The interactive file manager requires Javascript. Please enable it or use sftp or scp.
You may still browse the files here.

Name	Modified	Size	InfoDownloads / Week
Parent folder
image-sbom.cdx.json	2026-09-15	1.4 MB	0
image-sbom.cdx.json.sha256	2026-09-15	86 Bytes	0
openmed-image-sbom_f93836209f1ace6b99b50e765dbf2881959401a9.cyclonedx.json	2026-09-15	1.4 MB	0
openmed-2.5.0-py3-none-any.whl.sigstore.json	2026-09-15	11.2 kB	0
openmed-2.5.0.tar.gz.sigstore.json	2026-09-15	11.2 kB	0
python-distributions.intoto.json	2026-09-15	11.0 kB	0
release-artifact-digests.txt	2026-09-15	184 Bytes	0
release-source.json	2026-09-15	194 Bytes	0
sbom.cdx.json	2026-09-15	8.9 kB	0
OpenMed v2.5.0 source code.tar.gz	2026-09-15	41.0 MB	1
OpenMed v2.5.0 source code.zip	2026-09-15	43.2 MB	0
README.md	2026-09-15	21.6 kB	0
Totals: 12 Items		87.1 MB	1

OpenMed 2.5: Clinical Context. Local Control.

OpenMed 2.5 brings clinical privacy and extraction previews, richer FHIR and OMOP exports, safer document intake, and new privacy-policy and audit tools to the Python SDK. It preserves the v2.3 family-based model-registry API while adding an explicit API for managing separate model tiers and formats.

This release covers changes since v2.3.0. Python 3.10+ remains supported, and the core dependency set is unchanged.

Clinical text and privacy

Clinical-preserving de-identification, in preview. ClinicalPrivacyProcessor combines explicit language, category, and role controls with document-level review status. The clinical_preserve profile targets identifiers while retaining clinical context such as diagnoses, medications, doses, and measurements. Full-document ONNX batching preserves source offsets and input order, with bounded cancellation and per-document outcomes.
More structured clinical evidence. Context, medication instructions, quantities, measurements, relations, and temporal analysis retain links to the original text. Ambiguous or incomplete quantities are withheld. German clinical-context, date, address, and term-preservation handling gains additional regression coverage.
Persian language-pack support. The fa pack adds Persian patterns, Eastern-digit normalization, Jalali date handling, and Iranian national-ID surrogate support. Persian is also available in the REST language enum. Its default model route remains a compatibility placeholder; this release does not supply newly qualified dedicated Persian weights.
Consistent locale handling. Structural language-tag normalization and stricter alias validation make language and locale selection more explicit.

Clinical-preserving processing is opt-in and remains a preview. No language/model route is clinically qualified by default; applications must establish qualification for their intended data, language, policy, and runtime.

FHIR, OMOP, and terminology

A consistent FHIR export entry point. The grounded-span to_fhir() facade dispatches supported entities to Condition, Observation, MedicationStatement, and Procedure resources, assembles deterministic Bundles, and reports unmapped labels. A conservative DiagnosticReport exporter adds explicit status handling and reference normalization.
More export validation. Dependency-free base FHIR R4 structural checks and US Core 9.0.0 checks cover supported resources. Additional validators check reference targets, profile declarations, and Observation extensions. FHIR R5 Bundle comparison reports structural round-trip differences.
Expanded OMOP CDM 5.4 coverage. Exporters add measurement, procedure_occurrence, visit_occurrence, observation_period, and note_nlp rows, preserving supported numeric values, units, dates, and source offsets. Cohort validation checks keys, relationships, vocabulary references, and note provenance.
Explicit terminology sources. ValueSet expansion works with caller-loaded vocabulary snapshots or an explicitly configured terminology endpoint. A local terminology cache records versioned provenance and rejects stale-release reuse. Restricted terminology requires caller-supplied access and permissions.

Synthetic FHIR/OMOP conformance coverage includes the official HL7 R4 validator, a malformed-resource negative control, and OMOP column, key, and reference checks. The Java validator and restricted vocabulary content are not bundled.

Document and multimodal intake

preflight_asset brings manifest, media-type, profile, resource-limit, and digest checks into one ordered accept-or-abstain report. An unevaluable check is reported as an abstention rather than a successful validation.

Bounded BMP, GIF, and WebP header inspection extends image preflight coverage. WAV handling gains additional envelope and RF64 refusal checks.
MOBILE_V1 and DESKTOP_V1 profiles bound bytes, dimensions, pages, frames, and duration before decoding. Batch summaries detect duplicate identifiers and digests and validate aggregate limits.
Memory-streamed Tesseract OCR and PDF/PNG export improvements add source-line, reading-order, and redaction-fidelity checks. OCR still requires a caller-installed Tesseract runtime.
ONNX inference validates label metadata before optional runtime imports and adds tensor batching and execution limits.

Privacy policies, structured data, and audit

New local tools make privacy decisions and their supporting evidence easier to inspect:

Policy controls: versioned policy schemas, composition and migration checks, minimum-necessary field selection, and field-level FHIR/OMOP de-identification with patient-consistent date shifting.
Structured-data checks: tabular re-identification risk and schema-drift reports, surrogate-map integrity checks, and nested-redaction idempotence checks.
Audit evidence: structured access reviews, evidence-bundle integrity and replay verification, key-custody and rotation metadata, and bounded exception and retention reports.
Artifact protection: authenticated encryption for reversible surrogate mappings with caller-owned keys, local deletion planning and verification, dataset-upload guards, privacy scanning for Git/CI workflows, and completed session-trace scrubbing.

These reporting APIs use bounded metadata, counts, offsets, hashes, and provenance instead of retaining protected source values. Deletion, upload, trace cleanup, and other external effects remain explicit operations.

The optional Snowpark adapter adds caller-managed warehouse de-identification. The separate privacy-proxy application requires an injected transport, keeps mappings scoped to each request, and supports placeholder restoration across streaming responses.

Agent and training workflows

Agent integrations gain strict provider-result and run-summary schemas, opaque correlation identifiers, typed governance identifiers, and artifact references with digest and size metadata.

Federated-training helpers add scheduling windows, round-status summaries, validated update metadata, and aggregate metrics with clipping declarations and minimum-group suppression. Retraining tools rank aggregate evidence, score trigger signals, and prepare recipe proposals. Compute, cost, energy, and carbon accounting adds explicit budget reports.

These additions provide orchestration and evidence contracts. They do not automatically train, deploy, or qualify a model.

Compatibility and fixes

Existing v2.3 registry callers keep their family-based API. RegistryService, family CLI selectors, unambiguous model aliases, and caller-owned schema-v1 files remain supported. An SDK upgrade does not require migrating those files.

Applications that need separate channels for model tiers or formats can opt into SlotRegistryService and schema-v2 slots keyed by family::tier::format. Stored slot versions are independent of version-like tokens in model repository names. The compatibility adapter retains existing schema-v2 files and refuses ambiguous multi-slot family operations without writing.

The static Python API comparison against v2.3.0 reports zero removed symbols, zero narrowed signatures, and zero new deprecations. The REST API retains its existing paths and component schemas, with Persian added to the language enum.

Other fixes include compatible Swift dependency resolution, stricter ONNX metadata validation, request-scoped privacy-proxy restoration, Windows artifact-deletion identity checks, and refreshed documentation and container dependencies.

The clinical, governance, and orchestration additions described above are Python capabilities. Matching npm, Swift, and Android version numbers do not imply that every new Python API is implemented on those platforms.

Upgrade

Use these coordinates once the v2.5.0 packages and tag are published:

:::bash
# Python SDK
pip install --upgrade "openmed==2.5.0"

# Optional Hugging Face and FHIR integrations
pip install --upgrade "openmed[hf,fhir]==2.5.0"

# JavaScript package
npm install openmed@2.5.0

Swift Package Manager:

:::swift
.package(url: "https://github.com/maziyarpanahi/openmed.git", from: "2.5.0")

Android through JitPack:

:::kotlin
implementation("com.github.maziyarpanahi:openmed:v2.5.0")

Container:

:::bash
docker pull ghcr.io/maziyarpanahi/openmed:v2.5.0

For Helm deployments, select the release image with image.tag=v2.5.0. Review the migration guide before adopting slot registries, new serialized schemas, or the clinical-preserving preview.

Use and qualification

OpenMed is assistive software. Clinical outputs require qualified review and must not automatically determine diagnosis, treatment, or billing. Structural validation and synthetic regression tests do not establish clinical efficacy, deployment-specific privacy performance, or regulatory compliance.

Before deployment, evaluate identifier recall, critical leakage, source-span integrity, clinical-text preservation, date handling, and surrogate consistency on the intended languages and runtimes. Model artifacts and their existing release targets are unchanged by this SDK release.