|
--- |
|
title: The PubMed Topic Modeling Toolkit |
|
emoji: ⚕ |
|
colorFrom: gray |
|
colorTo: gray |
|
sdk: streamlit |
|
sdk_version: 1.26.0 |
|
python_version: 3.11.4 |
|
app_file: 01_Introduction.py |
|
pinned: false |
|
--- |
|
|
|
Configuration reference: https://huggingface.co/docs/hub/spaces-config-reference |
|
|
|
# Supplementary Scripts |
|
|
|
The `PubMedTM-Scripts.zip` file is a directory of additional scripts and functions written during the development of the GUI tool. The file includes the scripts used to clean and process the data as well as experimental work on network analysis and lexical bundle extraction. Many of these scripts are experimental or incomplete, but they might be a good starting point if you want to analyze the data further. |
|
|
|
# Local Install |
|
|
|
For optimal performance, run this application on a computer with a performant CPU and at least 16GB of RAM. Using Dev Containers within Visual Studio Code (VS Code) is the easiest way to run the application locally. To get started, follow [this](https://code.visualstudio.com/docs/devcontainers/tutorial) tutorial until the "Check Installation" step. |
|
|
|
Next, download all the files in this repository to your computer, preserving the file structure in the process. To my knowledge, Hugging Face does not provide a way to do this via the GUI other than downloading each file individually. If you are comfortable with the terminal, you can clone the repository using [Git](https://git-scm.com/book/en/v2/Getting-Started-Installing-Git) and [Git-lfs](https://docs.github.com/en/repositories/working-with-files/managing-large-files/installing-git-large-file-storage?platform=windows). |
|
|
|
Once you've downloaded the files, open the folder containing the files in VS Code. You should be prompted to reopen the folder in a container. If not, press F1, search for, then select the option "Dev Containers: Reopen in Container." |
|
|
|
Wait for the container to build and install the requirements. This could take five to ten minutes depending on the internet connection and hardware. Once the container is ready, run the application by typing `streamlit run 01_Introduction.py` in the terminal at the bottom of VS Code. The application should open in your default browser automatically. |
|
|
|
### Troubleshooting |
|
|
|
- If the container fails to build, look for a part of the error message that says `"bind": bind source path does not exist:` then delete the `wayland-0` and `wayland-0.lock` files in the path specified by that portion of the error. On a default Ubuntu installation, the path is `/run/user/1000/wayland-0`. This advice is taken from [here](https://dev.to/crumdev/vscode-dev-container-failing-to-build-ap4). |
|
- If the app starts to run in the Dev Container but spontaneously kills itself, increase the maximum RAM and SWAP allocations in the Docker Desktop application settings. |
