Skip to content

docs: set LOAD_ENTERPRISE_BLOCKS=TRUE in docs workflow build step#2386

Open
rvirani1 wants to merge 2 commits into
mainfrom
riaz/ent-1295-fix-enterprise-block-documentation-404-errors
Open

docs: set LOAD_ENTERPRISE_BLOCKS=TRUE in docs workflow build step#2386
rvirani1 wants to merge 2 commits into
mainfrom
riaz/ent-1295-fix-enterprise-block-documentation-404-errors

Conversation

@rvirani1
Copy link
Copy Markdown
Contributor

@rvirani1 rvirani1 commented May 29, 2026

Summary

Fixes Bug 1 from ENT-1295 (enterprise block doc pages never generated) in two parts:

  1. .github/workflows/docs.yml: the docs CI checks out and pip installs the private workflows-enterprise-blocks repo, but the "Build block docs" step never set LOAD_ENTERPRISE_BLOCKS, so blocks_loader.py silently dropped every enterprise block and no page was generated (links 404). Now sets LOAD_ENTERPRISE_BLOCKS: "TRUE" on that step.
  2. development/docs/build_block_docs.py: with enterprise blocks now loaded, get_source_link_for_block_class only handled inference/core/workflows/, so enterprise pages got href="None". It now resolves the path relative to the inference package root, producing valid links for both core (inference/core/workflows/…) and enterprise (inference/enterprise/workflows/…) blocks. Core-block links are unchanged.

After merge, trigger the workflow with deploy: true (or cut a release) to regenerate and publish. Bug 2 (frontend slugifier mismatch for OPC UA and 6 other names) is handled separately in the roboflow repo.

Test plan

  • Re-run docs workflow with deploy: true; confirm opcua_writer_sink, mqtt_writer, event_writer, plc_modbus_tcp pages resolve and their "Source" links point to valid GitHub URLs (not None).

Without this flag, the block-enumeration loader silently drops every
enterprise block, so the docs CI never generates pages for them and
links from the frontend 404. The workflow already checks out and
installs the private workflows-enterprise-blocks repo; this just makes
the build step actually load them.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
get_source_link_for_block_class only handled inference/core/workflows/,
so enabling enterprise blocks produced pages with href="None" (the
split raised IndexError and the link fell back to None). Resolve the
path relative to the inference package root instead, so both core
(inference/core/workflows/...) and enterprise
(inference/enterprise/workflows/...) blocks get valid links. Core-block
links are unchanged.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant