jaehunro

"""Match URLs from WARC records according to regular expression.

WARC records are those that match a SQL query on the columnar URL index.
"""

import re

from bs4 import BeautifulSoup
from bs4.dammit import EncodingDetector
from pyspark.sql import SparkSession

jaehunro

  _recordWrite: function(id, fnName, totalTime, args) {
    // TODO: totalTime isn't that useful since it doesn't count paints/reflows
    var writes =
      ReactDefaultPerf
        ._allMeasurements[ReactDefaultPerf._allMeasurements.length - 1]
        .writes;
    writes[id] = writes[id] || [];
    writes[id].push({
      type: fnName,
      time: totalTime,