sme-core 2.0.0

Sequence Method Engine — core validators, regression diff_report, and ingestion-pipeline base classes. Consumed by Scriptura SaaS and any other SME wrapper.

sme-core

Sequence Method Engine — core validators, regression diff_report, and ingestion-pipeline base classes.

sme-core is the Python runtime library for the Sequence Method Engine (SME). It packages canonical JSON schemas, schema validation, the §7.2.6 regression diff_report, and the abstract ingestion pipeline base used by P1 (PubMed), P2 (FDA), P3 (DEA), P4 (503A bulks), and future wrapper-defined pipelines.

Companion package: @scripturahealth/sme-schemas ships the same schemas as TypeScript types for the Scriptura SaaS frontend.

Install

pip install sme-core

Requires Python ≥ 3.10.

Usage

Validate a document

from sme_core import SchemaValidator

v = SchemaValidator()
errors = v.validate(my_compound_dict, schema_name="compound")
if errors:
    for e in errors:
        print(".".join(str(p) for p in e.absolute_path), "→", e.message)

Regression diff_report

from sme_core import generate_diff_report

report = generate_diff_report(
    baseline_protocol_json=old,
    new_protocol_json=new,
    baseline_id="2025-12-14-sabrina-v2.3",
    baseline_generated_at="2025-12-14T18:32:00Z",
)
print(report.drift_verdict)              # no_clinical_drift | clinical_drift_detected
print(report.requires_eric_approval_before_ship)

Build an ingestion pipeline

from sme_core import IngestionPipeline, RegulatoryRecord, ReviewQueueItem

class PubMedPipeline(IngestionPipeline):
    name = "p1_pubmed"

    def fetch(self, since):
        # call PubMed E-utilities, return raw items
        ...

    def transform(self, raw_items):
        records = [...]       # list[RegulatoryRecord]
        review_items = [...]  # list[ReviewQueueItem]
        return records, review_items

p = PubMedPipeline(
    cursor_loader=lambda: db.get_cursor("p1_pubmed"),
    cursor_saver=lambda v: db.set_cursor("p1_pubmed", v),
    record_sink=db.append_regulatory_records,
    review_queue_sink=webhook_to_wrapper,
)
result = p.run()

CLI tools

sme-validate path/to/doc.json --schema compound
sme-diff-report baseline.json new.json

sme-diff-report exits 1 when clinical drift is detected — designed for CI gating.

Versioning

The package version tracks the SME schema_version. Tagging the repo v1.1.0 publishes both sme-core==1.1.0 and @scripturahealth/sme-schemas@1.1.0.

License

UNLICENSED — internal Scriptura Health distribution.