qante 0.0.4

qante - Query ANnotated TExt

Motivation

Extracting the highly-valuable data from unstructured text often results in hard-to-read, brittle, difficult-to-maintain code. The problem is that using regular expressions directly embedded in the program control flow does not provide the best level of abstraction. We propose a query language (based on the tuple relational calculus) that facilitates data extraction. Developers can explicitly express their intent declaratively, making their code much easier to write, read, and maintain.

Solution

This package allows programmers to express what they are searching for by using higher-level concepts to express their query as tags, locations, and expressions on location relations. The location of a string of characters within the document is the interval defining its starting and ending position. Locations are grouped into sets named by tags. Tags can be used in conjunctions and disjunctions of interval relations to query for tuples of locations.

We invite you to view a YouTube Video of our talk on PyData Global 2022:

where we showed this material from our GitHub repo:

∙ pydataG22.pdf slides of our talk.

∙ ipynb/pydata.ipynb a jupyter notebook with examples.

Use pip or python (rev 3 or above) to install from PyPI:

pip install qante (or python -m pip install qante)

Use python docstrings for API Documentation:

python    # rev 3 or above
  from quante.tagger import Tagger
  from quante.query import Query
  from quante import loctuple as lt
  from quant.table import get_table

  help(Tagger)    # annotate text with tags using tagRE('tagname', regexp)
  help(Query)     # Syntax for querying annotated text
  help(lt)        # Predicates used by queries
  help(get_table) # extract table (as dictionaries) from text

See also: “API Documentation” at the end of our jupyter notebook.

We welcome your questions by electronic mail at: qante{at}asgard.com