Tax Authorities
A small desktop GUI for searching the full text of U.S. Tax Court opinions, IRS revenue rulings (also covering Notices, Announcements, and Revenue Procedures), and IRS private letter rulings / written determinations — no coding required.
Features
- Search across all three datasets at once, or toggle any combination off with checkboxes
- Plain-text search supports multiple terms (AND logic),
"quoted phrases"for exact word sequences, and a-termprefix to exclude documents containing that term - Full regex pattern search as an alternative to plain text
- Matching is case-insensitive
- Choose how many characters of surrounding context to show before/after each match
- Results are written to a text file that opens automatically when the search finishes
Installation
powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"
uv tool install tax-authorities
tax-auth
The first command installs uv, a fast Python package/tool manager. Already have Python set up your own way? pip install tax-authorities works too, as long as tax-auth ends up on your PATH.
The first time tax-auth runs, it automatically downloads all three datasets from Hugging Face. That download only happens once — new documents are added to the datasets periodically, but tax-auth won't pick them up on its own. Run tax-authorities download-data any time to pull the latest versions of all three.
Usage
Fill in exactly one of the two search fields, then press Enter or click Search:
- Search Term(s) — one or more words to match, without regex syntax:
- Every term must be present in a document for it to match (AND logic).
- Wrap multiple words in quotes (e.g.
"innocent spouse") to require that exact sequence. - Prefix a term with
-(e.g.-partnership) to exclude documents containing it.
- Regex Pattern — a standard Python regular expression, matched against each document's text.
Use the checkboxes to choose which of the three datasets to search: Tax Court Opinions, Revenue Rulings, and Private Letter Rulings. Use the Number of characters before/after fields to control how much surrounding context is shown around each match (300 characters on each side by default).
When the search finishes, the results file opens automatically. It's also saved under your user data folder if you want to find it again later.
Data
Three datasets, each hosted on Hugging Face:
- Tax Court Opinions — andrew-mitchel/tax-court-opinions — 14,830 opinions, 1997 to the present.
- Revenue Rulings — andrew-mitchel/revenue-rulings — 3,475 documents, also including Notices, Announcements, and Revenue Procedures, 1998 to the present.
- Private Letter Rulings — andrew-mitchel/private-letter-rulings — 45,348 IRS written determinations (PLRs, CCAs, TAMs, etc.), 1999 to the present.
These documents were originally published by the Tax Court and the IRS as PDFs, then converted to text and compiled into parquet files.
License
Metadata
Release files for tax-authorities 0.1.1
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | |
|---|---|---|---|---|
| tax_authorities-0.1.1.tar.gz | 62.3 kB | Without provenance | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| tax_authorities-0.1.1-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 70.5 kB
Security for tax-authorities 0.1.1
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.