Raw Versus Import Execution

Execution mode states where repository-controlled computation begins. It does not, by itself, establish vendor-native acquisition replay, chromatogram processing, external-engine parity, or scientific authority.

flowchart LR
    A["vendor or acquisition system"] --> X["exported or derived input"]
    X --> R["repository-controlled runtime lane"]
    R --> B["checked bundle and lineage"]
    B --> C["bounded runtime claim"]
    A -. "not implied by raw_executable" .-> R

Mode Contract

mode	repository guarantee	claim ceiling
import_only	the checked lane begins from imported exported-result evidence	no raw or external-engine rerun claim
raw_executable	the repository can execute its declared transformation from the checked input level	no automatic vendor-native, acquisition-native, or vendor-parity claim

The model field raw_rerun_supported distinguishes those two repository modes. On this page, the clearer reader-facing term is declared lane executable from checked input because the input may already be exported or derived.

Family Summary

workflow family	current run mode	declared lane executable from checked input	imported dependency count	blocked claim count
dda	import_only	no	3	2
dia	raw_executable	yes	5	2
lfq	raw_executable	yes	3	0
multiplex	raw_executable	yes	2	1
ptm	raw_executable	yes	2	0
targeted	raw_executable	yes	4	0

Family Boundaries

dda

• run contract: dda currently reruns through imported exported-result evidence instead of a raw in-repository execution lane.
• tracked imported dependencies: packages/bijux-proteomics-core/tests/fixtures/search_adapter_corpora/maxquant/maxquant_pipeline_export.tsv, packages/bijux-proteomics-core/tests/fixtures/search_adapter_corpora/maxquant/maxquant_evidence.tsv, packages/bijux-proteomics-core/tests/fixtures/search_adapter_corpora/maxquant/maxquant_settings.txt
• blocked claims: raw external-engine parity, vendor-native or engine-native reproducibility
• public claim ceiling: dda must not be described as raw-executable while dda-maxquant-pipeline-corpus still runs in import_only mode.

dia

• run contract: dia executes inside the runtime package, but its benchmark package can still include imported or derived evidence that does not prove vendor-native parity.
• tracked imported dependencies: packages/bijux-proteomics-core/benchmark-assets/flagship-public-packages/dia_library_review_package/primary/spectronaut_report.tsv, packages/bijux-proteomics-core/benchmark-assets/flagship-public-packages/dia_library_review_package/primary/spectronaut_pipeline_export.tsv, packages/bijux-proteomics-core/benchmark-assets/flagship-public-packages/dia_library_review_package/primary/spectronaut_settings.txt, packages/bijux-proteomics-core/benchmark-assets/flagship-public-packages/dia_library_review_package/comparator/diann_pipeline_export.tsv, packages/bijux-proteomics-core/benchmark-assets/flagship-public-packages/dia_library_review_package/comparator/diann_config.json
• blocked claims: chromatogram-native DIA authority, broad vendor-parity DIA replay
• public claim ceiling: DIA remains raw-executable in runtime terms, but the shipped package still stops short of chromatogram-native and vendor-parity claims.

lfq

• run contract: lfq executes inside the runtime package, but its benchmark package can still include imported or derived evidence that does not prove vendor-native parity.
• tracked imported dependencies: packages/bijux-proteomics-core/benchmark-assets/flagship-public-packages/lfq_cohort_review_package/evidence/study_scale_ms1_features.tsv, packages/bijux-proteomics-core/benchmark-assets/flagship-public-packages/lfq_cohort_review_package/evidence/study_scale.design.tsv, packages/bijux-proteomics-core/benchmark-assets/flagship-public-packages/lfq_cohort_review_package/evidence/quant_reproducibility_manifest.json
• blocked claims: none
• public claim ceiling: lfq should keep its stronger sentence behind the current benchmark package and downstream consequence limits.

multiplex

• run contract: multiplex executes inside the runtime package, but its benchmark package can still include imported or derived evidence that does not prove vendor-native parity.
• tracked imported dependencies: packages/bijux-proteomics-core/benchmark-assets/flagship-public-packages/multiplex_tmtpro_review_package/evidence/multiplex_ms1_features.tsv, packages/bijux-proteomics-core/benchmark-assets/flagship-public-packages/multiplex_tmtpro_review_package/evidence/multiplex.design.tsv
• blocked claims: outsider-auditable multiplex trust
• public claim ceiling: Multiplex may rerun in runtime terms while still failing the stronger outsider-facing claim boundary.

ptm

• run contract: ptm executes inside the runtime package, but its benchmark package can still include imported or derived evidence that does not prove vendor-native parity.
• tracked imported dependencies: packages/bijux-proteomics-core/benchmark-assets/flagship-public-packages/ptm_localization_review_package/evidence/localization_results.tsv, packages/bijux-proteomics-core/benchmark-assets/flagship-public-packages/ptm_localization_review_package/evidence/ptm_features.tsv
• blocked claims: none
• public claim ceiling: ptm should keep its stronger sentence behind the current benchmark package and downstream consequence limits.

targeted

• run contract: targeted executes inside the runtime package, but its benchmark package can still include imported or derived evidence that does not prove vendor-native parity.
• tracked imported dependencies: packages/bijux-proteomics-core/benchmark-assets/flagship-public-packages/targeted_transition_review_package/evidence/targeted_benchmark_qc.tsv, packages/bijux-proteomics-core/benchmark-assets/flagship-public-packages/targeted_transition_review_package/follow_up/supported_targeted_follow_up.json, packages/bijux-proteomics-core/benchmark-assets/flagship-public-packages/targeted_transition_review_package/follow_up/failed_targeted_transition_follow_up.json, packages/bijux-proteomics-core/benchmark-assets/flagship-public-packages/targeted_transition_review_package/follow_up/refused_targeted_follow_up.json
• blocked claims: none
• public claim ceiling: targeted should keep its stronger sentence behind the current benchmark package and downstream consequence limits.

Interpretation Discipline

A stronger execution mode changes the runtime statement only. Benchmark acceptance, grounding, recommendation, and lab consequence remain separately owned decisions. Imported dependencies stay visible even for raw_executable families so a repository rerun cannot be mistaken for vendor-parity replay.