feat: add czv, czv-wasm, and czv-python (init release)

czv-python/.gitignore

@@ -0,0 +1,72 @@
+/target
+
+# Byte-compiled / optimized / DLL files
+__pycache__/
+.pytest_cache/
+*.py[cod]
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+.venv/
+env/
+bin/
+build/
+develop-eggs/
+dist/
+eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+include/
+man/
+venv/
+*.egg-info/
+.installed.cfg
+*.egg
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+pip-selfcheck.json
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.coverage
+.cache
+nosetests.xml
+coverage.xml
+
+# Translations
+*.mo
+
+# Mr Developer
+.mr.developer.cfg
+.project
+.pydevproject
+
+# Rope
+.ropeproject
+
+# Django stuff:
+*.log
+*.pot
+
+.DS_Store
+
+# Sphinx documentation
+docs/_build/
+
+# PyCharm
+.idea/
+
+# VSCode
+.vscode/
+
+# Pyenv
+.python-version

czv-python/Cargo.toml

@@ -0,0 +1,16 @@
+[package]
+name = "czv-python"
+version = "0.0.0"
+authors = ["Mueez Khan"]
+description = "Python library for performing CSV-related functions for data engineering and analysis."
+repository = "https://github.com/rzmk/czv"
+edition = "2021"
+
+[lib]
+crate-type = ["cdylib", "rlib"]
+
+[dependencies]
+anyhow = "1.0.86"
+csv = "1.3.0"
+pyo3 = { version = "0.21.2", features = ["extension-module"] }
+thiserror = "1.0.61"

czv-python/README.md

@@ -0,0 +1,50 @@
+# czv-python
+
+Python library for [czv](https://github.com/rzmk/czv). czv is a library of utility functions for CSV-related data engineering and analysis tasks.
+
+## Installation and example
+
+```bash
+pip install czv
+```
+
+```python
+import czv
+
+data = """fruits,price
+apple,2.50
+banana,3.00
+strawberry,1.50"""
+
+output = czv.row_count(data, False)
+
+print(output)
+```
+
+## Development
+
+You'll need to have [maturin](https://github.com/PyO3/maturin/) and [uv](https://github.com/astral-sh/uv) installed. Set up a local virtual environment in the `czv-python` folder by running:
+
+```bash
+uv venv
+```
+
+Make sure to activate the virtual environment (instructions should be provided in your terminal after running the previous command).
+
+Once you've activated the virtual environment, install dependencies by running:
+
+```bash
+uv pip install -r requirements.txt
+```
+
+### Build package in local environment
+
+```bash
+maturin develop --uv --release
+```
+
+### Run tests
+
+```bash
+pytest
+```

czv-python/czv.pyi

@@ -0,0 +1,46 @@
+"""
+# czv
+
+Python library for [czv](https://github.com/rzmk/czv). CSV content manipulation and analysis.
+
+## Install
+
+```bash
+pip install czv
+```
+
+## Example
+
+```python
+from czv import row_count
+
+data = \"""fruits,price
+apple,2.50
+banana,3.00
+strawberry,1.50\"""
+
+output = row_count(data, False)
+
+print(output)
+```
+
+"""
+
+from typing import Optional
+
+def row_count(file_data: str, include_header_row: Optional[bool]) -> int:
+    """Returns a count of the total number of rows.
+    
+    ## Arguments
+     
+    * `file_data` - CSV file data.
+    * `include_header_row` - Specify whether to include the header row (first row) in the row count. Default is false.
+    """
+
+def column_count(file_data: str) -> int:
+    """Returns a count of the total number of columns (fields).
+    
+    ## Arguments
+    
+    * `file_data` - CSV file data.
+    """

czv-python/examples/simple_row_count.py

@@ -0,0 +1,10 @@
+import czv
+
+data = """fruits,price
+apple,2.50
+banana,3.00
+strawberry,1.50"""
+
+output = czv.row_count(data, False)
+
+print(output)

czv-python/pyproject.toml

@@ -0,0 +1,15 @@
+[build-system]
+requires = ["maturin>=1.6,<2.0"]
+build-backend = "maturin"
+
+[project]
+name = "czv"
+requires-python = ">=3.8"
+classifiers = [
+    "Programming Language :: Rust",
+    "Programming Language :: Python :: Implementation :: CPython",
+    "Programming Language :: Python :: Implementation :: PyPy",
+]
+dynamic = ["version"]
+[tool.maturin]
+features = ["pyo3/extension-module"]

czv-python/requirements.txt

@@ -0,0 +1,2 @@
+maturin
+pytest

czv-python/src/count.rs

@@ -0,0 +1,33 @@
+use crate::Result;
+use csv::ReaderBuilder;
+use pyo3::pyfunction;
+
+/// Returns a count of the total number of rows.
+///
+/// ## Arguments
+///
+/// * `file_data` - CSV file data.
+/// * `include_header_row` - Specify whether to include the header row (first row) in the row count.
+#[pyfunction]
+pub fn row_count(file_data: String, include_header_row: Option<bool>) -> Result<usize> {
+    let mut rdr = ReaderBuilder::new();
+
+    rdr.has_headers(!include_header_row.unwrap_or(false));
+    return Ok(rdr.from_reader(file_data.as_bytes()).records().count());
+}
+
+/// Returns a count of the total number of columns (fields).
+///
+/// ## Arguments
+///
+/// * `file_data` - CSV file data.
+#[pyfunction]
+pub fn column_count(file_data: Option<String>) -> Result<usize> {
+    let rdr = ReaderBuilder::new();
+
+    if let Some(file_data) = file_data {
+        return Ok(rdr.from_reader(file_data.as_bytes()).headers()?.len());
+    } else {
+        bail!("Could not determine a file path or file data for column_count_builder.");
+    }
+}

czv-python/src/lib.rs

@@ -0,0 +1,43 @@
+use pyo3::prelude::*;
+
+// Error-handling helpers
+#[derive(thiserror::Error, Debug)]
+#[error("{0}")]
+pub struct CzvError(anyhow::Error);
+
+impl From<pyo3::PyErr> for CzvError {
+    fn from(value: pyo3::PyErr) -> Self {
+        value.into()
+    }
+}
+
+impl From<csv::Error> for CzvError {
+    fn from(value: csv::Error) -> Self {
+        value.into()
+    }
+}
+
+impl From<CzvError> for pyo3::PyErr {
+    fn from(value: CzvError) -> Self {
+        value.into()
+    }
+}
+
+pub type Result<T> = anyhow::Result<T, CzvError>;
+
+#[allow(unused_macros)]
+macro_rules! bail {
+    ($err:expr $(,)?) => {
+        return Err(crate::CzvError(anyhow::anyhow!($err)))
+    };
+}
+
+// Command imports
+pub mod count;
+
+#[pymodule]
+fn czv(m: &Bound<'_, PyModule>) -> PyResult<()> {
+    m.add_function(wrap_pyfunction!(count::row_count, m)?)?;
+    m.add_function(wrap_pyfunction!(count::column_count, m)?)?;
+    Ok(())
+}

czv-python/tests/data/fruits.csv

@@ -0,0 +1,4 @@
+fruit,price
+apple,2.50
+banana,3.00
+strawberry,1.50

czv-python/tests/test_count.py

@@ -0,0 +1,24 @@
+import czv
+import pytest
+from .test_data import test_data
+
+class TestCountFunc:
+    @pytest.mark.parametrize(
+        "file_name,expected",
+        [("fruits.csv", 3), ("constituents_altnames.csv", 33971)],
+    )
+    def test_count(self, file_name, expected):
+        """Count the total number of non-header rows."""
+
+        result = czv.row_count(test_data[file_name].read_text())
+        assert result == expected
+
+    @pytest.mark.parametrize(
+        "file_name,expected",
+        [("fruits.csv", 4), ("constituents_altnames.csv", 33972)],
+    )
+    def test_include_header_row(self, file_name, expected):
+        """Count the total number of rows including the header row."""
+
+        result = czv.row_count(test_data[file_name].read_text(), include_header_row=True)
+        assert result == expected

czv-python/tests/test_data.py

@@ -0,0 +1,7 @@
+import pathlib
+
+data_path = pathlib.Path(__file__).parent.resolve().joinpath("data")
+test_data = {
+    file_name: data_path.joinpath(file_name)
+    for file_name in ["fruits.csv", "constituents_altnames.csv"]
+}