The tdda package provides Python support for test-driven data analysis (1-page summary, blog, book).

• Reference Testing (tdda.referencetest): extensions to unittest and pytest for testing data analysis pipelines. Supports file-based comparisons, semantic equivalence, automatic rewriting of reference results, and test tagging.

• Automatic Test Generation (tdda gentest): generates reference tests for any command-line script or program (Python, R, shell, Makefile, ...). *"Gentest writes tests, so you don't have to."*™

• Constraints (tdda.constraints): discovers constraints from Pandas DataFrames, Parquet files, flat files, and relational databases; verifies new data against those constraints; detects failing records.

• Regular Expression Inference (tdda.rexpy): automatically infers regular expressions from a column of string data.

• Data Diff (tdda diff): compares data frames in Parquet or flat files and reports differences in a visual format.

• Serial Format (tdda.serial): documents CSV and flat-file formats in .serial metadata files for accurate, portable reading and writing. Supports conversion to/from CSVW and Frictionless metadata.

• Utility Functions (tdda.utils): Unicode normalization (Normal Form TK), glyph counting, and RFC 9839 support.

Full documentation: tdda.readthedocs.io

pip install tdda

To upgrade an existing installation:

pip install -U tdda

git clone https://github.com/tdda/tdda.git
cd tdda
pip install .

pip install pygresql                  # PostgreSQL
pip install mysql-connector-python   # MySQL/MariaDB
pip install pymongo                  # MongoDB

tdda test

• TDDA Blog
• Book
• Quick Reference Guide
• 1-page summary
• Full documentation
• PyCon UK talk (video)
• Mastodon

• Nick Radcliffe
• Simon Brown