Possible UTF-8 conversion problems

[awead]

This item in Fedora3 has this description: https://scholarsphere.psu.edu/files/7d279232g#.VMfeyHDF_iE When converted, it displays as: https://scholarsphere-qa.dlt.psu.edu/files/7d279232g#.VMfetXDF_iE

[awead]

When viewing the descMetadata datastream via the Fedora3 web admin:

Objectives:\r\n•

When looking at it from console:

Objectives:\r\nâ\u0080¢

However, note that the contents appear the same when viewing their source and target from within FedoraMigrate::ObjectMover:

(byebug) source.datastreams["descMetadata"].content.split(/\n/)[5]
"<info:fedora/scholarsphere:7d279232g> <http://purl.org/dc/terms/description> \"Objectives:\\r\\nâ\u0080¢  Explain the role of a new genomic assay (Target Nowâ\u0084¢) in guiding oncology treatment plans.\\r\\nâ\u0080¢  Describe the Target Nowâ\u0084¢ assay.\\r\\nâ\u0080¢  Present a case study where Target Nowâ\u0084¢ was instrumental in the patientâ\u0080\u0099s treatment plan.\" ."
(byebug) mover.target.description.first
"Objectives:\r\nâ\u0080¢  Explain the role of a new genomic assay (Target Nowâ\u0084¢) in guiding oncology treatment plans.\r\nâ\u0080¢  Describe the Target Nowâ\u0084¢ assay.\r\nâ\u0080¢  Present a case study where Target Nowâ\u0084¢ was instrumental in the patientâ\u0080\u0099s treatment plan."

All of the above appears consistent, give than \u0080 is the Euro symbol according to: http://www.fileformat.info/info/unicode/char/0080/index.htm under "Java Data"

[awead]

@mjgiarlo @jcoyne This has me stumped. The problem is the character data is coming out of Rubydora looking this way and FedoraMigrate appears to be faithfully replicating it... warts and all.

[mjgiarlo]

It looks like something may be forcing an 8859 encoding and then encoding back to UTF-8. Here's the original string:

irb(main):036:0> original_string
=> "Objectives:\r\n•  Explain the role of a new genomic assay (Target Now™) in guiding oncology treatment plans.\r\n•  Describe the Target Now™ assay.\r\n•  Present a case study where Target Now™ was instrumental in the patient’s treatment plan."
irb(main):037:0> puts original_string
Objectives:
•  Explain the role of a new genomic assay (Target Now™) in guiding oncology treatment plans.
•  Describe the Target Now™ assay.
•  Present a case study where Target Now™ was instrumental in the patient’s treatment plan.
=> nil
irb(main):038:0> original_string.encoding
=> #<Encoding:UTF-8>
irb(main):039:0> original_string.bytes
=> [79, 98, 106, 101, 99, 116, 105, 118, 101, 115, 58, 13, 10, 226, 128, 162, 32, 32, 69, 120, 112, 108, 97, 105, 110, 32, 116, 104, 101, 32, 114, 111, 108, 101, 32, 111, 102, 32, 97, 32, 110, 101, 119, 32, 103, 101, 110, 111, 109, 105, 99, 32, 97, 115, 115, 97, 121, 32, 40, 84, 97, 114, 103, 101, 116, 32, 78, 111, 119, 226, 132, 162, 41, 32, 105, 110, 32, 103, 117, 105, 100, 105, 110, 103, 32, 111, 110, 99, 111, 108, 111, 103, 121, 32, 116, 114, 101, 97, 116, 109, 101, 110, 116, 32, 112, 108, 97, 110, 115, 46, 13, 10, 226, 128, 162, 32, 32, 68, 101, 115, 99, 114, 105, 98, 101, 32, 116, 104, 101, 32, 84, 97, 114, 103, 101, 116, 32, 78, 111, 119, 226, 132, 162, 32, 97, 115, 115, 97, 121, 46, 13, 10, 226, 128, 162, 32, 32, 80, 114, 101, 115, 101, 110, 116, 32, 97, 32, 99, 97, 115, 101, 32, 115, 116, 117, 100, 121, 32, 119, 104, 101, 114, 101, 32, 84, 97, 114, 103, 101, 116, 32, 78, 111, 119, 226, 132, 162, 32, 119, 97, 115, 32, 105, 110, 115, 116, 114, 117, 109, 101, 110, 116, 97, 108, 32, 105, 110, 32, 116, 104, 101, 32, 112, 97, 116, 105, 101, 110, 116, 226, 128, 153, 115, 32, 116, 114, 101, 97, 116, 109, 101, 110, 116, 32, 112, 108, 97, 110, 46]

And here's how things look when I force 8859 then encode as UTF-8:

irb(main):044:0> forced = original_string.force_encoding(Encoding::ISO8859_1).encode(Encoding::UTF_8)
=> "Objectives:\r\nâ\u0080¢  Explain the role of a new genomic assay (Target Nowâ\u0084¢) in guiding oncology treatment plans.\r\nâ\u0080¢  Describe the Target Nowâ\u0084¢ assay.\r\nâ\u0080¢  Present a case study where Target Nowâ\u0084¢ was instrumental in the patientâ\u0080\u0099s treatment plan."
irb(main):045:0> puts forced
Objectives:
•  Explain the role of a new genomic assay (Target Now™) in guiding oncology treatment plans.
•  Describe the Target Now™ assay.
•  Present a case study where Target Now™ was instrumental in the patient’s treatment plan.
=> nil
irb(main):046:0> forced.encoding
=> #<Encoding:UTF-8>
irb(main):047:0> forced.bytes
=> [79, 98, 106, 101, 99, 116, 105, 118, 101, 115, 58, 13, 10, 195, 162, 194, 128, 194, 162, 32, 32, 69, 120, 112, 108, 97, 105, 110, 32, 116, 104, 101, 32, 114, 111, 108, 101, 32, 111, 102, 32, 97, 32, 110, 101, 119, 32, 103, 101, 110, 111, 109, 105, 99, 32, 97, 115, 115, 97, 121, 32, 40, 84, 97, 114, 103, 101, 116, 32, 78, 111, 119, 195, 162, 194, 132, 194, 162, 41, 32, 105, 110, 32, 103, 117, 105, 100, 105, 110, 103, 32, 111, 110, 99, 111, 108, 111, 103, 121, 32, 116, 114, 101, 97, 116, 109, 101, 110, 116, 32, 112, 108, 97, 110, 115, 46, 13, 10, 195, 162, 194, 128, 194, 162, 32, 32, 68, 101, 115, 99, 114, 105, 98, 101, 32, 116, 104, 101, 32, 84, 97, 114, 103, 101, 116, 32, 78, 111, 119, 195, 162, 194, 132, 194, 162, 32, 97, 115, 115, 97, 121, 46, 13, 10, 195, 162, 194, 128, 194, 162, 32, 32, 80, 114, 101, 115, 101, 110, 116, 32, 97, 32, 99, 97, 115, 101, 32, 115, 116, 117, 100, 121, 32, 119, 104, 101, 114, 101, 32, 84, 97, 114, 103, 101, 116, 32, 78, 111, 119, 195, 162, 194, 132, 194, 162, 32, 119, 97, 115, 32, 105, 110, 115, 116, 114, 117, 109, 101, 110, 116, 97, 108, 32, 105, 110, 32, 116, 104, 101, 32, 112, 97, 116, 105, 101, 110, 116, 195, 162, 194, 128, 194, 153, 115, 32, 116, 114, 101, 97, 116, 109, 101, 110, 116, 32, 112, 108, 97, 110, 46]

This is what we're doing here, right? https://github.com/projecthydra-labs/fedora-migrate/blob/master/lib/fedora_migrate/rdf_datastream_mover.rb#L33

[awead]

@mjgiarlo yes, this is what @jcoyne did to correct for objects that were throwing RDF errors. See #9