d939b1015f806cfc709b29c
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/7.0.0 CPython/3.14.6
|
Security for omnesight 1.1.0
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.