Troubleshooting & FAQ¶
FAQ¶
- Do I need hardware to try Thalamus?
No. You can install the package, run the pipeline, generate signals with a WAVE node, record, and analyze – all in software. The Examples are runnable with no acquisition hardware.
- What is a
.thafile? Thalamus’s capture/recording format: an append-only log of length-prefixed protobuf records. See Concepts and Architecture and convert it to HDF5/CSV/Parquet with the bundled tools.
- Which wheel do I download?
Pick the file on the Releases page that matches your OS and Python – e.g.
...manylinux...for Linux,...win_amd64...for Windows,...macosx...arm64...for Apple Silicon. The version in the docs is just an example; always grab the current release.- Do Windows and Linux differ?
The tools and node types are the same; only the virtual-environment activation command and some third-party device drivers differ.
Install & startup¶
Verify your install. After installing the wheel, check the import works:
python -c "import thalamus; print('thalamus OK')"
Then launch the pipeline; you should see a window with an empty node list (see Quick Start):
python -m thalamus.pipeline
grpc version conflicts. Thalamus bundles a specific grpc build. Installing
into a fresh virtual environment (as the Quick Start shows) avoids a mismatched
system grpcio that can cause import or connection errors.
No window appears. The pipeline and task controller are Qt GUI applications and need a display. On a headless machine or over SSH without X forwarding no window can open – run on a desktop session, or forward the display. If you only need data conversion/analysis, the command-line tools and examples do not require a display.
Running examples¶
The example scripts live in the examples/ folder of the
repository (they are not shipped inside the installed wheel). Clone the repo, or
copy a script, then run it from the repo root, e.g. python
examples/synthetic_recording.py.
Connecting tools to a pipeline¶
registry and other clients connect to a running pipeline over gRPC (default
localhost:50050). Start python -m thalamus.pipeline (or
thalamus.task_controller) first; if a client cannot connect, confirm the pipeline
is running and that the --address / --port match.