43-EFT.WP.Data.DatasetCards v1.0 | Chapter 5 Optional Extensions

Home ／ Docs-Technical WhitePaper ／ 43-EFT.WP.Data.DatasetCards v1.0

Chapter 5 Optional Extensions

I. Chapter Purpose & Scope

II. Layering & Activation Principles

• Optional: Enable when the scenario demands higher auditability/reproducibility/compliance transparency. If enabled, the field must satisfy this chapter’s type/regex/dependency constraints and pass Lint.
• Cross-volume dependencies: Data contracts & export per Core.DataSpec v1.0; units & dimensions per Core.Metrology v1.0; path/arrival-time equations per Core.Equations v1.1.

III. Optional Extensions — Master Table

IV. Field Specs & Examples

1) lineage

• Semantics: Upstream datasets recorded as dataset_id@version.
• Constraint: Each entry must be resolvable; public citations should prefer stable minor line (e.g., v1.*).
• Example:
• lineage:
• - "eift.obs.sky_survey@v1.2"
• - "eift.sim.noise_bank@v1.0"

2) related_artifacts

• Semantics: Scripts, configs, baseline models strongly bound to the dataset.
• Constraint: URI/relative path must resolve; artifacts should also appear in export_manifest.artifacts[] with sha256.
• Example:
• related_artifacts:
• - "scripts/preprocess_v2.py"
• - "configs/card_lint.yaml"
• - "baselines/frb_cls_v1.onnx"

3) sensor_profile

• Semantics: Sensor/array/channel profiling and calibration.
• Fragment:
• sensor_profile:
• instruments: [{name:"LOFAR", station:"DE601"}]
• channels: [{id:"LBA-110", fs_hz: 195312.5, bw_hz: 3.9e6, pol:"XY"}]
• calibration: {method:"noise-diode", date:"2025-06-01"}
• Dependency: Align with Methods/Metrology series.

4) privacy / ethics

• Semantics: Privacy policy, de-identification, ethical disclosures, usage restrictions.
• Constraint: Enumerated policies with traceable explanations; external posture consistent with Core.DataSpec v1.0.
• Example:
• privacy:
• policy: "no-PII"
• deidentification: ["hash-id","mask-location"]
• ethics:
• intended_use: ["academic","benchmark"]
• restrictions: ["no_biometric_identification"]

5) uncertainty

• Semantics: Components, propagation, and combination; system/random split.
• Fragment:
• uncertainty:
• components:
• - {name:"cal_gain", type:"systematic", value:0.8, unit:"%", method:"lab"}
• - {name:"thermal", type:"random", value:2.1, unit:"K", method:"propagate"}
• combine: "rss"
• coverage: "k=2"
• Dependency: Dimensional/units baseline & error-budget workflow.

6) path_corrections

• Semantics: Apply medium/line corrections and distribute uncertainty for path-dependent quantities (e.g., T_arr).
• Fragment:
• path_corrections:
• models: ["iono","troposphere","instrumental"]
• parameters: {pwv_mm: 6.0, tec: 8.5}
• uncertainty: {iono:0.7e-9, tropo:0.3e-9, instr:0.2e-9}
• see:
• - "EFT.WP.Core.Equations v1.1:S20-1"
• Dependency: Arrival-time anchors.

7) timebase_sync

• Semantics: Timebase calibration, sync strategy, and switchover thresholds across platforms.
• Fragment:
• timebase_sync:
• references: ["GPSDO","Rb-clock"]
• drift_ppb_max: 5
• switchover: {threshold_ppb: 10, holdover_s: 1800}
• audits: ["TIE","MTIE"]
• Dependency: Metrology timebase/sync.

8) distribution

• Semantics: Mirrors, sharding, rate limits, and regional compliance.
• Fragment:
• distribution:
• mirrors: ["https://mirror-a.example/datasets/foo/","s3://bucket/foo/"]
• shards: [{path:"train-000.tgz", bytes:134217728}, {path:"train-001.tgz", bytes:134217728}]
• rate_limit: {mbps: 50}
• Dependency: Export strategy & manifest.

V. Consistency for Metrology & Path-Dependent Data (when enabled)

1. Any occurrence of T_arr must register in the card: delta_form, path="gamma(ell)", measure="d ell", and pass check_dim. Two equivalent expressions:
   • T_arr = ( 1 / c_ref ) * ( ∫ n_eff d ell )
   • T_arr = ( ∫ ( n_eff / c_ref ) d ell ).
2. No Chinese in formulas; wrap all inline symbols with backticks; any division/integral/composite operator must use parentheses.

VI. Machine-Readable Schema (Excerpt, Normative)

# I15-5 Optional Extensions (excerpt)

properties:

lineage:

type: array

items: {type: string, pattern: "^[a-z0-9_\\-\\.]+@v\\d+\\.\\d+(\\.\\d+)?$"}

related_artifacts:

type: array

items: {type: string, minLength: 1}

sensor_profile:

type: object

properties:

instruments: {type: array, items: {type: object}}

channels: {type: array, items: {type: object, properties:{id:{type:string}, fs_hz:{type:number}, bw_hz:{type:number}, pol:{type:string}}}}

calibration: {type: object}

privacy:

type: object

properties: {policy:{type:string}, deidentification:{type:array, items:{type:string}}}

ethics:

type: object

properties: {intended_use:{type:array, items:{type:string}}, restrictions:{type:array, items:{type:string}}}

uncertainty:

type: object

properties:

components: {type: array}

combine: {type: string, enum: ["rss","linear","bayesian"]}

coverage: {type: string}

path_corrections:

type: object

properties:

models: {type: array, items:{type:string}}

parameters: {type: object}

uncertainty: {type: object}

see: {type: array, items:{type:string, pattern:"^[^:]+ v\\d+\\.\\d+:[A-Z].+$"}}

timebase_sync:

type: object

properties:

references: {type: array, items:{type:string}}

drift_ppb_max: {type: number}

switchover: {type: object, properties:{threshold_ppb:{type:number}, holdover_s:{type:number}}}

audits: {type: array, items:{type:string}}

distribution:

type: object

properties:

mirrors: {type: array, items:{type:string}}

shards: {type: array, items:{type: object, properties:{path:{type:string}, bytes:{type:integer}}}}

rate_limit: {type: object, properties:{mbps:{type:number}}}

VII. Coupling to Export Manifest (supplement)

export_manifest:

references:

- "EFT.WP.Core.DataSpec v1.0:EXPORT"

- "EFT.WP.Core.Equations v1.1:S20-1"

VIII. Chapter Compliance Checklist

• If any optional field is enabled, corresponding schema/regex and dependency anchors are complete and pass Lint.
• see[]/references[] use “Volume+Version+Anchor”; no shortcodes or missing versions.
• Any T_arr entries include delta_form, path, measure, and pass check_dim; path corrections and timebase sync cite the appropriate metrology entries.
• Math follows backticks/parentheses/no-Chinese rules; T_fil/T_trans and n/n_eff are not conflated.