Umut CANandClaude Opus 5
b03a7615b1
Tool count 131 -> 154; 474 -> 1242 collected tests. The release exists for two measured reasons. A 149-tool catalog costs a client ~40,300 tokens before it has asked for anything, and a drafter searching the way drafters think -- FILLET, BPOLY, QSELECT, WBLOCK -- scored df=0 against the stock index, because those words were simply not in the data. Discovery and batching - DISCOVERY_MODE=search replaces the catalog with search_tools + call_tool, ranked over an AutoCAD command and synonym corpus (discovery/aliases.py) on top of fastmcp's BM25 index. Measured: 40,305 -> 356 tokens advertised. - cad_batch runs a step list in one round trip; fields= projects result-heavy responses. The `core` profile is gone -- discovery replaced it. Honesty debt (Tier 0) - OCS/ECS: `extrusion` appeared nowhere in this codebase, so mirrored and rotated geometry reported coordinates that were quietly wrong. backends/ocs.py translates at the backend boundary; every tool coordinate is WCS on both engines, and a tilted plane refuses rather than answering with a number that looks fine. - drawing_save_as(fmt="dwg") wrote DXF bytes to a .dwg path. drawing_audit never repaired. analysis_measure_area shoelaced points the caller typed in. - ezdxf calls gained a per-call timeout, so one hung call can no longer wedge a server whose document lock is a single asyncio.Lock. Architecture - The 1854-line backend ABC is 19 contract modules; @capability(key, reason=) makes optional methods declarative and a gate fails on undeclared holes. - 8 layout/viewport tools, CHSPACE, CAD_PROGID escape hatch. - Wave A: selection filters, boundary tracing, hatch depth, annotation objects, measurement -- 13 tools of a planned 39. The REGION and 2D-boolean family was cut on evidence: add_region() yields zero ACIS bytes and the greiner_hormann substitute loses 28.2% of the area on a square with one semicircular edge. Verified against a live AutoCAD 2026, which found four COM defects no amount of reading would have caught -- AcadPViewport has no ViewCenter, and the line that set it had been swallowed by a bare except since v1.4. Benchmark v3 (15 tasks) and what it caught - v2 scored this server 10/10, which carried no information: every task in it exercised something the server was built around. Five tasks were added because they can fail, and three did while being written -- a hatch could not be measured though the refusal listed HATCH as measurable, the boundary tools and the measure tool disagreed about the same shape (139.2177 vs 139.2699), and a circle stored as two bulged vertices measured 0.0. - Competitor reports are NOT re-scored. They are pinned v1.4 runs against v2 and were never asked these five questions; the chart shows those rows `not run`. - Correctness A/B: 21/24 -> 24/24, zero regressions. Performance, stated plainly The published v1.4 perf report was recorded on CPython 3.14 and the new one on 3.11.15, so the apparent 2-6x speedup is the interpreter, not this release. Re-measured on one interpreter, v1.5.0 is 28% slower creating entities and 22% slower on the 10k roundtrip. Fully attributed: EZDXF_CALL_TIMEOUT=0 restores v1.4.0's exact numbers. That is the deadlock fix being paid for, the knob is documented, and winning it back is on the 1.6 roadmap. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
78 lines
3.4 KiB
Bash
78 lines
3.4 KiB
Bash
# AutoCAD MCP Pro — Environment Variables
|
|
|
|
# Rendering (view_screenshot / drawing_export_pdf) is an install-time choice,
|
|
# not an env var — the core install cannot render at all. ezdxf's drawing
|
|
# add-on imports Pillow unconditionally, so every render path needs it:
|
|
# pip install "autocad-mcp-pro[pdf]" ezdxf backend, matplotlib renderer
|
|
# (also pulls Pillow in transitively)
|
|
# pip install "autocad-mcp-pro[com]" live AutoCAD Win32 window capture
|
|
# pip install "autocad-mcp-pro[full]" both
|
|
# With only [com] the ezdxf backend still cannot render (no matplotlib).
|
|
|
|
# Backend selection: auto | com | ezdxf
|
|
AUTOCAD_MCP_BACKEND=auto
|
|
|
|
# COM ProgID the live backend attaches to. Escape hatch for AutoCAD-compatible
|
|
# applications that expose the same ActiveX object model.
|
|
#
|
|
# CAD_PROGID only changes which COM application the backend attaches to — every
|
|
# tool beyond the connection itself is developed and tested against AutoCAD
|
|
# alone, and AutoCAD-clone compatibility is unverified, so treat a non-default
|
|
# ProgID as experimental. An unrecognised ProgID fails loudly rather than
|
|
# falling back to AutoCAD, because the fallback would launch the very
|
|
# application you said you were not using.
|
|
#
|
|
# Candidates below are UNVERIFIED — none was tested against a running seat.
|
|
# List what is actually registered on your machine with:
|
|
# Get-ChildItem Registry::HKEY_CLASSES_ROOT |
|
|
# Where-Object { $_.PSChildName -like '*.Application*' } |
|
|
# Select-Object -Expand PSChildName
|
|
# CAD_PROGID=BricscadApp.AcadApplication # UNVERIFIED
|
|
# CAD_PROGID=ZWCAD.Application # UNVERIFIED
|
|
# CAD_PROGID=GstarCAD.Application # UNVERIFIED
|
|
CAD_PROGID=AutoCAD.Application
|
|
|
|
# Comma-separated list of allowed directories for file operations.
|
|
# Leave empty to allow all paths (not recommended for production).
|
|
# Example: ALLOWED_PATHS=C:\Drawings,D:\Projects\CAD
|
|
ALLOWED_PATHS=
|
|
|
|
# Maximum undo stack size (number of snapshots to keep)
|
|
MAX_UNDO_STACK=50
|
|
|
|
# Logging level: DEBUG | INFO | WARNING | ERROR
|
|
LOG_LEVEL=INFO
|
|
|
|
# Enable dangerous AutoCAD commands (ERASE ALL, QUIT, etc.)
|
|
# Set to true only if you trust all MCP clients
|
|
DANGEROUS_COMMANDS_ENABLED=false
|
|
|
|
# Exposed tool profile: full (default) | lean
|
|
# lean is a curated ~47-tool drafting/inspection set for clients with tight
|
|
# tool caps. The "core" profile was removed in v1.5.0 — the discovery layer
|
|
# below solves the crowded-surface problem it existed for. TOOL_PROFILE=core
|
|
# still boots, falling back to full with a warning.
|
|
TOOL_PROFILE=full
|
|
|
|
# Tool discovery: off (default) | search
|
|
# "search" replaces the advertised catalog with search_tools + call_tool, so a
|
|
# client sees two tools instead of ~126 and searches by AutoCAD command name
|
|
# ("QSELECT") or plain English ("round off this corner").
|
|
DISCOVERY_MODE=off
|
|
|
|
# Per-call backend timeouts in seconds; 0 disables the check.
|
|
# A hung call would otherwise hold the backend lock forever.
|
|
COM_CALL_TIMEOUT=60
|
|
EZDXF_CALL_TIMEOUT=120
|
|
|
|
# Undo/redo history for the headless backend, in steps. 0 disables it.
|
|
# Not free: ezdxf has no journal, so each step is a full DXF snapshot.
|
|
# Measured 37x slower entity creation with it on (0.18 -> 6.65 ms/call),
|
|
# and a snapshot of a 5000-entity drawing costs 130 ms / 658 KB. For a
|
|
# single checkpoint use transaction_begin/rollback instead.
|
|
# The live COM backend uses AutoCAD's own undo and ignores this.
|
|
EZDXF_UNDO_DEPTH=0
|
|
|
|
# Opt-in 3D solids (COM backend only; ezdxf reports them as unsupported)
|
|
ENABLE_3D=false
|