Open Stats
OpenStats is a library built on top of streamlit that extracts data from the Github API and shows the main KPIs for an OpenSource project:
- Star evolution: What is the popularity of the repo?
- Good First issues: Is there a focus towards the community?
- Recurrent collaborators: How many people are involved?
- Repository traffic: How many visits and clones do we have?
While there many other things to take into account, these metrics help us get a taste on how our project is doing in a single view.
If you'd like to see other metrics or graphics, open an issue or jump into the action!
- Requirements
- How does it work?
- Config
- Secrets
- Caching
- Publishing
- Contributing
- Acknowledgements
- License
Requirements
- Python 3.6+
- The main dependencies are
streamlit
andpandas
. The configuration is managed with Levy. - In terms of permissions, the traffic data requires an account (token) with
write
to the repo.
You can install OpenStats with:
$ pip install openstats
---> 100%
Successfully installed openstats
How does it work?
OpenStats is a helper tool to build an amazing dashboard from a config file. You can check an example here:
To run a streamlit
app, we need the following ingredients:
app.py
file that triggers the application.requirements.txt
, where we will just addopenstats
- Optionally, a
.streamlit/config.toml
config file with the theme.
By creating an openstats.yaml
file, we will pick up the necessary information, build the streamlit components and help you generate the theme file
This means that the final setup can look like this:
- An
app.py
with:from openstats.app import run if __name__ == "__main__": run()
- A
requirements.txt
file withopenstats
. - An
openstats.yaml
file following the examplesπ
Theme
To generate the theme file based on the config, you can run openstats-theme
after installing openstats
.
This will create the .streamlit/config.toml
file with the properties defined in our openstats.yaml
.
Config
Let's take a look at how to configure openstats.yaml
. You can see an example here:
title: "Dashboard title"
logo_file: "Image file to show at the sidebar"
client: # Information about the repository
root: "api.github.com" # We only support GitHub API
owner: "e.g., pmbrull"
repo: "e.g., OpenStats"
start_date: "Start counting stars from this date" # Format "Aug 1 2021" (`%b %d %Y`)
style: # To generate the streamlit theme
primary_color: "#7147E8" # Also used for the charts coloring
background_color: "#F9F8FD"
secondary_background_color: "#EEEAF8"
text_color: "#37352F"
font: "sans serif"
social: "
Free markdown text! Show your badges πͺ
"
Note that the style
section is only to centralise and generate the config.toml
file for streamlit
. The only added value here is that we will use the primary_color
for the theme and charts.
If you don't want to add any image to the sidebar, just remove the YAML entry.
More on streamlit themes
π blog
Minimum Config
The app can run with as minimum configuration as:
title: "Levy"
client:
root: "api.github.com" # We only support GitHub API
owner: "pmbrull"
repo: "levy"
start_date: "Aug 1 2021" # Format `%b %d %Y`
Secrets
To show the traffic data and to have a higher API query rate, we need to identify ourselves to the GitHub API.
OpenStats only supports authenticated requests. To make things work, there are two options:
- Prepare an
API_TOKEN
environment variable before running the app - Use streamlit secrets when publishing the app. The secret should also be named
API_TOKEN
.
The app will first try to obtain the token from the environment variables and will fall back to using streamlit secrets.
How to create an access token
π docs
Caching
Not all computations are lightning fast. In order to provide the best possible UX, we cache the API results using streamlit
memoization features. If you want to refresh the data, there is a clear cache
button available.
Publishing
You can create and manage your streamlit
apps at https://share.streamlit.io/. You can follow the docs for more information.
Contributing
Take a look at our CONTRIBUTING guide.
Acknowledgements
Thanks to streamlit for an amazing library and the GitHub API for sharing all the information!
License
OpenStats is released under Apache License, Version 2.0