colss

colss is a lightweight expression evaluator that evaluates math-style string expressions on NumPy, Pandas, Polars, and standard Python arrays, reducing verbosity compared to native syntax

Source: https://github.com/SivaPA08/colss

eg: $\Sigma\left(e^{\sin(a)} + \log^2(b+1) - c^3\right)$

colss.sumof("exp(sin(a)) + log(b+1)^2 - c^3")

---

Installation

pip install colss

---

Requirements

All arrays must be 1D, preferably float64, and C-contiguous. If you have a 2D array:

a = a.ravel()

---

Usage

All functions accept a string expression. Just pass the expression — no need to register variables or pass arrays manually.

query

Evaluates an expression element-wise and returns a NumPy array.

import numpy as np
import colss
a = np.array([1.0, 2.0, 3.0], dtype=np.float64)
b = np.array([4.0, 5.0, 6.0], dtype=np.float64)
colss.query("a + b + 7")
colss.query("a > 1 ? 100 : 0")       # ternary
colss.query("sqrt(a) + sin(b)")
colss.query("a ^ 2")                  # a squared
colss.query("a ^ 2 + b ^ 0.5")       # complex expression

---

mean

Returns the arithmetic mean of the evaluated expression as a scalar.

colss.mean("a")
colss.mean("a + b")
colss.mean("a ^ 2 + b")

---

sumof

Returns the sum of the evaluated expression as a scalar.

colss.sumof("a")
colss.sumof("a * 2")
colss.sumof("a + b")

---

prod

Returns the product of all elements of the evaluated expression as a scalar.

colss.prod("a")
colss.prod("a + 1")
colss.prod("a * b")

---

median

Returns the median of the evaluated expression as a scalar. For even-length arrays, returns the average of the two middle values.

colss.median("a")
colss.median("a + b")
colss.median("a ^ 2 + b")

---

sd

Returns the population standard deviation of the evaluated expression as a scalar.

colss.sd("a")
colss.sd("a + b")
colss.sd("sqrt(a) + b ^ 2")

---

variance

Returns the population variance of the evaluated expression as a scalar.

colss.var("a")
colss.var("a + b")
colss.var("a ^ 2 - b")

---

NumPy Compatibility

colss.query() returns a NumPy array, so you can directly use NumPy array methods and operations like .mean(), .sum(), .reshape(), .astype(), .max(), .min(), .std(), .var(), and more on the result.

import numpy as np
import colss

a = np.array([1.0, 2.0, 3.0], dtype=np.float64)
b = np.array([4.0, 5.0, 6.0], dtype=np.float64)

colss.query("a + b").mean()
colss.query("a ^ 2").sum()
colss.query("sqrt(a)").reshape((3, 1))
colss.query("a * b").astype(np.float32)
colss.query("a + 5").max()
colss.query("a + b").min()
colss.query("a ^ 2").std()
colss.query("a ^ 2").var()

---

Using with Pandas

import pandas as pd
import numpy as np
import colss
df = pd.DataFrame({
    "a": [1.0, 2.0, 3.0],
    "b": [4.0, 5.0, 6.0]
})
a = df["a"].to_numpy(dtype=np.float64)
b = df["b"].to_numpy(dtype=np.float64)
df["c"] = colss.query("a + b + 7")
m       = colss.mean("a + b")
med     = colss.median("a")
s       = colss.sd("a + b")

---

Operators

Operator	Description
+ - * /	Arithmetic
^	Exponentiation — a ^ 2 raises a to the power of 2
%	Modulo
> < >= <= == !=	Comparison
and or not	Logical
condition ? a : b	Ternary

import numpy as np
import colss

a = np.array([1.0, 2.0, 3.0, 4.0], dtype=np.float64)
b = np.array([4.0, 3.0, 2.0, 1.0], dtype=np.float64)

# comparison
colss.query("a > 2")
colss.query("a <= b")
colss.query("a == b")

# logical
colss.query("(a > 1) and (b < 4)")
colss.query("(a > 2) or (b == 1)")
colss.query("not(a > 2)")

# ternary
colss.query("a > 2 ? a * 10 : a + 1")

# conditional
colss.query("if(a>b,a,b)")

---

Available Functions

abs(x)       sqrt(x)      
log(x)       log10(x)     exp(x)
sin(x)       cos(x)       tan(x)
floor(x)     ceil(x)
min(x, y)    max(x, y)

---

Notes

• ^ is exponentiation, not bitwise XOR.
• All arrays used in an expression must be the same length.
• You can use both scalar and vector at the same time.
• A NaN or Inf result raises an error identifying the first offending index.