BQC is simple web application for checking the quality of BIM datasets, specifically IFC and COBie files. It supports IFC, COBie for the data integrity using the predefine rule, code, BCF, LLM(Large Language Model). It provides a user-friendly interface to upload these files, validate them for quality issues, and generate a comprehensive report.
BIM quality checker web application link

• Validate IFC files for compliance with BIM standards.
• Validate COBie files for data integrity and completeness.
• Customizable rules for data validation.
• Supports multiple file formats, including .xlsx, .csv, and .ifc.
• Built-in checks for:
  • Attribute ranges
  • Equations
  • Functional categories
  • Custom Python scripts
• Generate PDF reports summarizing validation results. This project will continue to be developed in the future.

• IFC 4.0 support.
• Error message and Bug fixed.

• BCF 2.0 support (without visualization).
• Bug fixed.

• Clash (collision) detection support. To test, use voids.ifc and walls.ifc in sample folder.

• Support to check BIM elements which have issues in 3D viewer like below

• Support LLM such as ChatGPT to check ruleset. It's still not perfect as it's an experimental feature. In detail, refer to LLM based quality check in Rule Condition chapter.

Add more features including LandXML, IFC various types, visualzation support.

To install source code and run it, follow the below and open http://127.0.0.1:7860

git clone https://github.com/mac999/BIM-quality-checker.git
pip install -r requirements.txt
cd ./src
python app.py

• 1. Upload BIM Check Ruleset JSON Configuration File
• 2. Upload BIM Files (COBie xlsx, csv. IFC. LandXML)
• 3. Click 'Run' to check BIM data quality
• 4. Download the Quality Report

The below is SAD(Software Architecture Design)(UML).

The tool uses a JSON configuration file to define the validation rules. Below are the main sections:

"project": {
  "name": "BIM quality check project (example)",
  "description": "This project is for BIM quality check example.",
  "version": "1.0",
  "author": "Taewook Kang",
  "email": "laputa99999@gmail.com"
}

Defines basic project information, including the name, description, version, author, and contact email.

The validation_rules consists of checks which have section defines multiple validation check for different BIM data types.

"validation_rules": [
{
  "name": "COBie data check",
  "entity": "Space",
  "file_format": [".xlsx", ".xls", ".csv"],
  "classification_system": "uniclass(2015), omniclass",
  "notify": "BIM-coordinator@gmail.com",
  "checks": [
    {
      "name": "Area Check",
      "description": "Validate the gross floor area of spaces",
      "attribute": "GrossArea",
      "condition": { ... }
    },
  ]
}
]

• entity: entity to check it's value. entity can be IFC element type (ex. IfcDoor), COBie worksheet's name (ex. Space) etc.
• file_format: checking files format.
• classification_system: information classification of the files. it's just explaination.
• notify: if there are problem in the checks, input the responsible person contact email.
• checks: consists of name, description and attribute for checking the value in properties. The entity has default properties category "Element" as default properties which have "Name", "OperationType", "Tag" in case of IFC.

{
  "name": "COBie data check",
  "entity": "Space",
  "file_format": [".xlsx", ".xls", ".csv"],
  "classification_system": "uniclass(2015), omniclass",
  "checks": [ ... ]
}

Validates COBie Space tables. Includes checks for:

• Gross floor area validation
• Functional category alignment
• Space classification via scripts

{
  "name": "IFC space data check",
  "entity": "IfcSpace",
  "file_format": [".ifc"],
  "classification_system": "uniclass(2015), omniclass",
  "checks": [ ... ]
}

Validates IfcSpace entities in IFC files. Includes checks for:

• Functional category validation
• Classification alignment

Supported Condition Types

Ensures a numeric attribute falls within a specific range, with optional tolerances.

{
  "type": "range",
  "min": 300,
  "max": 320,
  "tolerance_min": -2,
  "tolerance_max": 2,
  "units": "m2"
}

• min / max: Defines the acceptable range.
• tolerance_min / tolerance_max: Expands the range with tolerances.
• units: Specifies the unit of measurement (e.g., m2).

Validates an attribute using a custom mathematical condition.

{
  "type": "equation",
  "equation": "value > 100",
  "units": "m2"
}

• equation: A mathematical formula or logic (e.g., value > 100).
• units: Specifies the unit of measurement.

Clash (collision) detection between two elements IFC class in IFC files. In example,

"clash_check_file": "walls.ifc",
"condition": {
	"type": "collision",
	"option": "each_file",
	"IFC_entity": "IfcWall"
}

• type: clash detection type (only collision in 0.4 version).
• option: each_file for checking between one and another BIM file.
• IFC_entity: Specifies target IFC entity type of clash_check_file IFC file.

Validates that an attribute matches a predefined list of categories.

{
  "type": "list",
  "categories": [
    {
      "code": ".*SL_25_10_14.*",
      "name": "Classrooms"
    },
    {
      "code": ".*SL_20_15_59.*",
      "name": "Offices"
    }
  ]
}

• categories: A list of valid category codes and their names.
• code: A regex pattern matching the attribute value.
• name: A human-readable label for the category.

Here’s a complete example of a checks array with multiple condition types:

  "validation_rules": [
    {
      "name": "COBie Data Check",
      "COBie_table": "Space",
      "file_format": [".xlsx", ".xls", ".csv"],
      "checks": [
        {
          "name": "Area Check",
          "description": "Validate the gross floor area of spaces.",
          "attribute": "GrossArea",
          "condition": {
            "type": "range",
            "min": 300,
            "max": 320,
            "tolerance_min": -2,
            "tolerance_max": 2,
            "units": "m2"
          }
        }
      ]
    }
  ]

Custom scripts can be used for advanced validation. In the python code, you can use math, numpy as np, bim_rule_issues as rule_issues with the predefined variables like below.

• object_id: object id in case of IFC
• entity_name: object name
• check_name: check name
• check_cond: rule check's condition
• cat: cateogry
• attribute: attribute of value
• req_type: check value's type
• value: object's value which can be integer, float, list, dictionary and model like IFC
• guids: object GUID list
• check: check rule consists of condition. refer to the Ruleset Configuration File (JSON).
• self: condition object
• add_result_in_check: Function of condition object. Prototype = self.add_result_in_check(object_id, check, issue, passed, GUIDs=[], models=[])

Example:

{
  "condition": {
    "type": "code",
    "code": "print('code check')/nprint(value)/nself.add_result_in_check(object_id, check, f'{attribute} satisfies equation', False)"
  }
}

or

{
  "condition": {
    "type": "script",
    "script_file": "cobie_model_check.py"
  }
}

Support to check the dataset of models using LLM such as ChatGPT. To use, you need to enter your OpenAI api_key and enter the appropriate prompt. Note that depending on the size of the quality check target dataset, many tokens may be used. This is currently an experimental feature, and research will be conducted to reduce the number of tokens used effectively.

"checks": [			
  {
    "name": "Check Door Clearance Compliance Using LLM",
    "description": "Ensures that all doors meet the minimum clearance standards.",
    "attribute": "Element.Name",	
    "condition": {
      "type": "LLM",
      "model": "gpt-4o",
      "api_key": "", 
      "prompt": "Verify the value which has a clearance width of at least 900mm and height of 2100mm."
  }
}	
]

• model: model name such as gpt-4o
• api_key: your OpenAI API key
• prompt: input your prompt to check the data value of the entity in the check element. ex) "entity": "IfcDoor"

Develop by Taewook Kang (laputa99999@gmail.com) This project is licensed under the MIT License. See the LICENSE file for more details.

Welcome your contribution.

• UK BIM Alliance
• BCF-XML
• BCF-API
• BCF, Leica
• OpenCDE
• BCF viewer
• BCFViewer – BIM Collaboration Format tool

Thanks for the contributions.

• fastapi
• pydantic
• Gradio
• reportlab
• pandas
• ifcopenshell
• openpyxl
• fpdf