45-EFT.WP.Data.Pipeline v1.0 | Chapter 2 Terms & Dependencies

Home ／ Docs-Technical WhitePaper ／ 45-EFT.WP.Data.Pipeline v1.0

Chapter 2 Terms & Dependencies

I. Chapter Position & Coverage

• Position: Define terminology sources, the incremental glossary, unified symbols and dimensional baselines; fix cross-volume dependencies and anchor syntax, covering pipeline-specific terms such as pipeline/layers/edges/stage/operator/contract/schema/DQ gate/lineage/SLA/SLO/orchestrator.
• Coverage: Naming style & reserved keys, metrology & units posture, cross-volume dependency map, citation & anchor rules, path-quantity posture (e.g., T_arr), machine-readable dependencies and export manifest.

II. Terminology Sources & Version Pinning

• General terminology: Follow EFT Technical Whitepaper & Notes — Comprehensive Template v0.1; this volume defines incremental pipeline-specific terms only.
• Pinned minimal dependency set: EFT.WP.Core.DataSpec v1.0, EFT.WP.Core.Metrology v1.0, EFT.WP.Core.Equations v1.1, EFT.WP.Data.DatasetCards v1.0, EFT.WP.Data.ModelCards v1.0. If higher versions are required, declare them explicitly in export_manifest.references[].
• Clause-level citations: Cross-volume citations must include “Volume vX.Y: Chapter/Anchor”, preferably P/S/M/I anchors; whole-volume generic cites and aliases are forbidden.

III. Unified Symbols & Notation (Pipeline additions)

• Pipeline/topology: π (pipeline ID), S (set of stages), E (set of edges), Σ_in/Σ_out (input/output schema).
• Resources/perf: C_cpu (vCPU), C_mem (GiB), C_gpu (#/model), QPS (throughput), T_inf (per-sample latency, ms), ρ (utilization).
• Scheduling/reliability: τ_cron (Cron expression), p99 (99th pct), RTO/RPO.
• Privacy/security: k_anon (k-anonymity), ε_dp (DP budget).
• Math & symbol rules: Wrap inline symbols with backticks; any division/integral/composite operator must use parentheses and declare path gamma(ell) and measure d ell; no Chinese in formulas/symbols/definitions.

IV. Field Naming Style & Reserved Keys

• Style: Keys use snake_case; arrays denote plural entities.
• Reserved (must not be repurposed): pipeline.id, pipeline.version, layers[], edges[], resources, scheduling, quality_gates, export_manifest, metrology, monitoring, alert_rules, lineage.
• I/O contract keys: schema_ref, compat_mode, evolution_policy are reserved for contract management.

V. Cross-Volume Dependency Map (Minimal Set)

• Data contracts/exports: EFT.WP.Core.DataSpec v1.0 (file org, release, export_manifest).
• Metrology/units/error: EFT.WP.Core.Metrology v1.0 (units="SI", check_dim=true, error budget).
• Equations/path quantities: EFT.WP.Core.Equations v1.1 (Ch.2 S20-* arrival time T_arr and other path-dependent forms).
• Data facts & splits: EFT.WP.Data.DatasetCards v1.0 (provenance/sampling/splits/quality).
• Feature & I/O assumptions: EFT.WP.Data.ModelCards v1.0 (feature space, task I/O).

VI. Citation & Anchor Rules

• Fixed format: See "<Volume> vX.Y" Ch.<n> <Anchor>; anchor types: Pxx-? (postulates), Sxx-? (minimal equations), Mx-? (metrology workflows), Ixx-? (implementation bindings).
• Machine-readable references: Use "Volume vX.Y:Anchor" in both see[] and export_manifest.references[]; shortcodes/aliases/missing versions are prohibited.
• Examples: "EFT.WP.Core.DataSpec v1.0:EXPORT", "EFT.WP.Core.Metrology v1.0:check_dim", "EFT.WP.Core.Equations v1.1:S20-1".

VII. Arrival Time T_arr Posture (if applicable)

1. Two equivalent expressions:
   • T_arr = ( 1 / c_ref ) * ( ∫ n_eff d ell )
   • T_arr = ( ∫ ( n_eff / c_ref ) d ell )
2. Registration: Whenever the pipeline involves path quantities (e.g., TOA processing/correction), register delta_form, path="gamma(ell)", and measure="d ell" in the relevant stage config, and pass check_dim.

VIII. Machine-Readable Dependencies & Implementation Binding

• see[] example:
• see:
• - "EFT.WP.Core.DataSpec v1.0:EXPORT"
• - "EFT.WP.Core.Metrology v1.0:check_dim"
• - "EFT.WP.Core.Equations v1.1:S20-1"
• export_manifest excerpt:
• export_manifest:
• version: "v1.0"
• references:
• - "EFT.WP.Core.DataSpec v1.0:EXPORT"
• - "EFT.WP.Core.Metrology v1.0:check_dim"
• - "EFT.WP.Data.DatasetCards v1.0:Ch.11"
• artifacts:
• - {path:"pipeline.yaml", sha256:"..."}
• - {path:"dq/report.json", sha256:"..."}

IX. Metrology Baseline for Entities & Metrics

X. Chapter Compliance Checklist

• All cross-volume citations include “Volume vX.Y: Chapter/Anchor,” preferably P/S/M/I.
• snake_case naming is used; reserved keys are not repurposed; I/O contract fields use schema_ref/compat_mode/evolution_policy.
• Metrology baseline is units="SI" and check_dim=true; performance/resource units are explicit and consistent.
• If path quantities like T_arr are involved, delta_form/path/measure are registered and pass check_dim.
• see[] and export_manifest.references[] use the fixed format; no shortcodes/aliases/missing-version cites.