A Python package for running an ensemble of sentiment analysis models and comparing their results. Wraps functions for text cleaning and the VADER, TextBlob, DistilBERT, and SentimentR sentiment analysis models, and provides a new function for plotting the results with various adjustments.
Created for SentimentArcs_WebApp and other uses.
Clone this GitHub repository, or download it as a .zip and unzip it. Use this console shell command to install the package:
$ python3 -m pip install /path/to/SentimentArcsPackage To reinstall after an update to the existing local copy of the package, run:
$ pip install --upgrade /path/to/SentimentArcsPackageImport and use within a python script, say my_script.py:
import imppkg as sa
def main():
with open("scollins_thehungergames.txt", "r") as file:
text = file.read()
title = "The Hunger Games"
clean_df = sa.preprocess_text(text)
distilbert_df = sa.compute_sentiments(clean_df, models=["distilbert"])
sa.download_df(distilbert_df, title, filename_suffix="_distilbert_raw_sentiments")
sentiment_results_df = sa.compute_sentiments(clean_df, title, models=["vader", "textblob", "sentimentr"])
smoothed_no_adjustments_df = sa.plot_sentiments(sentiment_results_df, title,
adjustments="none", plot = "save")
smoothed_zero_mean_df = sa.plot_sentiments(sentiment_results_df, title, models = ["vader", "textblob", "distilbert",
"sentimentr_jockers_rinker", "sentimentr_jockers", "sentimentr_huliu"],
plot = "display")
if __name__ == "__main__":
main()Then, in a console shell:
$ python3 /path/to/my_script.pyOr, import and use within an interactive python notebook through the interface of your choice (e.g., Google Colab, Jupyter Notebook) using the code in the main() function above.