CADnection “Smart Parsing” for Normalizing Measurements and Labels

In our last blog post we talked about CADnection and looked at the solution’s core capabilities comprised of extraction, data enhancement, and CAD model visualization. This post takes a closer look at the first of these capacities to better understand a key piece of the extraction capability referred to as “smart parsing”.  We’ll look at the technology and how it solves the issue of maintaining data’s context when extracting information from dissimilar CAD model formats.

If you haven’t already, our prior post on CADnection will help you get the most out of this post. You can read it here.

Learn More About CADnection

Today there are about 50 different CAD tools on the market, and each tool contributes unique capabilities and information. Some of these tools are simple 2D sketching applications typically used by individuals or students. On the other end of the spectrum are the more sophisticated 3D modeling solutions like Autodesk’s Inventor, Dassault’s Catia and SolidWorks, PTC’s Creo, and Siemen’s NX.

The resultant files produced by these various CAD applications are, in effect, mini proprietary databases. Like any database, it is comprised of so-called tag/value pairs. That is, an attribute name coupled with its value.  For example, the attribute name might be “Title” and the value is “Project ABC.”

So unlike text files that are primarily a string of words and/or numbers, CAD files inherently are filled with a wealth of information which is especially true for the high-end 3D modeling solutions. These files can contain characteristics defining the design structure, purchasing details, quality and inspection criteria, manufacturing techniques, and so on. There are hundreds if not thousands of tag value combinations within CAD models.

Accessing this data is not straight forward. Current techniques include exporting the desired data from within the CAD application, integrating the CAD application with a product data management (PDM) or product lifecycle management (PLM) solution, and even creating custom extraction tools. In other words, accessibility can be arduous, and context is often lost during the process.

On top of all that, CAD vendors have created their own tag/value dictionaries.  So, if you are a company that uses more than one CAD application, making sense of extracted data would require maintaining many mapping tables. These multi-CAD environments are very common for built-to-order (BTO), engineering-to-order (ETO), and tiered suppliers’ companies where their customers often dictate what CAD application they need to use.

As companies increasingly view their data as an asset, many have implemented, or plan to implement, search capabilities with solutions like SOLR, Elasticsearch, or search ecosystems like Lucidworks. How do you find or retrieve the needed CAD models based on their content, aside from maybe the folder name and file name?  You can still use keywords, but the richness of the tag/value properties is completely lost in the “proprietary databases” of each CAD model. Imaging searching for all CAD models that use part 123 version B with effectivity after a certain data. How do you do that? Or, as a global manufacturer, how can a user who’s search criteria in Imperial units present results that are in Metric units. Or, put plainly, how does someone accustomed to metric find a part weighing between 20 and 30lbs?

So, the challenge is not only getting access to all the data but delivering it in user-relevant formats. Here is where CADnection’s “smart parsing” helps. When CADnection extracts the CAD model data, the related tag/values are maintaining. This allows for dynamic normalization against a mapping dictionary to deliver contextual results in formats and units specific to the user preferences. These mapping dictionaries come pre-configured in CADnection but also allow for customization. Its worth noting the original tag/values remain and continue to be retrievable if users choose to view a model’s native values.   

In our next blog, we’ll be talking more about how smart parsing works and how it enables the enrichment process. We’ll also look at how data from other sources can be retrieved and fused with CAD model data for a richer user experience.