2026-05-27
I have blogged in the past that I downloaded all of the Tax Court opinions and converted them to text files. A while ago, I created a Python app that searches the opinions. I have found it to be very useful. I decided to make the app open source.
If you aren’t familiar with Python, you can still install and run the app. On Windows, press the Start key, type PowerShell, and hit enter. Then type:
This runs an executable file from Astral, a well-respected company that was recently acquired by OpenAI. Close out of PowerShell, and then reopen it. Then type:
Next time to open the app, just open PowerShell and type tc-search. The app screen looks like:
As an example, type “Arrowsmith” into the Normal Text Search box. This returns 6 hits. Each hit shows 300 characters before "Arrowsmith" and 300 characters after "Arrowsmith." You can adjust the number of characters before and after the first search term. Each hit shows the date, opinion type, number of pages, docket number, judge, taxpayer's name, and search results. The second hit for Arrowsmith conveniently summarizes the holding:
In Arrowsmith v. Commissioner, 344 U.S. 6 (1952), the Supreme Court held that the characterization of a transaction may require examination of prior, related transactions.
If you want to open the original PDF file, navigate to https://dawson.ustaxcourt.gov, click on opinions, and type in the Docket number.
You can find the source code for the app on my GitHub page. The app is also on PyPI. If you are familiar with Python, you can “pip” install it with:
If you are on a Mac, in the Terminal, type:
The underlying data is saved as a parquet file on Hugging Face. The dataset can be viewed and manipulated via the Python Pandas library.
I hope others find this useful.