Changeset 33914 for other-projects/maori-lang-detection/mongodb-data/ManualShortlisting.txt

Timestamp:

2020-02-13T17:09:07+13:00 (4 years ago)

Author:

ak19

Message:

Shortlisted just the domain sites by country into ManualShortlist2.txt after taking the reingest into MongoDB into account. And then put all these shortlisted domains for which containsMRI=true as per manual inspection into a separate new file.

File:

• other-projects/maori-lang-detection/mongodb-data/ManualShortlisting.txt (modified) (2 diffs)

other-projects/maori-lang-detection/mongodb-data/ManualShortlisting.txt

@@ -1762 +1762 @@
         "http://teaohou.natlib.govt.nz", 4/4, 2/4
         "http://www.tuwharetoa.iwi.nz", 2/3 0/3
-+        "http://auturoa.nz", 0/4 0/3 [lots of MRI terms among English] - COMMUNITY (But there are pages inMRI to be found by non-random sampling, e.g. http://auturoa.nz/KarakiaMoKuaToRangiTeRaa.html)
+X        "http://auturoa.nz", 0/4 0/3 [lots of MRI terms among English] - COMMUNITY (But there are pages inMRI to be found by non-random sampling, e.g. http://auturoa.nz/KarakiaMoKuaToRangiTeRaa.html)
         "https://www.terito.school.nz", 3/3, 0/2 total
         "https://ttw1.cwp.govt.nz", 3/3 3/3
@@ -1991 +1991 @@
 3. GRAND TOTALS
 
-Count per country of web SITES that contain at least 1 web page containing at least 1 genuine MRI sentence:
-
+Count per country of web SITES that contain at least 1 web page containing at least 1 genuine MRI sentence. (Number in brackets for overseas is number of sites of that geolocation if nz TLDs were NOT grouped with NZ geolocation under "NZ". Number in brackets for NZ indicates the number of sites that are only of NZ geolocation ignoring nz TLDs hosted overseas.)
+
+OLD
 countryCode, num manually inspected sites as having pages containing MRI, num sites openNLP detected as having pages containing MRI
-NZ: 126 actual sites out of 176 detected sites
-US: 29 actual out of 486 detected sites
-AU: 2 actual out of 21 detected sites
+NZ: 126 actual sites out of 176 (89) detected sites
+US: 29 actual out of 422 (486) detected sites
+AU: 2 actual out of 5 (21) detected sites
 DE, Germany: 2 actual out of 27 detected sites
 DK, Denmark: 2 out of 8
 BG, Bulgaria: 1 out of 1
 CZ, Czech Republic: 1 out of 4
-ES, Spain: 1 out of 7
-FR, France: 1 out of 36
+ES, Spain: 1 out of 5 (7)
+FR, France: 1 out of 35 (36)
 IE, Ireland: 1 out of 2
+
 
 TOTAL: 166 sites of all the crawled sites where the crawled set of pages per site actually contained at least one sentence in Māori based on manual inspection.