OpenInjectability

OpenInjectability is a transparent Python toolkit for estimating the pressure drop and predicted fluid-resistance force for a declared Newtonian liquid flowing through a specified needle.

Scientific boundary

The model is intentionally narrow:

• In scope: idealized Hagen–Poiseuille needle-fluid resistance for Newtonian liquids with traceable viscosity, needle ID/length, barrel ID, volume, and rate.
• Out of scope: syringe friction, break-loose force, device drivetrain losses, tissue backpressure, human factors, and non-Newtonian rheology.
• A needle gauge label is metadata only; a traceable needle inner diameter is required.
• The result is always named predicted fluid-resistance force, never total device force, “safe,” or “compliant.”

Non-Newtonian or otherwise unsupported inputs fail closed (typed rejection; no fallback rheology model).

Quick start

python -m pip install -e ".[dev,reports]"
openinjectability assess examples/formulation.csv `
  --config examples/assessment.json `
  --results assessment.json `
  --report assessment.md `
  --plot force-vs-needle.png `
  --audit-bundle audit.zip

PDF and plot output require the reports extra. Markdown and HTML reports require no optional dependency beyond the core install.

openinjectability version
openinjectability schema --format json
openinjectability validate-input examples/formulation.csv
openinjectability validation-status --json
openinjectability assess-one --viscosity-value 35 --viscosity-temperature 25 `
  --use-temperature 25 --rheology-evidence RHEO-1 --needle-id 0.21 `
  --needle-length 12.7 --needle-source drawing --barrel-id 6.35 `
  --barrel-source syringe --volume 2 --injection-time 15

Literature / experimental panel comparison

Compare a panel JSON of measured fluid-side force (or pressure × barrel area) to the model. Report status is only experimental_comparison or insufficient_data — this command never claims independently_validated and does not advance package validation_status.

openinjectability validate-experimental `
  validation/experimental/panel_allmendinger2014_glycerol_digitized.json `
  --out validation/experimental/reports/allmendinger2014_digitized

Panel schema and rules: validation/experimental/README.md.

Documentation: docs/scientific-basis.md, docs/input-schema.md, docs/interpretation.md, docs/roadmap.md, docs/phase-status.md, VALIDATION.md, PROJECT_SPECIFICATION.md, project.md.

Backend engine API

For services, notebooks, or another application backend, use the in-memory engine. It does not read files, produce plots or reports, or invoke the CLI. It preserves the same Newtonian-only scientific boundary as the command line.

from openinjectability import AssessmentInput, OpenInjectabilityEngine

engine = OpenInjectabilityEngine()
result = engine.assess(case)  # `case` is a validated AssessmentInput
batch = engine.assess_batch([case_a, case_b])
metadata = engine.metadata.to_dict()

assess_batch preserves accepted input order and records each invalid or unsupported case as a typed RejectedAssessment; it never applies a fallback rheology model. The returned value remains the predicted fluid-resistance force, not total device force.

Status

Version 0.1.0 is alpha. Development status on PyPI classifiers is 3 - Alpha.

• Do not describe this release as experimentally validated.
• Digitized figures, geometry setups from papers, or exploratory literature comparisons under validation/experimental/ are not accepted independent experimental validation evidence.
• See VALIDATION.md for the claims-to-evidence map and status vocabulary, and docs/phase-status.md for phase tracking.