liwca 0.1.0a3

Linguistic Inquiry Word Count Assistant

liwca (Linguistic Inquiry Word Count Assistant) offers helper functions for working with LIWC dictionaries. Useful when you want end-to-end pipelines or notebook workflows that don't require the LIWC-22 app to be open, or when you just need reusable .dic[x] file I/O without writing it from scratch every project.

Features:

• Reading and writing dictionary files (.dic/.dicx)
• Merging dictionary files
• Fetching public LIWC-format dictionaries from remote repositories
• Pure-Python word counting (no LIWC-22 needed)
• Calling liwc-22-cli from Python

Installation

pip install --upgrade liwca

Quick Start

import liwca

# Fetch a public dictionary and count words (no LIWC-22 needed)
dx = liwca.fetch_threat()
results = liwca.count(["danger lurks ahead"], dx)

Usage

Fetching dictionaries

import liwca

dx = liwca.fetch_sleep()           # Fetch and load a public dictionary
dx = liwca.fetch_bigtwo()          # Versioned dictionary (version="a" by default)
dx = liwca.read_dx("./my.dicx")    # Read a local dictionary file
liwca.write_dx(dx, "./my.dic")     # Write to a different format

Word counting

Pure-Python word counting using LIWC-style dictionaries (no LIWC-22 needed).

texts = ["I feel happy today", "This is a sad story"]
results = liwca.count(texts, dx)                      # percentages (default)
results = liwca.count(texts, dx, as_percentage=False)  # raw counts

LIWC-22 wrapper (requires LIWC-22)

The LIWC-22 desktop application (or its license server) must be running when you call the CLI. See the LIWC CLI documentation and Python CLI example for more details.

import pandas as pd

liwc = liwca.liwc22.Liwc22(encoding="utf-8", count_urls=True)
df = pd.DataFrame({"doc_id": ["a", "b"], "text": ["...", "..."]})
out_path = liwc.wc(
    input=df,                      # DataFrame or path
    output="results.csv",
    row_id_indices=["doc_id"],
    include_categories=["anger", "joy"],
)
results = pd.read_csv(out_path, index_col=0)

Similar Projects

• liwc-python
• lingmatch
• sentibank