Skip to content

progress_roadmap.md

Latest commit

 

History

History
77 lines (59 loc) · 4.93 KB

progress_roadmap.md

File metadata and controls

77 lines (59 loc) · 4.93 KB
Raw

Progress & Roadmap

Legend: [ ] not started · [~] in progress · [x] done

Current phase

Validation / QA — core parser and takeoff pipeline are built, VCNGC live-drawing checks are running, and comparison tooling now exists to quantify gaps between raw takeoff output and the Task 1 ground-truth schedule.

Milestones

  • Set up storage architecture (log, description, roadmap, instructions-for-AI)
  • Define project description and scope
  • Market/gap research (DWG takeoff landscape, pricing, competitor focus)
  • Set up dev environment (Python, ezdxf, ODA File Converter)
  • Get a sample structural DWG/DXF drawing to test against
  • Build DXF parser: extract LINE/LWPOLYLINE entities grouped by layer
  • Compute lengths per entity, group/sum per layer
  • LLM layer-classification step: identify which layers are likely beams even with inconsistent naming
  • Output beam schedule to Excel/CSV
  • Test against a real/messy drawing (not just a clean sample)
  • Build takeoff cross-check / comparison utility against the Task 1 ground-truth schedule
  • [~] (stretch) Extract member sizes from block attributes/text (extractor implemented; source drawing has incomplete per-member annotations)
  • (stretch) Accept DWG directly, automate the DXF conversion step

Next up

Tackle the stretch goal: extract member tags, widths, and depths from block attributes/text so the comparator can perform true missing/extra member QA and dimension mismatch checks on VCNGC.

Known blockers / open questions

  • The current VCNGC takeoff export is aggregate-only (raw layer/entity totals), so it cannot yet identify individual missing/extra members or dimension mismatches.
  • Member-level tag/dimension extraction is implemented in parser_pipeline/member_size_extractor.py; the remaining gap is incomplete size annotation coverage in the source drawing, which leaves unresolved records for review.
  • Column ground truth is not yet modeled in the Task 1 schedule artifacts.

Phase 2: Automated BOQ System

Milestone 1: Database & Workspace Initialization

  • Create project workspaces and technical specs.
  • Define PostgreSQL relational database schema (boq_schema.sql).
  • Run database migration script in Supabase SQL Editor.
  • Seed initial Fajardo Concrete, Masonry, and Rebar mix factors.

Milestone 2: Input & Extraction Pipeline

  • Set up CAD conversion utility (DWG to DXF conversion script; local ODA integration is a shipping task).
  • Implement DXF parser using ezdxf to extract LINE, POLYLINE, TEXT, and schedule block attributes.
  • Map extracted entities into the Drawing Object Model (DOM) schema.

Milestone 3: Quantity Takeoff & Rule Engine

  • Develop Concrete Volume module (Fajardo Class A/B/C ratios).
  • Develop Steel Reinforcement module (weight conversion and GI tie wire rates).
  • Develop Masonry module (CHB wall areas and 16mm cement plaster).

Phase 3 engine QA verification — 2026-07-21

  • Added independent concrete volume, rebar-weight, and CHB area-versus-layer-count cross-checks with warnings above 2% divergence.
  • Applied distinct Masonry Mortar and Plaster Class A/B/C/D scales; these no longer derive from concrete mix factors.
  • Added opening deductions for CHB walls and expanded focused unit coverage to 13 passing tests.
  • [~] Unfinished — final QA sign-off: Retrieve and review the authoritative v2.0 technical specification from cloud storage, then reconcile each engine rule and factor against it. The local tech_spec.md is v1.0 and is not sufficient for v2.0 compliance sign-off.

Milestone 4: Costing & Consolidation

  • Integrate local unit cost database (materials + labor prices).
  • Generate detailed backup computations per drawing element.
  • Implement checklist rollout summary combining quantities and costs.

Milestone 5: Verification and Exports

  • Build Excel output module (openpyxl exporting Backup Computation, Checklist BOQ Summary, and Unit Cost Derivation sheets).
  • Implement manual validation/override workflow checks.

Phase 1 completion note — 2026-07-21

  • Added parser_pipeline/dom_mapper.py to map converted DXF entities into a traceable Drawing Object Model JSON export.
  • Verified against Structural_Drawings_Residential_House.dxf: 10,278 supported entities mapped; 8 parser tests passed.
  • Uploaded the Phase 1 mapper and DOM output to Supabase after explicit cloud-write approval.

Shipping / deployment notes

  • Configure the installed local ODA File Converter path in the shipped application and run DWG-to-DXF conversion automatically before parsing. This is a deployment task, not a Phase 1 parser blocker.
  • Provide a user-facing DWG import flow with a clear conversion-status message and DXF failure handling.
  • Keep PDF files available for drawing review and validation. Direct PDF entity extraction is deferred to a later phase; it is not part of the current CAD parser.