NativeExtractor module for Python

This is official Python binding for the NativeExtractor project.

Installation

Requirements

• Python >=2.7 (>3 usage is highly recommended)
• pip
• build-essential (gcc, make)
• libglib2.0, libglib2.0-dev, libpythonX-dev

We recommend to use virtual environments.

virtualenv myproject
source myproject/bin/activate

or

python -m venv myproject
source myproject/bin/activate

Instant PyPi solution

pip install pynativeextractor

Manual

• Clone the repo git clone --recurse-submodules https://github.com/SpongeData-cz/pynativeextractor.git

• Install via pip or pip3

  pip install -e ./pynativeextractor/
  

Typical usage

import os
from pynativeextractor.extractor import BufferStream, Extractor, DEFAULT_MINERS_PATH

# Construct new Extractor instance
ex = Extractor()
# Add fictional miner from web_entities.so with name match_url matching all URLs
ex.add_miner_so(os.path.join(DEFAULT_MINERS_PATH, 'web_entities.so'), 'match_url')
text = '{}'.format("https://spongedata.cz")

# Make from hw stream (you can also do the stream from files - use FileStream - mmap is used internally)
with BufferStream(text) as bf:
    # Initialize occurrences list as empty list
    occurrences = []
    # Set the stream to the extractor
    with ex.set_stream(bf):
        # Mine all occurrences of URLs
        while not ex.eof():
            # Summarize occurrences
            occurrences += ex.next()

print(occurrences) # Prints [{'label': 'URL', 'value': 'https://spongedata.cz', 'pos': 0, 'len': 13, 'prob': 1.0}]