Encoding of topictitles

[iboeckma]

I don't know why it didn't strike my attention earlier. We made tests with topics that have german umlauts, maybe the topics had the same numbers of num_ret coincidentally..

1. only the titles where searched, not the narratives
2. boolean OR
3. with synonyms

solr-version: 8.4.1

topicid: 20, topictitle: Lateinamerikanische Tänze

occurrence	num_ret	encoding	link
in solr	116	Lateinamerikanische%20T%C3%A4nze	http://localhost:8983/solr/gelic-3/select?fl=score%2Cid&q=subject_gnd_ss%3A(Lateinamerikanische%20T%C3%A4nze)&rows=300&wt=json
by script	3	Lateinamerikanische%20Ta%CC%88nze	http://localhost:8983/solr/gelic-3/select?fl=score%2Cid&q=subject_gnd_ss%3A(Lateinamerikanische%20Ta%CC%88nze)&rows=300&wt=json

topicid: 21, topictitle: Gesprächsführung

occurrence	num_ret	encoding	link
in solr	41	Gespr%C3%A4chsf%C3%BChrung	http://localhost:8983/solr/gelic-3/select?fl=score%2Cid&q=subject_gnd_ss%3A(Gespr%C3%A4chsf%C3%BChrung)&rows=300&wt=json
by script	0	Gespra%CC%88chsfu%CC%88hrung	http://localhost:8983/solr/gelic-3/select?fl=score%2Cid&q=subject_gnd_ss%3A(Gespra%CC%88chsfu%CC%88hrung)&rows=300&wt=json

tests

test 1: solr-version 8.2.0

topicid: 20, topictitle: Lateinamerikanische Tänze

occurrence	num_ret	encoding	link
in solr	116	Lateinamerikanische%20T%C3%A4nze	http://localhost:8983/solr/gelic-3/select?fl=score%2Cid&q=subject_gnd_ss%3A(Lateinamerikanische%20T%C3%A4nze)&rows=300&wt=json
by script	3	Lateinamerikanische%20Ta%CC%88nze	http://localhost:8983/solr/gelic-3/select?fl=score%2Cid&q=subject_gnd_ss%3A(Lateinamerikanische%20Ta%CC%88nze)&rows=300&wt=json

-> solr-version didn't change anything with the encoding

test 2: basic managed-schema, no changes to solrconfig.xml or synonyms.txt

topicid: 20, topictitle: Lateinamerikanische Tänze

occurrence	num_ret	encoding	link
in solr	0	Lateinamerikanische%20T%C3%A4nze	http://localhost:8983/solr/gelic-3/select?fl=score%2Cid&q=subject_gnd_ss%3A(Lateinamerikanische%20T%C3%A4nze)&rows=300&wt=json
by script	0	Lateinamerikanische%20Ta%CC%88nze	http://localhost:8983/solr/gelic-3/select?fl=score%2Cid&q=subject_gnd_ss%3A(Lateinamerikanische%20Ta%CC%88nze)&rows=300&wt=json

-> configurations didn't change the encoding (num_ret is 0 because of almost no configurations)

test 3: changed topic 20 in topics.xml "Lateinamerikanische Tänze" -> "Lateinamerikanische%20T%C3%A4nze"

topicid: 20, topictitle: Lateinamerikanische Tänze

occurrence	num_ret	encoding	link
in solr	116	Lateinamerikanische%20T%C3%A4nze	http://localhost:8983/solr/gelic-3/select?fl=score%2Cid&q=subject_gnd_ss%3A(Lateinamerikanische%20T%C3%A4nze)&rows=300&wt=json
by script	3	Lateinamerikanische%2520T%25C3%25A4nze	http://localhost:8983/solr/gelic-3/select?fl=score%2Cid&q=subject_gnd_ss%3A(Lateinamerikanische%2520T%25C3%25A4nze)&rows=300&wt=json

-> maybe it's a sign to not work with german umlauts anymore.. But I hope it's just my script that isn't working or a problem that I overlooked.

zip-file

1. I changed evalpertopic.py a bit to get the same url-structure as one gets when searching in the solr interface. This version is in the zip.
   The zip also includes a fieldnames.json file that only searches gnd_ss, so that the right url is quicker to find in the terminal window.
   These two files have to be in [solrinstallationname]/scripts.
2. topics.xml and assessments.txt need to be in [solrinstallationname]/components.
3. Used solrconfig.xml, managed-schema and synoyms.txt is also in the zip.

filesfortesting.zip