Providing specklepy docs feedback

JoostGevaert · 25 November 2025 17:59

Hi ,

I’m giving the specklepy SDK a go now that it has some nice docs. I’d be happy to provide some feedback, and create PRs to make the docs better as I go.
What’s the best way to do this?
I don’t really get in what repo I can find the the docs, and there’s no “Edit this page” or “Create issue” buttons in the docs.

So far I’ve come across:

Docs on QGIS Objects, which don’t seem relevant anymore since you’ve terminated the QGIS plugin
Code that doesn’t work: info.workspaces.workspacesEnabled
I’m assuming this attribute doesn’t exist anymore now that Speckle servers always have workspaces?

jonathon · 25 November 2025 18:03

fair. GitHub - specklesystems/speckle-docs-NEW

Although

github.com/specklesystems/specklepy

src/specklepy/core/api/models/current.py

2f2e8ba73


      
              name: str
              url: str
          
          
          class ServerConfiguration(GraphQLBaseModel):
              blob_size_limit_bytes: int
              object_multipart_upload_size_limit_bytes: int
              object_size_limit_bytes: int
          
          
          class ServerWorkspacesInfo(GraphQLBaseModel):
              workspaces_enabled: bool
          
          
          # Keeping this one all Optionals at the minute,
          #  because its used both as a deserialization model for GQL and Account Management
          class ServerInfo(GraphQLBaseModel):
              name: str | None = None
              company: str | None = None
              url: str | None = None
              admin_contact: str | None = None

yes this should be snake cased.

JoostGevaert · 25 November 2025 20:56

Ah, yea, great!

Next question:
Why does class ServerTransport (still) have a stream_id attribute instead of a project_id attribute?

github.com/specklesystems/specklepy

src/specklepy/transports/server/server.py

2f2e8ba73


      
              objectId = hash,
              modelId = model_id,
              projectId = project_id,
          )
          version = client.version.create(input)
          ```
          """
          
          def __init__(
              self,
              stream_id: str,
              client: Optional[SpeckleClient] = None,
              account: Optional[Account] = None,
              token: Optional[str] = None,
              url: Optional[str] = None,
              name: str = "RemoteTransport",
          ) -> None:
              super().__init__()
              if client is None and account is None and token is None and url is None:
                  raise SpeckleException(
                      "You must provide either a client or a token and url to construct a"

jonathon · 25 November 2025 21:27

Because if there weren’t crumbs of history like this, then archaeologists would be out of a job.

Essentially, because reasons but

This is there so noone asks this question

JoostGevaert · 26 November 2025 13:28

Love the answer hahaha

Why is there a specklepy.core.api as well as a specklepy.api module?

Jedd · 26 November 2025 15:44

The specklepy.api classes are wrappers around specklepy.core.api and should provide the same functionality.

The specklepy.api functions are designed to be used by third-party developers such as your self, and the specklepy.core.api are used by our first party applications (like Blender connector, Ifc Importer, and Speckle Automate).

The split exists to allow us to track different usage metrics; and differentiate usage of specklepy by third party developers from usage of specklepy by first party products.

In either case, the metrics track are lightweight usage analytics related to integration with speckle servers.
It allows us to improve Speckle’s service; e.g. measure impact of breaking changes, prioritize developer user needs, and understand the performance of our tech stack.

Any data is collected as per our privacy and data processing agreements

JoostGevaert · 26 November 2025 15:46

Oh wauw, interesting!