pyteryx 0.1.0

Replicate Alteryx Designer tools as independent Python functions — migrate from Alteryx to pandas with a familiar API.

🐦 PyTeryx: The Alteryx-to-Python Migration Engine

Replicate Alteryx Designer tools as independent Python functions.

---

PyTeryx is a Python package that mirrors every major Alteryx Designer tool as a simple, independent function. If you're migrating from Alteryx to Python, PyTeryx gives you a familiar API while leveraging the full power, speed, and ecosystem of pandas.

Note: This package is not affiliated with Alteryx, Inc. "Alteryx" is a registered trademark of Alteryx, Inc. PyTeryx is an independent, community-driven open-source project.

🌟 Why PyTeryx?

Migrating from a visual ETL tool like Alteryx to code-first Python is traditionally painful. Data engineers and analysts have to translate complex logic, visual anchors, and proprietary tool configurations into standard Pandas operations.

PyTeryx bridges this gap by providing:

• Zero-Friction Migration: Translating an Alteryx workflow to Python becomes a straightforward 1:1 mapping exercise.
• Escape Vendor Lock-In: Execute your data pipelines anywhere Python runs—locally, on Airflow, AWS Lambda, or Databricks—without expensive proprietary licensing.
• Pure Python, Pure Pandas: Under the hood, PyTeryx leverages optimized pandas operations. No bloated dependencies, no proprietary file formats.
• CI/CD Ready: Since your workflows are now standard Python scripts or declarative YAML files, you can version control them, write unit tests, and integrate them into standard CI/CD pipelines.

✨ Core Architecture & Features

• 1:1 Alteryx Tool Mapping: Each core Alteryx tool has a corresponding Python function (e.g. Preparation.filter, Join.join, Transform.summarize).
• Familiar Output Anchors: Tools that have multiple output anchors in Alteryx return tuples in Python.
  • Filter returns (True_Data, False_Data)
  • Join returns (Left_Unjoined, Joined, Right_Unjoined)
  • Unique returns (Unique_Records, Duplicate_Records)
• Pure Functions: All methods are @staticmethod. There is no hidden instance state and no side effects.
• Immutable by Default: Original DataFrames are never mutated. Every tool returns a brand new DataFrame.
• Type-Hinted & Documented: Full type annotations and Google-style docstrings enable rich IDE autocomplete.

📦 Installation

Install PyTeryx via pip:

pip install pyteryx

(Requires Python 3.9+ and pandas 1.5.0+)

---

🚀 How to Use PyTeryx

PyTeryx is designed for both software engineers (Python API) and data analysts (Declarative YAML).

Option 1: Python API (For Developers)

Use PyTeryx exactly like you use pandas, but with the familiar Alteryx tool names and behaviors. You can chain these together natively.

from pyteryx import InOut, Preparation, Join, Transform, Parse, Developer

# 1. Read data (Alteryx: Input Data tool)
sales = InOut.input_data("sales.csv")
customers = InOut.input_data("customers.csv")

# 2. Filter rows (Alteryx: Filter tool → T/F anchors)
# Returns a tuple representing the True and False output anchors
high_revenue, low_revenue = Preparation.filter(sales, "Revenue > 1000")

# 3. Add a calculated column (Alteryx: Formula tool)
high_revenue = Preparation.formula(high_revenue, "Profit", "Revenue - Cost")

# 4. Join with another dataset (Alteryx: Join tool → L/J/R anchors)
# Returns (Left Unjoined, Joined, Right Unjoined)
left_only, joined, right_only = Join.join(high_revenue, customers, on="CustomerID")

# 5. Summarize (Alteryx: Summarize tool)
# Supports native Alteryx aggregation aliases!
summary = Transform.summarize(
    joined,
    group_by="Region",
    aggregations={
        "Profit": ["sum", "mean"], 
        "CustomerID": "count distinct"
    }
)

# 6. Verify data integrity (Alteryx: Test / Expect Equal tool)
Developer.test(summary, lambda df: df["Sum_Profit"].sum() > 0, "Profit must be positive!")

# 7. Output results (Alteryx: Output Data tool)
InOut.output_data(summary, "summary.parquet")

Option 2: YAML Pipelines (No-Code ETL)

For users who prefer a declarative approach, you can build end-to-end data pipelines without writing any Python code using PyTeryx's declarative YAML engine. This is perfect for defining pipelines as configuration files.

1. Create a pipeline.yaml

name: "Sales Filter Pipeline"
steps:
  - id: "load"
    tool: "InOut.input_data"
    args:
      path: "sales.csv"
  
  - id: "filter_high"
    tool: "Preparation.filter"
    inputs:
      df: "load"
    args:
      condition: "Revenue > 1000"
  
  - id: "save"
    tool: "InOut.output_data"
    inputs:
      df: "filter_high.0"  # The '.0' grabs the first tuple output (the 'True' anchor)
    args:
      path: "high_revenue.csv"

2. Run via Command Line Interface (CLI)

pyteryx run pipeline.yaml

---

📚 Complete Tool Reference

PyTeryx has fully audited and implemented the 6 core Alteryx data-manipulation palettes, giving you comprehensive parity for data transformation.

🔌 In/Out Palette

Read, write, browse, and generate data.

PyTeryx Method	Alteryx Tool	Description
InOut.input_data()	Input Data	Read CSV, Excel, JSON, Parquet (auto-detect format)
InOut.output_data()	Output Data	Write DataFrame to CSV, Excel, JSON, Parquet
InOut.text_input()	Text Input	Create DataFrame from inline dictionary/list data
InOut.browse()	Browse	Display rich summary statistics, types, and head to stdout
InOut.directory()	Directory	List files in a path with creation/modification metadata
InOut.date_time_now()	DateTime Now	Return the current timestamp as a DataFrame

🔧 Preparation Palette

Cleanse, filter, sort, and transform rows/columns.

PyTeryx Method	Alteryx Tool	Description
Preparation.filter()	Filter	Split data based on condition into (true_df, false_df)
Preparation.formula()	Formula	Add/update a column using a string expression
Preparation.select()	Select	Select, rename, or cast data types of columns
Preparation.data_cleansing()	Data Cleansing	Strip whitespace, modify case, remove punctuation/nulls
Preparation.sort()	Sort	Sort dataframe by one or multiple columns
Preparation.unique()	Unique	Split data into (unique_df, duplicate_df)
Preparation.sample()	Sample	Extract first N, last N, random N, or percentage
Preparation.record_id()	Record ID	Attach an auto-incrementing integer ID column
Preparation.generate_rows()	Generate Rows	Create sequential rows programmatically
Preparation.auto_field()	Auto Field	Optimize column data types to save memory
Preparation.multi_field_formula()	Multi-Field Formula	Apply a single formula across multiple columns
Preparation.multi_row_formula()	Multi-Row Formula	Apply formulas that reference previous/next rows
Preparation.tile()	Tile	Group data into quantiles or uniform bins
Preparation.imputation()	Imputation	Fill missing values (mean, median, mode, or custom value)
Preparation.create_samples()	Create Samples	Split data into estimation/validation/holdout sets
Preparation.date_filter()	Date Filter	Filter dataset by a specific date range
Preparation.oversample_field()	Oversample Field	Perform stratified sampling to balance a target class
Preparation.rank()	Rank	Assign numeric ranks (supports group-by ranking)

🔗 Join Palette

Blend and match multiple datasets together.

PyTeryx Method	Alteryx Tool	Description
Join.join()	Join	Standard join. Returns (Left_Unjoined, Joined, Right_Unjoined)
Join.join_multiple()	Join Multiple	Merge 3 or more DataFrames on a common key
Join.union()	Union	Stack DataFrames vertically (by column name or position)
Join.find_replace()	Find Replace	Lookup-based string replacement (or append lookup value)
Join.append_fields()	Append Fields	Cross (Cartesian) join to append all rows
Join.fuzzy_match()	Fuzzy Match	Approximate string matching (Levenshtein/Jaro-Winkler)
Join.make_group()	Make Group	Group relationship keys using DFS connected-components

📊 Transform Palette

Reshape, pivot, and aggregate data.

PyTeryx Method	Alteryx Tool	Description
Transform.summarize()	Summarize	GroupBy with named aggregations
Transform.transpose()	Transpose	Wide-to-long reshaping (unpivot)
Transform.cross_tab()	Cross Tab	Long-to-wide reshaping (pivot)
Transform.running_total()	Running Total	Cumulative sum calculation (supports grouping)
Transform.count_records()	Count Records	Output total row count as a single-value DataFrame
Transform.arrange()	Arrange	Manually transpose and rearrange multiple columns
Transform.make_columns()	Make Columns	Wrap sequential rows into multiple columns
Transform.weighted_average()	Weighted Average	Calculate weighted average (supports grouping)

Pro-Tip: Transform.summarize() and Transform.cross_tab() natively support familiar Alteryx aggregation aliases like "count distinct", "count null", "count blank", "concatenate", "longest", "shortest", and "mode" out-of-the-box (in addition to all standard pandas agg strings like "sum" and "mean"). Output columns are automatically named Agg_ColumnName (e.g. Sum_Profit), just like Alteryx.

📝 Parse Palette

Extract and parse dates, XML, and text.

PyTeryx Method	Alteryx Tool	Description
Parse.date_time()	DateTime	Convert strings to DateTime (or infer formats automatically)
Parse.regex_match()	RegEx (Match)	Create a boolean flag if a pattern is found
Parse.regex_parse()	RegEx (Parse)	Extract regex capture groups directly into new columns
Parse.regex_replace()	RegEx (Replace)	Replace text based on a regular expression
Parse.regex_tokenize()	RegEx (Tokenize)	Split a string by a delimiter into rows or columns
Parse.text_to_columns()	Text to Columns	Split delimited text into columns or rows
Parse.xml_parse()	XML Parse	Extract XML nodes, outer XML, and auto-flatten child tags

🛠️ Developer Palette

Advanced data manipulation utilities.

PyTeryx Method	Alteryx Tool	Description
Developer.base64_encode()	Base64 Encoder	Encode string columns to Base64
Developer.base64_decode()	Base64 Encoder	Decode Base64 columns to strings
Developer.download()	Download	Perform an HTTP GET request directly into a DataFrame
Developer.column_info()	Column Info	Generate a rich metadata/schema report of your data
Developer.dynamic_rename()	Dynamic Rename	Rename columns dynamically via a lookup table mapping
Developer.json_parse()	JSON Parse	Flatten a JSON string column dynamically into multiple columns
Developer.dynamic_select()	Dynamic Select	Subset columns dynamically by data type or regex pattern
Developer.test()	Test	Assert a condition evaluates to True on the dataset
Developer.test_equal()	Expect Equal	Strictly validate that two DataFrames are identical

(Note: PyTeryx intentionally skips Alteryx GUI workflow orchestrators like "Detour" and "Block Until Done", as well as external execution nodes like the "R Tool", since PyTeryx workflows natively leverage standard Python script execution flow).

---

🧪 Testing & Development

PyTeryx boasts an extensive test suite verifying 1:1 parity with Alteryx tools.

# Clone the repository
git clone https://github.com/pyteryx/pyteryx.git
cd pyteryx

# Install development dependencies
pip install -e ".[dev]"

# Run the test suite with coverage
pytest tests/ -v --cov=pyteryx --cov-report=term-missing

🤝 Contributing

Contributions are heavily encouraged! PyTeryx is community-driven. If you find a missing edge-case, want to optimize a pandas operation, or want to add support for a new Alteryx Marketplace tool, please open an issue or submit a pull request on GitHub!

📄 License

MIT License — see the LICENSE file for details.