IFC Model Analysis

Summarize an IFC BIM model: what storeys exist, what elements are on them, and what the model contains. This is the standard first pass before any deeper BIM task (clash review, quantity takeoff, model checking).

When to use

• User gives you an .ifc / .ifcxml file and asks "what's in this model?"
• Comparing two models (e.g. architect vs. structural model completeness)
• Preparing for clash detection or quantity extraction
• Checking whether required elements (walls, slabs, doors, windows) are present per storey

Workflow

1. Locate the file. If the user provides a path, use it. If not, ask for the path or find .ifc files in the working directory.

2. Prefer ifcopenshell (accurate, fast):

   pip install ifcopenshell

   import ifcopenshell
   
   model = ifcopenshell.open("model.ifc")
   
   # All entities by type (top 20 by count)
   from collections import Counter
   types = Counter()
   for entity in model:
       types[entity.is_a()] += 1
   for t, c in types.most_common(20):
       print(f"{t}: {c}")
   
   # Storeys and their elements
   storeys = model.by_type("IfcBuildingStorey")
   for storey in sorted(storeys, key=lambda s: s.Elevation or 0):
       print(f"\n{storey.Name or 'Unnamed storey'} (elev {storey.Elevation})")
       if storey.IsDecomposedBy:
           for rel in storey.IsDecomposedBy:
               for el in rel.RelatedObjects:
                   print(f"  - {el.is_a()}: {el.Name}")

3. Fallback (no dependencies): parse the STEP-style text. Lines look like #123=IfcWallStandardCase('...',...);. Extract entity type names with a regex like ^#\d+\s*=\s*([A-Za-z0-9_]+)\(, count them, and infer storeys from IfcBuildingStorey entities.
4. Produce the summary in this order:
   • File info: name, size, IFC version (from FILE_SCHEMA header)
   • Total entities, unique entity types
   • Building storeys (sorted by elevation) with element counts
   • Element inventory: top element types with counts
   • Completeness check: flag missing typical elements (walls, slabs, doors, windows) per storey
5. Report any parsing errors explicitly instead of guessing.

Output format

Use a compact report, for example:

IFC: office_model.ifc (IFC4, 8.2 MB)
Entities: 148,203 | Types: 47
Storeys: 5 (B1 -4.2, L1 0.0, L2 4.2, L3 8.4, Roof 11.6)
Elements: Wall 1,204 | Slab 402 | Column 186 | Beam 312 | Door 84 | Window 96
Completeness: B1 missing windows (parking), L1-L3 OK

Edge cases

• IFC2x3 vs IFC4: both supported; IfcWallStandardCase vs IfcWall naming differs — normalize by stripping the Ifc prefix and matching suffixes (Wall, Slab, Column, Beam, Door, Window, Roof, Stair, Ramp, Floor, Covering, Plate, Footing, Pile, Member).
• Huge models (100k+ entities): prefer ifcopenshell; the text fallback is fine for inventory but slow on large files — warn the user.
• Corrupt or partial files: if the file fails to parse, report the first error line and offer to continue with the entity-count-only pass.
• Unnamed storeys: use elevation or index as label, note it in the report.

Files

• references/ifc-quick-reference.md — entity classes and relationships worth knowing during analysis.
• scripts/analyze_ifc.py — dependency-free parser producing the report above (also reads .ifcxml).