Convert 17-0574 (rev 2): read Designer WorkFlow exports, re-pair to Forms_17-0574 (AEM to Quad).xml, re-pin metrics

[devin-ai-integration]

Summary

Second convert-form pass for 17-0574 with a new reference export, Forms_17-0574 (AEM to Quad).xml, replacing 17-0574.xml. The form is still not converted and not ready: G3 and G4 fail, G5/G6 are external, all three decisions remain proposed.

Input note: the session was again launched as form_id: 17-2574; the XDP is 170574.xdp and pairs.yaml pairs it as 17-0574, so this PR is for 17-0574. Nothing was renamed (the target file keeps its supplied name, spaces and parentheses included).

What changed and why:

• The new export is a full Designer WorkFlow, not a bare layout. Root is <WorkFlow version="17.0.436.3"> with Property/Connect/Output/ExtWfdImpl children and one Layout module whose nested <Layout> holds the record list the inventory expects. target_inventory.inventory() iterated the root directly, so G4 saw 0 controls/variables and G5 lint produced 200 bogus IMP-003. Reusable fix (rule 8, not a per-form patch):

  def layout_root(root: ET.Element) -> ET.Element:
      Layout   -> root
      WorkFlow -> the single Layout module's single nested <Layout>; ValueError if 0 or >1
      other    -> ValueError("unsupported target export root <...>")
  inventory(...)["export_kind"] = "layout" | "workflow"   (+ "workflow_version" for WorkFlow)

  Tests: WorkFlow-wrapped records inventory identically to the bare <Layout>; multiple modules and unknown roots raise.
• src/fixtures/pairs.yaml: 17-0574 layout_xml re-paired to the new file (sha256 7202b804…, 4,089,931 B); note records that the earlier 17-0574.xml was a different revision and is superseded.
• src/decisions/17-0574.yaml: no new findings; D-001 rationale now cites both exports (the new one also has one *_BarcodeFlowArea_* per page and no NB6001 variable). Still 3 proposed, decided_by/decided_on null.
• knowledge/quadient-designer.md: the "full WFD vs partial layout export" question is now answered by observation (both variants exist and are read).
• tests/test_golden.py 17-0574 re-pins: G0 fail→pass; field match 0.5041→0.6341; exact-name 0.122→0.0894; static text 0.7707→0.9608; target controls 271→74; data variables 314→2494. 18-1026 untouched.
• No raw artifacts committed (may_commit_artifacts: no); hashes verified.

Readiness (out/readiness.json, 17-0574)

gate	verdict	key metrics
G0	pass	static-text coverage 0.9608 (threshold 0.95), 16 target-only tokens; hashes match pairs.yaml
G1	pass	41 containers / 123 fields / 71 statics
G2	pass	all source objects accounted; one minor G2-UNKNOWNSTATIC
G3	fail	123 in-scope fields, 120 bound / 3 unbound (97.6%), name evidence 100%; findings routable 3, unrouted 0, proposed 3, accepted 0, rejected 0; 3 nodes need human touch; 6 G3-DRAFT-RULE (major)
G4	fail	field match 0.6341 (floor 0.70), exact-name 0.0894, sections 0/13; target has 74 FormControls / 2494 data variables; 45 G4-UNMATCHED
G5	external	offline lint clean (IMP-001..005 all 0 on the new export); no Designer import log
G6	external	reference PDF inspected (9 pages, Designer 17.0.436.3); no composition run, no input payload, no AEM PDF

Routed findings by action: manual_design 3 (G3-UNBOUND x3, the footer PDF417 barcodes). No G3-STALE-DECISION. Minor, not routable: G3-DUPNAME x25.

Scripts (classified, never executed): 48 total, 100% mechanically transferable — mutual_exclusion 19, visibility 16, conditional_clear 10, uppercase 3. None manual.

Draft rules the form depends on: BND-04, FLD-01, OBJ-02, OBJ-05, OBJ-08, OBJ-12.

Why G4 is still under the floor with a matching revision: 45/123 fields find no target by name; the export carries 2,494 data variables (most under module/system groups) against 74 FormControls, and exact-name agreement is 9%, so the dictionary (BND-05) rather than fuzzy matching has to close the gap. No thresholds changed.

People / external gates

• Decision reviewer: none supplied — nothing accepted or rejected.
• Designer importer: none supplied — G5 stays external. Handoff package (handoff/17-0574/: plan sha256 66bc8e7c… unchanged, CimBuildPlanImport.groovy, filtered decisions, CHECKLIST.md with the migration-stack v17.0.30 deploy sequence) and out/readiness.md are attached to the session; handoff/ is gitignored.
• G6 external: no fixed input payload or AEM PDF.

Checks

ruff check ., ruff format --check . clean; pytest -q 96 passed / 7 skipped (18-1721 artifacts absent); python -m tools.pipeline -o out/ exit 0.

Clearance note: derived metrics and SOM paths from a partner-built test form, consistent with existing records; clearance status per knowledge/customer-clearance-boundaries remains TBD.

Link to Devin session: https://partner-workshops.devinenterprise.com/sessions/060b9626e9ad430186f22f341a34278b
Open in Devin Desktop: https://partner-workshops.devinenterprise.com/desktop/session/060b9626e9ad430186f22f341a34278b?variant=devin
Requested by: @sumitshatwara

[devin-ai-integration]

🤖 Devin AI Engineer

I'll be helping with this pull request! Here's what you should know:

✅ I will automatically:

• Address comments on this PR. Add '(aside)' to your comment to have me ignore it.
• Look at CI failures and help fix them

⚙️ Control Options:

• Disable automatic comment, CI, and merge conflict monitoring