Skip to content

Commit

Permalink
Merge pull request #429 from usc-isi-i2/development
Browse files Browse the repository at this point in the history 
update for rdflin 6.0.0, version
  • Loading branch information
@saggu
saggu committed Jul 23, 2021
2 parents 906e700 + 566d944 commit df79e1e
Show file tree
Hide file tree
Showing 3 changed files with 7 additions and 4 deletions.
7 changes: 5 additions & 2 deletions etk/unit_tests/test_html_content_extractor.py
View file
Original file line number Diff line number Diff line change
@@ -1,5 +1,8 @@
import unittest
import json

import sys

from etk.extractors.html_content_extractor import HTMLContentExtractor, Strategy


Expand All @@ -17,7 +20,7 @@ def test_html_content_extractor(self) -> None:
main_strict = hce.extract(test_html, strategy=Strategy.MAIN_CONTENT_STRICT)[0].value
main_relax = hce.extract(test_html, strategy=Strategy.MAIN_CONTENT_RELAXED)[0].value

expected_all_text = """ 323-452-2013 Los Angeles Escort Listing View Escorts in other cities smoothlegs24 28 chrissy391 27 My name is Helena height 160cms weight 55 kilos contact me at escort.here@gmail.com jefferson ave age: 23 HrumpMeNow 28 xxtradition 24 jumblyjumb 26 claudia77 26 gushinPuss 28 Littlexdit 25 PinkSweets2 28 withoutlimit 27 bothOfUs3 28 lovelylips 27 killerbod 27 Littlexdit 27 azneyes 23 323-452-2013 ESCORT ALERT! - Luna The Hot Playmate (323) 452-2013 - 23 Escort's Phone: 323-452-2013 Escort's Location: Los Angeles, California Escort's Age: 23 Date of Escort Post: Jan 02nd 6:46am REVIEWS: READ AND CREATE REVIEWS FOR THIS ESCORT There are 50 girls looking in . VIEW GIRLS Hey I'm luna 3234522013 Let's explore , embrace and indulge in your favorite fantasy % independent. discreet no drama Firm Thighs and Sexy. My Soft skin & Tight Grip is exactly what you deserve Call or text Fetish friendly Fantasy friendly Party friendly 140 Hr SPECIALS 3234522013. Call 323-452-2013 . Me and my friends are on EZsex soooo you can find us all on there if you want... skittlegirl Call me on my cell at 323-452-2013. Date of ad: 2017-01-02 06:46:00 More posts from 323-452-2013 323-452-2013 Nov 04, 2016 Guess whose New with 80Specials (323) 452-2013 - 21 Recent Escort Classifieds North Jersey, New Jersey (732-621-4443) :*: G O O D G I R L :*: G O N E *:**: B A D ;) LATINA - 21 Chicago, Illinois (773-412-2044) ( LAtE NiGHt ) UNRUShEd (ULTiMAtE) PLEASURE (*AmAziNg Azz*) CHOOSE..W... Chicago, Illinois (414-914-3777) Petite. and Sweet. Super new and Ready... in out call - 21 Chicago, Illinois (312-600-8628) CM Get PAMPERED - 22 Ft Lauderdale, Florida (786-504-1860) you will ME. GORGEOUSLY sexy LATIN - 21 Atlanta, Georgia (401-324-9388) WoW. MuSt TaKe A LoOk At ThIs. - 21 Atlanta, Georgia (347-940-1982) SMOKiNG HOT Specials BuSTy BaBe (( 5 SeRvICe )) Pretty 36DDDs ( ) ( ... Atlanta, Georgia (305-849-8140) Here to please all your desiresCALL ME 100% Real Curvy White Girl. - 21 Houston, Texas (832-914-9667) Beautiful Salvadorean The One And Only(: - 21 Phoenix, Arizona (623-500-7076) NEW GIRL PERSIAN Gem EXotIC Blend - 21 Toronto, Ontario (416-554-3337) (L) (L) ~~~Special 80 for 20 min:) 22YeAr oLd $$exyy LaTiNa BoMbSheLL~~(L... Toronto, Ontario (416-520-5198) **21 years old * $80 **real pictures ** A sian Kathy *** - 21 Toronto, Ontario (647-702-6825) **: :** 18 Yr Old Mercedes **: :** Incalls/Outcalls (647)702-6825 - 1... Toronto, Ontario (416-275-1296) 100% real pictures ** True college sweet heart ** Taiwan Kelly ** 416-275... Northern Virginia, DC (206-922-9303) SPECIALs 60/90/150 FuN & WiLd BrUnEtTe ThIcK & BuSty JUSt THe WAy U LIKe ... Chicago, Illinois (608-609-5899) New To Town Dont Miss Me 100$ Specials - 18 Dallas, Texas (469-510-5849) JENNY ...A New RedHeaded Firecracker. YeeHaw. - 21 New York, New York (442-222-0227) F R E A K Y __ O B S E S S I O N UPSCALE UNRUSHED-100%... New York, New York (917-676-1333) SuPeR SpEciALs Red BoNe ThiCk Incalls - 19 Houston, Texas (301-000-6018) (Jamaican Ebony Goddess) petite sexy spontaneous UPSCALE discreet dates M... Site Map Call this Escort at 323-452-2013 View escorts in other cities Register Here Login to your account Key for Escort Acronyms Top 10 Escort Practices Prostitution Laws Around the World Prostitution Laws In Your State Top Escort Cities New York, New York Toronto, Ontario Dallas, Texas Chicago, Illinois Atlanta, Georgia North Jersey, New Jersey Detroit, Michigan Los Angeles, California San Diego, California Orange County, California Houston, Texas Phoenix, Arizona Philadelphia, Pennsylvania Boston, Massachusetts Washington DC, DC Orlando, Florida Las Vegas, Nevada Miami, Florida Recent Blog Posts Sheriff candidate Minister and Detective Reno Fells arrested in prostitution bust Man gets 35 years for impersonating cop to get free sex from hooker Alexander Marino: Psychologist by Day, Pimp by Night Surfside Beach, SC Prostitution BUST: Video College Park, Maryland Prostitution Sting: 30 arrested SMU Players Robbed By Escort Study: 1 in 6 Sex Offenders Live Double Life Dark Knight Rises Suspect Killer James Holmes Bought Prostitutes Online Hookers Picked Up Off the Street Hookers paid extra for photos Links Escorts on Webcam Ad Removal Form Contact Us Search Box Search For Profiles Register Here Login to your account Non Mobile Version Escort Blog Key for Escort Acronyms Top 10 Escort Practices Escort Reviews See Escorts on Webcam Prostitution Laws Most Recently Viewed Today at 10:02am Pacific 346-908-0622 Killeen """
expected_all_text = """ 323-452-2013 Los Angeles Escort Listing View Escorts in other cities smoothlegs24 28 chrissy391 27 My name is Helena height 160cms weight 55 kilos contact me at escort.here@gmail.com jefferson ave age: 23 HrumpMeNow 28 xxtradition 24 jumblyjumb 26 claudia77 26 gushinPuss 28 Littlexdit 25 PinkSweets2 28 withoutlimit 27 bothOfUs3 28 lovelylips 27 killerbod 27 Littlexdit 27 azneyes 23 323-452-2013 ESCORT ALERT! - Luna The Hot Playmate (323) 452-2013 - 23 Escort's Phone: 323-452-2013 Escort's Location: Los Angeles, California Escort's Age: 23 Date of Escort Post: Jan 02nd 6:46am REVIEWS: READ AND CREATE REVIEWS FOR THIS ESCORT There are 50 girls looking in . VIEW GIRLS Hey I'm luna 3234522013 Let's explore , embrace and indulge in your favorite fantasy % independent. discreet no drama Firm Thighs and Sexy. My Soft skin & Tight Grip is exactly what you deserve Call or text Fetish friendly Fantasy friendly Party friendly 140 Hr SPECIALS 3234522013.\u00a0Call 323-452-2013 . Me and my friends are on EZsex soooo you can find us all on there if you want... skittlegirl Call me on my cell at 323-452-2013. Date of ad: 2017-01-02 06:46:00 More posts from 323-452-2013 323-452-2013 Nov 04, 2016 Guess whose New with 80Specials (323) 452-2013 - 21 Recent Escort Classifieds North Jersey, New Jersey (732-621-4443) :*: G O O D G I R L :*: G O N E *:**: B A D ;) LATINA - 21 Chicago, Illinois (773-412-2044) ( LAtE NiGHt ) UNRUShEd (ULTiMAtE) PLEASURE (*AmAziNg Azz*) CHOOSE..W... Chicago, Illinois (414-914-3777) Petite. and Sweet. Super new and Ready... in out call - 21 Chicago, Illinois (312-600-8628) CM Get PAMPERED - 22 Ft Lauderdale, Florida (786-504-1860) you will ME. GORGEOUSLY sexy LATIN - 21 Atlanta, Georgia (401-324-9388) WoW. MuSt TaKe A LoOk At ThIs. - 21 Atlanta, Georgia (347-940-1982) SMOKiNG HOT Specials BuSTy BaBe (( 5 SeRvICe )) Pretty 36DDDs ( ) ( ... Atlanta, Georgia (305-849-8140) Here to please all your desiresCALL ME 100% Real Curvy White Girl. - 21 Houston, Texas (832-914-9667) Beautiful Salvadorean The One And Only(: - 21 Phoenix, Arizona (623-500-7076) NEW GIRL PERSIAN Gem EXotIC Blend - 21 Toronto, Ontario (416-554-3337) (L) (L) ~~~Special 80 for 20 min:) 22YeAr oLd $$exyy LaTiNa BoMbSheLL~~(L... Toronto, Ontario (416-520-5198) **21 years old * $80 **real pictures ** A sian Kathy *** - 21 Toronto, Ontario (647-702-6825) **: :** 18 Yr Old Mercedes **: :** Incalls/Outcalls (647)702-6825 - 1... Toronto, Ontario (416-275-1296) 100% real pictures ** True college sweet heart ** Taiwan Kelly ** 416-275... Northern Virginia, DC (206-922-9303) SPECIALs 60/90/150 FuN & WiLd BrUnEtTe ThIcK & BuSty JUSt THe WAy U LIKe ... Chicago, Illinois (608-609-5899) New To Town Dont Miss Me 100$ Specials - 18 Dallas, Texas (469-510-5849) JENNY ...A New RedHeaded Firecracker. YeeHaw. - 21 New York, New York (442-222-0227) F R E A K Y __ O B S E S S I O N UPSCALE UNRUSHED-100%... New York, New York (917-676-1333) SuPeR SpEciALs Red BoNe ThiCk Incalls - 19 Houston, Texas (301-000-6018) (Jamaican Ebony Goddess) petite sexy spontaneous UPSCALE discreet dates M... Site Map Call this Escort at 323-452-2013 View escorts in other cities Register Here Login to your account Key for Escort Acronyms Top 10 Escort Practices Prostitution Laws Around the World Prostitution Laws In Your State Top Escort Cities New York, New York Toronto, Ontario Dallas, Texas Chicago, Illinois Atlanta, Georgia North Jersey, New Jersey Detroit, Michigan Los Angeles, California San Diego, California Orange County, California Houston, Texas Phoenix, Arizona Philadelphia, Pennsylvania Boston, Massachusetts Washington DC, DC Orlando, Florida Las Vegas, Nevada Miami, Florida Recent Blog Posts Sheriff candidate Minister and Detective Reno Fells arrested in prostitution bust Man gets 35 years for impersonating cop to get free sex from hooker Alexander Marino: Psychologist by Day, Pimp by Night Surfside Beach, SC Prostitution BUST: Video College Park, Maryland Prostitution Sting: 30 arrested SMU Players Robbed By Escort Study: 1 in 6 Sex Offenders Live Double Life Dark Knight Rises Suspect Killer James Holmes Bought Prostitutes Online Hookers Picked Up Off the Street Hookers paid extra for photos Links Escorts on Webcam Ad Removal Form Contact Us Search Box Search For Profiles Register Here Login to your account Non Mobile Version Escort Blog Key for Escort Acronyms Top 10 Escort Practices Escort Reviews See Escorts on Webcam Prostitution Laws Most Recently Viewed Today at 10:02am Pacific 346-908-0622 Killeen """
# if readability is improved in the future, the expectation should be:
# expected_main_strict = """\n \n \n \n \n \n smoothlegs24 28 \n \n \n chrissy391 27 \n \n \n My name is Helena height 160cms weight 55 kilos contact me at escort.here@gmail.com jefferson ave age: 23 HrumpMeNow 28 \n \n \n xxtradition 24 \n \n \n jumblyjumb 26 \n \n \n claudia77 26 \n \n \n gushinPuss 28 \n \n \n Littlexdit 25 \n \n \n PinkSweets2 28 \n \n \n withoutlimit 27 \n \n \n bothOfUs3 28 \n \n \n lovelylips 27 \n \n \n killerbod 27 \n \n \n Littlexdit 27 \n \n \n azneyes 23 \n \n \n \n \n \n Escort's Phone: \n \n \n323-452-2013 \n \n Escort's Location: \nLos Angeles, California \n Escort's Age: 23 Date of Escort Post: Jan 02nd 6:46am \n REVIEWS: \n READ AND CREATE REVIEWS FOR THIS ESCORT \n \n \n \n \n \nThere are 50 girls looking in .\n VIEW GIRLS \n \nHey I'm luna 3234522013 Let's explore , embrace and indulge in your favorite fantasy % independent. discreet no drama Firm Thighs and Sexy. My Soft skin & Tight Grip is exactly what you deserve Call or text Fetish friendly Fantasy friendly Party friendly 140 Hr SPECIALS 3234522013. Call 323-452-2013 . Me and my friends are on EZsex soooo you can find us all on there if you want... skittlegirl \n   \n \n   \n \n   \n Call me on my cell at 323-452-2013. Date of ad: 2017-01-02 06:46:00 \n \n \n"""
# expected_main_relax = """\n \n \n \n \n \n \n smoothlegs24 28 \n \n \n chrissy391 27 \n \n \n My name is Helena height 160cms weight 55 kilos contact me at escort.here@gmail.com jefferson ave age: 23 HrumpMeNow 28 \n \n \n xxtradition 24 \n \n \n jumblyjumb 26 \n \n \n claudia77 26 \n \n \n gushinPuss 28 \n \n \n Littlexdit 25 \n \n \n PinkSweets2 28 \n \n \n withoutlimit 27 \n \n \n bothOfUs3 28 \n \n \n lovelylips 27 \n \n \n killerbod 27 \n \n \n Littlexdit 27 \n \n \n azneyes 23 \n \n \n \n \n \n Escort's Phone: \n \n \n323-452-2013 \n \n Escort's Location: \nLos Angeles, California \n Escort's Age: 23 Date of Escort Post: Jan 02nd 6:46am \n REVIEWS: \n READ AND CREATE REVIEWS FOR THIS ESCORT \n \n \n \n \n \nThere are 50 girls looking in .\n VIEW GIRLS \n \nHey I'm luna 3234522013 Let's explore , embrace and indulge in your favorite fantasy % independent. discreet no drama Firm Thighs and Sexy. My Soft skin & Tight Grip is exactly what you deserve Call or text Fetish friendly Fantasy friendly Party friendly 140 Hr SPECIALS 3234522013.\u00a0Call 323-452-2013 . Me and my friends are on EZsex soooo you can find us all on there if you want... skittlegirl \n \u00a0\u00a0\n \n \u00a0\u00a0\n \n \u00a0\u00a0\n Call me on my cell at 323-452-2013. Date of ad: 2017-01-02 06:46:00 \n \n \n \n" != "\n \n \n \n \n \n Escort's Phone: \n \n \n323-452-2013 \n \n Escort's Location: \nLos Angeles, California \n Escort's Age: 23 Date of Escort Post: Jan 02nd 6:46am \n REVIEWS: \n READ AND CREATE REVIEWS FOR THIS ESCORT \n \n \n \n \n \nThere are 50 girls looking in .\n VIEW GIRLS \n \nHey I'm luna 3234522013 Let's explore , embrace and indulge in your favorite fantasy % independent. discreet no drama Firm Thighs and Sexy. My Soft skin & Tight Grip is exactly what you deserve Call or text Fetish friendly Fantasy friendly Party friendly 140 Hr SPECIALS 3234522013.\u00a0Call 323-452-2013 . Me and my friends are on EZsex soooo you can find us all on there if you want... skittlegirl \n \u00a0\u00a0\n \n \u00a0\u00a0\n \n \u00a0\u00a0\n Call me on my cell at 323-452-2013. Date of ad: 2017-01-02 06:46:00 \n \n \n \n"""
Expand All @@ -30,4 +33,4 @@ def test_html_content_extractor(self) -> None:


if __name__ == '__main__':
unittest.main()
unittest.main()
2 changes: 1 addition & 1 deletion requirements.txt
View file
Original file line number Diff line number Diff line change
Expand Up @@ -48,6 +48,6 @@ pyexcel-io>=0.5.6
pyexcel-xls>=0.5.6
pyexcel-xlsx>=0.5.6
SPARQLWrapper>=1.8.2
pyshacl==0.9.10
pyshacl>=0.9.10
tabula-py>=1.4.3
requests>=2.18.4
2 changes: 1 addition & 1 deletion setup.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -12,7 +12,7 @@

setuptools.setup(
name="etk",
version="2.2.6",
version="2.2.7",
author="Amandeep Singh",
author_email="amandeep.s.saggu@gmail.com",
description="extraction toolkit",
Expand Down

0 comments on commit df79e1e

Please sign in to comment.