Nakladanie

Category model export

Category model export is the JSON export for Expert Categories. Use it when you need to study, compare, redesign, or share the structure of categories, category inheritance, direct attributes, effective templates, universal links, product naming rules, identity rules, and referenced dictionaries.

This export is designed for application administrators and expert taxonomy work. It is created through the normal Actualog export queue, stored in private export storage, and downloaded from Expert exports after the file is ready.

When to use it

Use category model export when you need to:

  • design universal categories;
  • compare concrete categories with universal category templates;
  • analyze where category attributes come from;
  • understand inherited effective templates;
  • review category tree context for selected or filtered categories;
  • send category model data to another expert, analyst, or AI-assisted design workflow;
  • troubleshoot why a product-level category has particular mandatory, optional, variant, measure, or unit rules.

Use CSV category export instead when you only need a spreadsheet list of category names, identifiers, type, workflow state, and summary counts.

Where to start

  1. Open Expert -> Categories.
  2. Apply filters if you need a specific category scope.
  3. Click Select items.
  4. Choose visible rows or Select all matching filters.
  5. Open Export and choose Category model JSON.
  6. Choose a preset and adjust the options if needed.
  7. Click Queue JSON export.
  8. Wait for the notification that the export completed.
  9. Open Expert -> Exports or use the notification link to download the file.

The export is queued. The page does not download one category at a time and it does not need to stay open while the worker builds the file.

Selection and filters

Category model export uses the same selection model as the Expert Categories list.

You can export:

  • selected visible rows;
  • the current visible page by selecting visible rows;
  • all categories matching the applied filters;
  • all matching filters minus visible rows you exclude after selecting the filtered scope.

Only applied filters are used. If you change filters but do not click Apply filters, Actualog treats the list as dirty and blocks all-matching-filter export until you apply the current filter state.

Filter metadata is written into the JSON file so the file explains how it was produced later. This can include search text, category type, level, workflow state, community, facets, sort order, selection mode, and exclusions.

Presets

The export dialog has three presets.

Lean

Use Lean when you want a compact file that is easier to read.

Lean includes:

  • core category identity and hierarchy fields;
  • direct attributes;
  • effective template rows.

Lean omits:

  • category tree context;
  • communities and facets;
  • sample product names;
  • product name templates;
  • identity policy;
  • referenced dictionaries;
  • effective-template hashes.

Lean is a good first choice when the full model feels too large or when you only need the category and template essentials.

Design

Use Design for most universal category design work. This is the default.

Design includes:

  • category tree context;
  • communities and facets;
  • universal links;
  • direct attributes;
  • effective templates;
  • product name templates;
  • identity policy;
  • referenced attributes, attribute groups, measures, and units.

Design omits:

  • sample product names;
  • internal effective-template hash fields.

Design is usually the best balance between context and readability.

Full

Use Full when you need deep troubleshooting evidence.

Full includes every optional section:

  • category tree context;
  • communities and facets;
  • sample product names;
  • universal links;
  • direct attributes;
  • effective templates;
  • product name templates;
  • identity policy;
  • referenced dictionaries;
  • full inheritance paths;
  • value filters;
  • inheritance analysis hints;
  • effective-template hash fields.

Full can be large. Use it when you explicitly need analysis details, not as the default handoff format.

Detail level

The Detail level control changes how much inheritance and template analysis is written.

Basic keeps row values and requirement metadata. It is easiest to read and does not include source lineage.

Standard adds useful source context: source category, source type, source distance, and override status.

Full analysis adds full source paths, value filters, inheritance hints, and optional template hashes when the hash switch is enabled.

Optional sections

You can turn individual sections on or off.

Category tree context adds ancestors, children, descendants, and linked universal categories around the exported targets. Tree nodes include isExportTarget so you can distinguish actual targets from context.

Communities and facets adds community and facet references for each category.

Universal links adds existing links from concrete categories to universal categories.

Sample product names adds a capped list of product names for context. It never exports full product payloads.

Product name templates adds raw product, variant group, and variant product naming templates plus parsed attribute token IDs.

Identity policy adds category identity rules used to distinguish products or variants.

Direct attributes adds attributes directly linked to each category.

Effective template adds the current materialized effective template for product-level categories.

Template hashes adds internal hash fields used for deep diagnostics. Leave this off unless you are investigating template generation or change detection.

Referenced dictionaries add only the attributes, groups, measures, and units used by the exported categories and templates. Actualog does not dump the whole library.

JSON structure

Category model JSON uses:

{
  "schemaVersion": "1.1",
  "documentType": "ActualogCategoryModelExport",
  "export": {},
  "selection": {},
  "filters": {},
  "modelOptions": {},
  "categories": [],
  "categoryTree": [],
  "attributes": [],
  "attributeGroups": [],
  "measures": [],
  "units": []
}

categories contains only the selected or filtered export targets.

categoryTree is context. It can contain ancestors, descendants, child categories, and linked universal categories that were not selected as export targets.

attributes, attributeGroups, measures, and units are reference dictionaries. They contain enough information to understand the category model without opening separate exports.

If an optional section is disabled, Actualog omits that JSON section instead of writing an empty full-library dump.

Requirements and legacy fields

Category model export writes requirement information in business terms such as:

  • Mandatory;
  • Recommended;
  • Optional.

It does not write retired legacy fields such as isRecommendedRequired.

Access and storage

Category model export is an Expert workspace export.

It appears in Expert -> Exports for the user who queued it. Application Administrators can also inspect live export files from Admin -> Integrations -> Exports. Category model export does not belong to the current company workspace and it does not count against company export quotas.

Generated files are stored in private export storage. Actualog does not expose public blob URLs. Downloads go through authenticated routes and require access to the export operation.

Deleted or expired exports cannot be downloaded.

Practical tips

Start with Design. Switch to Lean if the file is too large or you want to read it manually. Switch to Full only when you need diagnostic detail.

Turn off Sample product names unless product examples help the design task.

Turn off Template hashes unless an engineer or administrator asks for them.

Keep Referenced dictionaries on when sending the file to another tool or person, because dictionary sections make attributes, groups, measures, and units understandable without additional exports.

Use applied filters to narrow the file before exporting. A smaller, precise category model is usually more useful than a full platform-wide snapshot.