Difference between revisions of "User:Andreas Plank/Import issues with CETAF identifiers"

Revision as of 15:26, 10 October 2022

Screenshot of the Firefox RESTED plugin (steps to retrieve an RDF data source)

Note: Unresolved or pending issues are on top and issues that are done get to the end. To check for RDF in your browser or on command line:

1. you can use https://www.w3.org/RDF/Validator/ in general
   or use command line tools from Apache Jena (see Documentation), e.g. on Linux:
   /path/to/your/apache-jena-3.15.0/bin/rdfxml --validate "Testfile.rdf"
   # or with log file
   /path/to/your/apache-jena-3.15.0/bin/rdfparse -R "Testfile.rdf" > "Testfile.rdf.ttl" 2> "Testfile.rdf.log"
2. you can more specifically use the CETAF Specimen URI Tester (http://herbal.rbge.info)
3. you can use a plugin in your browser, to basically evaluate redirection to the source RDF, e.g. RESTED Client and then adding Header Accept: application/rdf+xml (see example aside)

data.biodiversitydata.nl (Naturalis)

In some RDF files are invalid URI entries that is, they are not URL-encoded, e.g. <rdf:Description rdf:about="http://data.biodiversitydata.nl/naturalis/specimen/L 0934036`"> having bare spaces or accent characters; URIs having spaces there are many (about ≈278.900), having accent characters there are a view, e.g. with error messages like:

[line: …, col: 68] Illegal character in IRI (codepoint 0x60, '`'): <http://data.biodiversitydata.nl/naturalis/specimen/L%20%200934036[`]...>
[line: …, col: 80] Illegal character in IRI (codepoint 0x60, '`'): <http://data.biodiversitydata.nl/naturalis/specimen/L%20%200799429%20%20%20%20[`]...>
[line: …, col: 68] Illegal character in IRI (codepoint 0x60, '`'): <http://data.biodiversitydata.nl/naturalis/specimen/L%20%200979378[`]...>
[line: …, col: 63] Illegal character in IRI (codepoint 0x60, '`'): <http://data.biodiversitydata.nl/naturalis/specimen/L.4305564[`]...>

These URI entries have to be fixed otherwise it would not be imported, find and replace is fixing this issue manually for the import. --Andreas Plank (talk) 12:30, 8 July 2020 (CEST)

 Done in any case https://github.com/infinite-dao/glean-cetaf-rdfs/blob/main/bin/fixRDF_before_validateRDFs.sh checks URL errors and tries to fix it --Andreas Plank (talk) 15:09, 10 October 2022 (CEST)

coldb.mnhn.fr (MNHN)

Unicode/XML issues:

Notes for a general work around during harvest/import:

• invalid unicode characters break rdfparse and subsequent import, so the harvested RDF must be fixed first manually at this point --Andreas Plank (talk) 12:33, 8 June 2020 (CEST)
• characters that can not be guessed properly will be replaced by a question mark “?” at that position where the wrong unicode character was before

Pending unicode/XML issues (but most will be ignored as these are no plant records to be used for the botany pilot):

• http://coldb.mnhn.fr/catalognumber/mnhn/f/dac98.2 see ? <dwc:municipality>Szirdokpisp?Ki</dwc:municipality>
rdfparse found: [line: …, col: 34] An invalid XML character (Unicode: 0x19) was found in the element content of the document
• http://coldb.mnhn.fr/catalognumber/mnhn/ic/1966-0058 see ? <dwc:occurrenceRemarks>LOC.: TRAPEANG-REPOU, ROUTE VEAL-RENG, KAMP?T</dwc:occurrenceRemarks>
rdfparse found: [line: …, col: 71] An invalid XML character (Unicode: 0x5) was found in the element content of the document.
• http://coldb.mnhn.fr/catalognumber/mnhn/ic/1966-0062 see ? in <dwc:occurrenceRemarks>LOC.: TRAPEANG-REPOU, ROUTE VEAL-RENG, KAMP?T</dwc:occurrenceRemarks>
rdfparse found: [line: …, col: 71] An invalid XML character (Unicode: 0x5) was found in the element content of the document.
• http://coldb.mnhn.fr/catalognumber/mnhn/ic/1966-0061 see ? in <dwc:occurrenceRemarks>LOC.: TRAPEANG-REPOU, ROUTE VEAL-RENG, KAMP?T</dwc:occurrenceRemarks>
rdfparse found: [line: …, col: 71] An invalid XML character (Unicode: 0x5) was found in the element content of the document.
• http://coldb.mnhn.fr/catalognumber/mnhn/ic/1986-0545 see ? <dwc:occurrenceRemarks>1986-545 A 547 DANS LE M?ME BOCAL</dwc:occurrenceRemarks>
rdfparse found: [line: …, col: 52] An invalid XML character (Unicode: 0x14) was found in the element content of the document
  M?ME => MÊME (AP: probably »DANS LE MÊME BOCAL«)
• http://coldb.mnhn.fr/catalognumber/mnhn/ic/1963-0398 see ? <dwc:occurrenceRemarks>1963-398 A 400 DANS LE M?ME BOCAL</dwc:occurrenceRemarks>
rdfparse found: [line: …, col: 52] An invalid XML character (Unicode: 0x14) was found in the element content of the document
  M?ME => MÊME (AP: probably »DANS LE MÊME BOCAL«)
• http://coldb.mnhn.fr/catalognumber/mnhn/ic/1980-1259 see ? <dwc:occurrenceRemarks>M?ME BOCAL QUE 1980-1260</dwc:occurrenceRemarks>
rdfparse found: [line: …, col: 29] An invalid XML character (Unicode: 0x14) was found in the element content of the document
  M?ME => MÊME (AP: probably »MÊME BOCAL QUE 1980-1260«)
• http://coldb.mnhn.fr/catalognumber/mnhn/ic/1963-0399 see ? in <dwc:occurrenceRemarks>1963-398 A 400 DANS LE M?ME BOCAL</dwc:occurrenceRemarks>
rdfparse found: [line: …, col: 52] An invalid XML character (Unicode: 0x14) was found in the element content of the document.
  M?ME => MÊME
• http://coldb.mnhn.fr/catalognumber/mnhn/ic/1980-1260 see ? in <dwc:occurrenceRemarks>M?ME BOCAL QUE 1980-1259</dwc:occurrenceRemarks>
rdfparse found: [line: …, col: 29] An invalid XML character (Unicode: 0x14) was found in the element content of the document.
  M?ME => MÊME
• http://coldb.mnhn.fr/catalognumber/mnhn/ic/1963-0400 see ? in <dwc:occurrenceRemarks>1963-398 A 400 DANS LE MME BOCAL</dwc:occurrenceRemarks>
rdfparse found: [line: …, col: 52] An invalid XML character (Unicode: 0x14) was found in the element content of the document.
  M?ME => MÊME
• http://coldb.mnhn.fr/catalognumber/mnhn/ic/1995-0897 see ? <dwc:occurrenceRemarks>don du northern territory museum extrait du n°13530-003. Proc. Biol. Soc. Wash. v. 109 (no. 2). B?ocal a cote de la 230-0-0-1.</dwc:occurrenceRemarks>
rdfparse found: [line: …, col: 125] An invalid XML character (Unicode: 0x16) was found in the element content of the document
• http://coldb.mnhn.fr/catalognumber/mnhn/ic/b-2510 see ? <dwc:occurrenceRemarks>PARALECTOTYPE DESIGNE PAR SPRINGER, 1962 IN COPEIA No 2? 2 : 4321 EX. EXTRAIT DE A.2024 / D.XII-23 , A.II-23 / VOIR SMITH. CONTR. TO ZOOL., No 73,</dwc:occurrenceRemarks>
rdfparse found: [line: …, col: 83] An invalid XML character (Unicode: 0x1b) was found in the element content of the document
• http://coldb.mnhn.fr/catalognumber/mnhn/ic/a-4687 see ? in <dwc:occurrenceRemarks>SYNTYPE?DE BATRACHUS POROSISSIMUS CUVIER, 1829 IN REGNE ANIMAL (ed. 2) V. 2 : 254</dwc:occurrenceRemarks>
rdfparse found: [line: …, col: 35] An invalid XML character (Unicode: 0x13) was found in the element content of the document.
• http://coldb.mnhn.fr/catalognumber/mnhn/ic/a-4718 see ? in <dwc:occurrenceRemarks>SYNTYPES ?DE BATRACHUS POROSISSIMUS CUVIER, 1829 IN REGNE ANIMAL (ed. 2) V. 2 : 254 / LS = 69 - 71 et 82 mm / LT = 78 - 80,5 et 92 mm</dwc:occurrenceRemarks>
rdfparse found: [line: …, col: 37] An invalid XML character (Unicode: 0x13) was found in the element content of the document.
• http://coldb.mnhn.fr/catalognumber/mnhn/ra/1991.4878 see ? in <dwc:locality>R?mire</dwc:locality>
rdfparse found: [line: …, col: 20] An invalid XML character (Unicode: 0x1a) was found in the element content of the document.
  R?mire => R?mire (perhaps Rémire ?))
• http://coldb.mnhn.fr/catalognumber/mnhn/ra/1991.4926 see ? in <dwc:locality>piste de St H?lie</dwc:locality>
rdfparse found: [line: …, col: 32] An invalid XML character (Unicode: 0x1a) was found in the element content of the document..
  H?lie => H?lie

data.rbge.org.uk (RBGE)

The RDF embedded XML contains pure “&” which is not properly escaped in the XML realm (the propper escape is &). Many RDF files e.g. http://data.rbge.org.uk/herb/E00011206, seems a generic problem --Andreas Plank (talk) 16:33, 16 March 2020 (CET)

• it will be fixed during harvesting routine, but provided XML should be valid including escaped ampersand & --Andreas Plank (talk) 16:33, 16 March 2020 (CET)
•  Done in any case https://github.com/infinite-dao/glean-cetaf-rdfs/blob/main/bin/fixRDF_before_validateRDFs.sh checks & errors and tries to fix it --Andreas Plank (talk) 15:24, 10 October 2022 (CEST)

data.nhm.ac.uk (NHM)

( Pending (minor issue does not block)) Requesting “Content-Type: application/rdf+xml” results in 404 (not found) instead of getting RDF (see https://github.com/NaturalHistoryMuseum/ckanext-nhm/issues/458) --Andreas Plank (talk) 14:06, 18 February 2020 (CET)

• minor issue not relevant because header “Content-Type: application/rdf+xml” is meant for the (returned) resource, not the request --Andreas Plank (talk) 10:40, 20 February 2020 (CET)

No or mixed up RDF description of CETAF-ID

See perhaps the example of CETAF Specimen Preview Profile (CSPP) in general.

id.luomus.fi (LUOMUS)

The requested RDF does not describe the requested CETAF-ID http://id.luomus.fi/GL.749 itself, the ID “hangs somewhat in the air” (from a descriptive point of view):

1. http://id.luomus.fi/GL.749 gets redirected to http://id.luomus.fi/GL.749?format=RDFXML and
2. by analysing the RDF via Apache Jena’s rdfparse it reveals that it describes <http://id.luomus.fi/GL.749?format=RDFXML> <http://purl.org/dc/terms/subject> <http://id.luomus.fi/GL.749> just to be related, but
3. http://id.luomus.fi/GL.749 itself has no related description (rdf:Description) but there are two descriptions http://tun.fi/MY.275076 and http://tun.fi/MY.881682 which do not relate to http://id.luomus.fi/GL.749. So CETAF-ID http://id.luomus.fi/GL.749 “hangs somewhat in the air” because it is not described.

--Andreas Plank (talk) 12:10, 20 February 2020 (CET)

•  Done --Andreas Plank (talk) 16:19, 10 October 2022 (CEST)

id.zfmk.de (ZFMK)

( Pending) The requested RDF does not describe the requested CETAF-ID http://id.zfmk.de/collection_ZFMK/1650/733377/90217 itself, the ID “hangs somewhat in the air” (from a descriptive point of view):

1. http://id.zfmk.de/collection_ZFMK/1650/733377/90217 gets redirected to https://id.zfmk.de/collection_ZFMK/rdf/xml/CollectionSpecimen/1650/733377/90217/?shorturl=1 and
2. by analysing the RDF via Apache Jena’s rdfparse it reveals that it describes something other: https://id.zfmk.de/collection_ZFMK/1650, but unrelated to the ID
3. http://id.zfmk.de/collection_ZFMK/1650/733377/90217 itself has no related description (rdf:Description) and “hangs somewhat in the air”
4. checking the website states a stable URL https://id.zfmk.de/collection_ZFMK/page/CollectionSpecimen/1650 but this very URL does not return any RDF

--Andreas Plank (talk) 12:29, 20 February 2020 (CET)

specimens.kew.org (RBGK)

Requested RDF is instead HTML but RDF --Andreas Plank (talk) 14:32, 18 February 2020 (CET)

fixing seems in progress, which is good, but some IDs from the GBIF API return no specimen but a 404 page (which is possibly an old data record) --Andreas Plank (talk) 10:56, 16 July 2020 (CEST)

seems  Done --Andreas Plank (talk) 16:26, 10 October 2022 (CEST)

For instance under Linux:

wget --header='Accept: application/rdf+xml'  --content-on-error --output-document="specimens.kew.org⁄herbarium⁄1.000.rdf" "http://specimens.kew.org/herbarium/1.000"
head specimens.kew.org⁄herbarium⁄1.000.rdf 
# <!DOCTYPE HTML PUBLIC "-//IETF//DTD HTML 2.0//EN">
# <html><head>
# <title>404 Not Found</title>
# </head><body>
# <h1>Not Found</h1>
# <p>The requested URL /herbarium/1.000 was not found on this server.</p>
# </body></html>

For instance under Linux:

#!/bin/bash
cetaf_uri='http://specimens.kew.org/herbarium/K000001999'
wget --quiet --output-file="${cetaf_uri##*/}.log" --header='Accept: application/rdf+xml' --content-on-error --output-document="${cetaf_uri##*/}.rdf" "${cetaf_uri}"
#   download quietly RDF into file 'K000001999.rdf'

/path/to/local/downloaded/apache-jena-3.14.0/bin/rdfxml --validate 'K000001999.rdf' 
# validate rdf via Apache-Jena command line tool
# K000001999.rdf :: 12:34:44 ERROR riot                 :: [line: 37, col: 67] Invalid byte 2 of 3-byte UTF-8 sequence.

file 'K000001999.rdf'
# show file generic properties
# K000001999.rdf: XML 1.0 document, ISO-8859 text, with very long lines

Comments (AP 2020-08-11 12:51:26):

• K000001999.rdf contains no UTF-8 but ISO-8859 encoded characters
• manual work around would be: iconv -f ISO_8859-1 -t UTF-8 K000001999.rdf

( Done) Requested RDF has dc:relation nesting mistake: it is meant to be only inside <rdf:Description rdf:about="..." ><!-- data --><dc:relation><!-- related rdf:Description nests here --></dc:relation><!-- data --></rdf:Description>, e.g.:

Perhaps develop RDF via TriG format (on Questions, problem solutions and further discussions (Guide of best practices)) helps here ? --Andreas Plank (talk) 13:30, 16 July 2020 (CEST)

The following example compares the actual RDF (left) and the diff command line tool (right) from Linux of wget --header='Accept: application/rdf+xml' --content-on-error --output-document="K000001005.rdf" "http://specimens.kew.org/herbarium/K000001005"

The </rdf:Description> (of the CETAF-ID) ending in line 36 should end much later and must envelop all the <dc:relation> and all other elements accordingly: 35 <dwc:locationRemarks>in umbrosis.</dwc:locationRemarks> 36 </rdf:Description> 37 <!-- Image associated with the specimen --> 38 <dc:relation> 39 <rdf:Description rdf:about="http://www.kew.org/herbcatimg/588771.jpg"> 40 <dc:identifier rdf:resource="http://www.kew.org/herbcatimg/588771.jpg"/> 41 <dc:type rdf:resource="http://purl.org/dc/dcmitype/Image"/> 42 <dc:subject rdf:resource="http://specimens.kew.org/herbarium/K000001005"/> 43 <dc:format>image/jpeg</dc:format> 44 <dc:description xml:lang="en">Image of herbarium specimen</dc:description> 45 <dc:license rdf:resource="https://creativecommons.org/licenses/by/4.0/"/> 46 </rdf:Description> 47 </dc:relation> 48 <dwc:associatedMedia rdf:resource="http://www.kew.org/herbcatimg/588771.jpg"/> 49 </rdf:RDF>

Using diff to illustrate it, the </rdf:Description> counting from line 33 on in line 36, moves to the very bottom before </rdf:RDF>: --- K000001005.rdf 2020-07-16 10:25:35.236116113 +0200 +++ K000001005-fixed.rdf 2020-07-16 10:40:30.246263344 +0200 @@ -33,7 +33,6 @@ <dwc:recordNumber>0</dwc:recordNumber> <dwc:country>Bahia</dwc:country> <dwc:locationRemarks>in umbrosis.</dwc:locationRemarks> -</rdf:Description> <!-- Image associated with the specimen --> <dc:relation> <rdf:Description rdf:about="http://www.kew.org/herbcatimg/588771.jpg"> @@ -46,4 +45,5 @@ </rdf:Description> </dc:relation> <dwc:associatedMedia rdf:resource="http://www.kew.org/herbcatimg/588771.jpg"/> +</rdf:Description> </rdf:RDF>

Done --Andreas Plank (talk) 12:23, 10 August 2020 (CEST)

purl.org/nhmuio (NHMUO)

( Pending) The requested RDF does not describe the requested CETAF-ID http://purl.org/nhmuio/id/41d9cbb4-4590-4265-8079-ca44d46d27c3 itself, the ID “hangs somewhat in the air” (from a descriptive point of view):

1. http://purl.org/nhmuio/id/41d9cbb4-4590-4265-8079-ca44d46d27c3 gets redirected to https://data.gbif.no/resolver/O:L:14 and
2. by analysing the RDF via Apache Jena’s rdfparse it reveals that it describes something other: http://purl.org/gbifnorway/id/O:L:14, but unrelated to the ID
3. http://purl.org/nhmuio/id/41d9cbb4-4590-4265-8079-ca44d46d27c3 itself has no related description (rdf:Description) and “hangs somewhat in the air”

--Andreas Plank (talk) 13:30, 20 February 2020 (CET)

No RDF but HTML

col.smns-bw.org (SMNS)

( Pending) Requested RDF is instead an HTML fragment but RDF.--Andreas Plank (talk) 14:38, 18 February 2020 (CET)

For instance under Linux:

wget --header='Accept: application/rdf+xml'  --content-on-error --output-document="col.smns-bw.org⁄object⁄S10000227722006.rdf" "http://col.smns-bw.org/object/S10000227722006"
file col.smns-bw.org⁄object⁄S10000227722006.rdf
# col.smns-bw.org⁄object⁄S10000227722006.rdf: HTML document, ISO-8859 text, with very long lines, with CRLF line terminators

Fixed Issues

herbarium.bgbm.org (BGBM)

( Done) In some RDF files are invalid URI entries i.e. there is a tab/space character in the URI in owl:sameAs and this would break the whole import of data. The error log of triple store loader (tdbloader2) shows something like:

Bad URI: < http://purl.oclc.org/net/edu.harvard.huh/guid/uuid/a86596ea-6f4d-4b97-bf6f-8d492c0fc8b2> Code: 0/ILLEGAL_CHARACTER in SCHEME: The character violates the grammar rules for URIs/IRIs. ERROR Bad character in IRI (space): <[space]...>

… see for instance in line 63:

62 <rdf:Description rdf:about="http://www.wikidata.org/entity/Q6382619">
63                     <owl:sameAs rdf:resource="	http://purl.oclc.org/net/edu.harvard.huh/guid/uuid/a86596ea-6f4d-4b97-bf6f-8d492c0fc8b2" />
64                 <owl:sameAs rdf:resource="http://viaf.org/viaf/233473288" />
65           </rdf:Description>

The following objects were detected:

• http://herbarium.bgbm.org/data/rdf/B100000580 --Andreas Plank (talk) 16:21, 30 January 2020 (CET)  Done --Andreas Plank (talk) 11:45, 3 February 2020 (CET)
• http://herbarium.bgbm.org/data/rdf/B100000503 --Andreas Plank (talk) 16:21, 30 January 2020 (CET)  Done --Andreas Plank (talk) 11:45, 3 February 2020 (CET)
• http://herbarium.bgbm.org/data/rdf/B100000627 --Andreas Plank (talk) 16:21, 30 January 2020 (CET)  Done --Andreas Plank (talk) 11:45, 3 February 2020 (CET)

snsb.info (SNSB)

 Done: Mistakes of naming RDF elements/properties: they are sometimes mixed with CSPP-Element names, e.g. dwc:kindOfMaterial but with “kindOfMaterial” being meant the CSPP element name only, not the designated property (code) term; I found the following mistakes (considered using the following prefixes) :
PREFIX dwc: <http://rs.tdwg.org/dwc/terms/> and PREFIX dcterms: <http://purl.org/dc/terms/>
… the following elements are mistaken for instance and do not resolve:

• <dwc:kindOfMaterial> => <dcterms:type>
• <dwc:collectionDate> => <dcterms:created>
• <dwc:sourceLink> => <dcterms:publisher>

Perhaps there are more RDF elements to fix. --Andreas Plank (talk) 15:22, 8 October 2020 (CEST)

 Done --Andreas Plank (talk) 10:25, 14 October 2020 (CEST)