backend

Introduction

The backend container calls important analytics functions. We have 3 sets of functions: Required functions need to run to generate specific variables (e.g. colors) or aggregate speech to text chunks. Optional functions also run automatically, but might be commented out in the backend source code. Switch functions can be set on or off in the settings.txt file. Please note the dependencies between some of the functions.

In detail:

• required_functions
  • functions for (iteratively) updating mongoDB collocations
  • specify coherent speaker colors based on md5 hash values
  • reformat speech to text input to generic data collocation
  • aggregate speech to text chunks to speakers’ turns
• optional_functions
  • calculate frequency of words per chunk
  • extract most frequent words
• switch_functions
  • calculate language complexity
  • calculate speech readability
  • annotate chunks using spacy
  • annotate chunks using targer argument mining

The backend module also merges the two files .env and settings.txt to a new file .Renviron. This file needs to be mounted to the analytics and visualization services. In case this file is missing, the modules can’t access the (hidden) settings; i.e. no environmental variables are defined. Forwarding the (hidden) settings is, however, very important as these settings define the required variables to access the database.

Technical Framework

Most functions are based on R; some functions do need Python.

Main Contributor

Felix Süttmann, Valentin Gold