Digitizing a Spare Parts Catalog: From PDF to a Clickable Drawing in the Shop
Almost every spare parts dealer is sitting on catalogs: PDFs from the manufacturer, data packages from suppliers, lists from the ERP system. In the shop, they're only useful once they become clickable drawings with linked products. Which path gets you there depends on what you already have. This article goes through each path in turn, from the PDF catalog to the drawing you set up by hand.
What You Have Determines the Path
| Source material | Typically from | Path into the shop |
|---|---|---|
| PDF catalog with drawing and bill of materials | Manufacturer | PDF parser: frame, read, match |
| XML data with images as a ZIP | Supplier, importer | XML import: items, tree, and assemblies in one run |
| Spreadsheet from your inventory management system | own system | CSV import of the products, drawings afterward |
| Single drawing as an image | Engineering, scan | by hand: upload the image, place hotspots with a double-click |
Several paths often come together: the products arrive via CSV from the ERP system, then the drawings and hotspots come from the manufacturers' PDF catalogs.
Before You Start: Clarify Rights and Product Data
Two questions should be answered before the first import.
Are you allowed to show the drawings? Manufacturer catalogs and their drawings are, as a rule, protected by copyright. Many manufacturers explicitly make catalogs available to their dealers for selling, but what's binding is your agreement with the manufacturer or supplier.
How do the catalog and the shop find each other? Every position in the catalog carries a SKU, and it has to match a product in the shop. The matching runs via the SKU or the barcode. If the numbering systems don't line up, for example because the shop assigns its own numbers, it's worth adding the OEM number to a dedicated field beforehand.
The PDF Path: Framing the Drawing and the Bill of Materials
How the PDF parser works in detail is also shown on the importing a PDF spare parts catalog page. It opens the manufacturer catalog in the browser. The file stays on your own computer throughout; reading happens locally. Per assembly, it takes two steps:
- Frame the drawing. The frame determines which section will later be visible in the shop.
- Frame the bill of materials. The parser reads position numbers and SKUs from it and matches them to the products in the shop.
There are two ways to get the position numbers in the drawing, and the file decides which: if the numbers sit in the PDF's text layer, they're read directly. If the sheet is just an image, as with scanned catalogs, they're recognized from the image. A status line shows which method currently applies.
A few details that matter day to day:
- Bills of materials spanning several pages are merged together.
- A hotspot can carry several products, such as variants of the same part.
- A bill of materials sitting in the middle of the drawing is painted over in white on the image. Otherwise it would show up twice in the shop.
- Externally processed images: the cropped section can be downloaded, for example to cut it out or sharpen it, and played back in as the processed image. The hotspots still come from the PDF and can't end up misaligned.
- The destination in the catalog is chosen level by level, and you can also create a new one at any level.
Once the run is done, you know exactly how many assemblies, hotspots and linked products were created, and which numbers weren't found in the shop. This list is the basis for the next step: creating missing products or correcting numbers.
The XML Path: Supplier Data in One Run
Some suppliers and importers deliver their catalogs as a data package: a ZIP with an XML file per model and a folder of drawings. Techparts' XML import recognizes the manufacturer format on its own and shows in advance what's inside the package: models, assemblies, drawings, SKUs, and also what's missing, such as images or products with no hotspot in the drawing.
The import runs in two steps:
- Products. The SKUs are looked up in the shop. Missing ones can be created, optionally with the prices from the file and active right away. Matches that are found get remembered, which makes the next run fast.
- Catalog. Models and assemblies are created, drawings are uploaded and hotspots are placed, in rounds with a progress bar.
Shopify allows at most 128 different products per drawing. The import warns in advance which assemblies exceed that, and the rest still goes through. For a trial run, the app generates a sample ZIP with two drawn sheets.
The CSV Path: Products from the ERP System
Before drawings can be linked to products, the products need to exist in the shop. The CSV import brings them in from the ERP system:
- The delimiter is detected, and the columns are matched, with a suggestion based on the column names.
- The SKU decides whether a row creates a product or matches an existing one, not the title.
- Three modes: create and update, create only, update only.
- Empty cells are skipped. A run fills gaps instead of deleting data that's already maintained.
- Images via the file name:
1234.jpgbelongs to product 1234,1234_2.jpgis its second image.
The Manual Path
For individual drawings, for example from your own engineering, no import is needed. Upload the image, place hotspots with a double-click, assign products via Shopify's product picker. The editor can be operated entirely with the keyboard, and hotspots can be duplicated, moved and assigned several products.
Image Quality: What a Drawing Needs in the Shop
- Resolution: the smallest position number has to stay sharp when zoomed in. For scans of paper catalogs, 300 dpi is a good starting point.
- Cropping: just the drawing, without header, page number or bill of materials.
- Contrast: yellowed scans become easier to read if the background is lightened before importing.
Techparts warns in the editor when an image is borderline for sharp zooming in the shop.
After the Import: Checking What's Missing
No import is complete right away. That's why the overview counts what's still missing: hotspots with no assigned product and drawings with no image, and every number links straight to the list of affected entries. As long as nothing is published, the shop sees none of it: draft categories stay invisible, and publishing only writes what has actually changed. Measured on a real catalog, that took 8.8 seconds instead of 40.1.
How Long Does It Take?
Most of the time goes into creating products, not placing hotspots. Creating a product costs roughly fifteen times as much as looking one up. A catalog whose products are already in the shop is therefore taken over much faster than one where everything has to be created from scratch. Long runs proceed in rounds: if the connection drops, the work already done isn't lost, and a new click picks up right where it left off.
The sensible way to start is with a single model: import it, look at it in the shop, test it with real requests. Only then does the rest follow. What the catalog then looks like in the shop is shown in the article spare parts catalog in Shopify.
Frequently Asked Questions
My PDF has no text layer. Does this still work?
Yes. If the position numbers aren't in the text layer, the parser recognizes them from the image. The status line shows whether it's currently reading or recognizing.
Which XML formats are recognized?
The import recognizes the supported manufacturer formats on its own; otherwise the format can be chosen by hand. If your format isn't included, get in touch: a new manufacturer format can be added.
Are existing products overwritten during import?
For the CSV import, you control this via the mode, and empty cells never overwrite anything. The XML import looks up existing products by SKU and links them; it only creates missing ones if you choose to.
Which plan do I need?
Manual work is possible on every plan. CSV import belongs to PRO, PDF parser and XML import to MAX. Details are under Plans.