tlfs 0.0.4

Classes and scripts to create TLFs (Table, Listing, Figures shells) quickly and in an opinionated way.

tlfs

---

Table of Contents

• Installation
• License

Installation

pip install tlfs

License

tlfs is distributed under the terms of the MIT license.

Example usage

For some "standard" (opinionated) stuff you could do the following (which is what is done in the web version on streamlit):

tlf = TLF("Test")
tlf.from_xlsx("path-to-your.xlsx")
tlf
tlf.to_docx("path-to-your.docx")

For custom stuff one has to program by hand:

# overloading * Quant/Quali/Itemset to create Tables quickly
age = Quant("Age")
sex = Quali("Sex", groups = ["M", "F"])
trt = Quali("Treatment", groups = ["EXP", "CTRL"])
prices = Itemset("Unit costs",
                 items = ["Dentist", "Hospice", "Blood test"],
                 contents = ["unit cost", "per", "source"])
nation = Quali("Nation", groups = ["UK", "ITA"])

age * trt
[var * trt for var in [age, sex]] + [prices * nation]


# Changing display default in a table
age2 = Quant("Age", display = ["median", "25pct", "75pct"], unit = 'years')
age3 = Quant("Age", display = "median (iqr)", cell_content = "xx (xx - xx)")
sex2 = Quali("Sex", groups = ["M", "F"], display = "n", cell_content = "x")


univ = Section("Univariate tables", [Table(age), Table(sex)])
changed_def = Section("Some changed defaults", [Table(age2), age2 * trt, age3 * trt, Table(sex2)])
listings = Section("Listings", [Table(prices), prices * nation])
tlf2 = TLF("Table, Listings, Figure examples", [univ, biv, changed_def, listings])
tlf.to_doc("whatever.docx")

TODO

• overload + in Table to stack them
• overload + in Sections to concatenate them
• just do Section("Univariate tables", [age, sex]) instead of Section("Univariate tables", [Table(age), Table(sex)])