Description
Matthias Meusburger
2024-10-10 15:58:59 UTC
Created attachment 172684 [details] [review] Bug 38142: Choose language to report from authority to bibliographic record. Currently, when reporting authority field to the bibliographic records, the merge function uses the first matching authority field. However, authorities can provide multiple versions of the same field, with different languages, identified by a code in the $7 field. French reference for languages codes for example, but any code system could be used: https://documentation.abes.fr/sudoc/formats/unmb/DonneesCodees/CodesZone104.htm#$d This patch introduces the LanguageToReportOnMerge system preference, which allows to specify a language code. When set, if there is an authority field with the given language code, it will be used to report to the bibliographic record. If the system preference is not set, or there is no matching field, the default behavior will apply: the first field will be used. The $7 in the authority field can be in a short or long form: - the short form is 2 characters long. Example: ba - the long form is 8 characters long, with the short form in position 5 and 6 (starting from 1). Example: ba0yba0y Test plan: ---------- 1) Edit the 'Personal Name' authority type framework to make the 100 field repeatable 2) Add an authority with several languages, like this (the order is important): 100 _7ba0yda0y _a伊東豊雄 100 _7ba0yba0y _aItō 3) Add a bibliographic record linked to that authority. (Click on the "tag editor" icon next to the 100$a field in the bibliographic record to search for the authority, then Search main heading ($a only): contains Itō) Note that you can choose which version you want when searching for the authority, under the "Get it!" column For this example, choose the latin version: Itō You should now have this in your bibliographic record: 100 _aItō _92646 4) Edit the authority, you don't have to change anything, just edit and save. Note that your bibliographic record now looks like this: 100 _7ba0yda0y _a伊東豊雄 _92646 The first 100 field of the authority has been reported back to the bibliographic record, regardless of what you chose when selecting the authority. 5) Apply the patch, set the LanguageToReportOnMerge system preference to: 'ba' 6) Edit the authority again, and check that your bibliographic record now looks like this, according to the system preference: 100 _7ba0yba0y _aItō _92646 7) Edit the authority again, but set the $7 of the latin field to its short form: 100 _7ba0yda0y _a伊東豊雄 100 _7ba _aItō 8) Check that the bibliographic record still has the latin field reported: 100 _7ba _aItō _92646 9) prove koha/t/db_dependent/Authority/Merge.t 10) Sign-off :) Sponsored-by: ENSA Very interesting! Two remarks/demands: - Could you develop the syspref description? Something with sample, and explaining the 2 language code forms, short and long. - Wouldn't it make sense to have a repeatable languages selection? Let say I select 'ba,fa'. It means that I want first 'ba' authority, if not available then 'fa' authority, otherwise the 1st one. Created attachment 172802 [details] [review] Bug 38142: Choose language to report from authority to bibliographic record. Currently, when reporting authority field to the bibliographic records, the merge function uses the first matching authority field. However, authorities can provide multiple versions of the same field, with different languages, identified by a code in the $7 field. French reference for languages codes for example, but any code system could be used: https://documentation.abes.fr/sudoc/formats/unmb/DonneesCodees/CodesZone104.htm#$d This patch introduces the LanguageToReportOnMerge system preference, which allows to specify a language code. When set, if there is an authority field with the given language code, it will be used to report to the bibliographic record. If the system preference is not set, or there is no matching field, the default behavior will apply: the first field will be used. The $7 in the authority field can be in a short or long form: - the short form is 2 characters long. Example: ba - the long form is 8 characters long, with the short form in position 5 and 6 (starting from 1). Example: ba0yba0y Test plan: ---------- 1) Edit the 'Personal Name' authority type framework to make the 100 field repeatable 2) Add an authority with several languages, like this (the order is important): 100 _7ba0yda0y _a伊東豊雄 100 _7ba0yba0y _aItō 3) Add a bibliographic record linked to that authority. (Click on the "tag editor" icon next to the 100$a field in the bibliographic record to search for the authority, then Search main heading ($a only): contains Itō) Note that you can choose which version you want when searching for the authority, under the "Get it!" column For this example, choose the latin version: Itō You should now have this in your bibliographic record: 100 _aItō _92646 4) Edit the authority, you don't have to change anything, just edit and save. Note that your bibliographic record now looks like this: 100 _7ba0yda0y _a伊東豊雄 _92646 The first 100 field of the authority has been reported back to the bibliographic record, regardless of what you chose when selecting the authority. 5) Apply the patch, set the LanguageToReportOnMerge system preference to: 'ba' 6) Edit the authority again, and check that your bibliographic record now looks like this, according to the system preference: 100 _7ba0yba0y _aItō _92646 7) Edit the authority again, but set the $7 of the latin field to its short form: 100 _7ba0yda0y _a伊東豊雄 100 _7ba _aItō 8) Check that the bibliographic record still has the latin field reported: 100 _7ba _aItō _92646 9) prove koha/t/db_dependent/Authority/Merge.t 10) Sign-off :) Sponsored-by: ENSA (In reply to Frédéric Demians from comment #2) > Very interesting! Two remarks/demands: > > - Could you develop the syspref description? Something with sample, and > explaining the 2 language code forms, short and long. Agreed. The patch has been updated with a more descriptive syspref. > > - Wouldn't it make sense to have a repeatable languages selection? Let say I > select 'ba,fa'. It means that I want first 'ba' authority, if not available > then 'fa' authority, otherwise the 1st one. This is an interesting idea, but maybe perfect is the enemy of good :) I this this could be added later in another BZ if needed. Created attachment 172923 [details] [review] Bug 38142: Choose language to report from authority to bibliographic record. Currently, when reporting authority field to the bibliographic records, the merge function uses the first matching authority field. However, authorities can provide multiple versions of the same field, with different languages, identified by a code in the $7 field. French reference for languages codes for example, but any code system could be used: https://documentation.abes.fr/sudoc/formats/unmb/DonneesCodees/CodesZone104.htm#$d This patch introduces the LanguageToReportOnMerge system preference, which allows to specify a language code. When set, if there is an authority field with the given language code, it will be used to report to the bibliographic record. If the system preference is not set, or there is no matching field, the default behavior will apply: the first field will be used. The $7 in the authority field can be in a short or long form: - the short form is 2 characters long. Example: ba - the long form is 8 characters long, with the short form in position 5 and 6 (starting from 1). Example: ba0yba0y Test plan: ---------- 1) Edit the 'Personal Name' authority type framework to make the 100 field repeatable 2) Add an authority with several languages, like this (the order is important): 100 _7ba0yda0y _a伊東豊雄 100 _7ba0yba0y _aItō 3) Add a bibliographic record linked to that authority. (Click on the "tag editor" icon next to the 100$a field in the bibliographic record to search for the authority, then Search main heading ($a only): contains Itō) Note that you can choose which version you want when searching for the authority, under the "Get it!" column For this example, choose the latin version: Itō You should now have this in your bibliographic record: 100 _aItō _92646 4) Edit the authority, you don't have to change anything, just edit and save. Note that your bibliographic record now looks like this: 100 _7ba0yda0y _a伊東豊雄 _92646 The first 100 field of the authority has been reported back to the bibliographic record, regardless of what you chose when selecting the authority. 5) Apply the patch, set the LanguageToReportOnMerge system preference to: 'ba' 6) Edit the authority again, and check that your bibliographic record now looks like this, according to the system preference: 100 _7ba0yba0y _aItō _92646 7) Edit the authority again, but set the $7 of the latin field to its short form: 100 _7ba0yda0y _a伊東豊雄 100 _7ba _aItō 8) Check that the bibliographic record still has the latin field reported: 100 _7ba _aItō _92646 9) prove koha/t/db_dependent/Authority/Merge.t 10) Sign-off :) Sponsored-by: ENSA Works smoothly, both with UNIMARC/MARC21 Signed-off-by: Frédéric Demians <f.demians@tamil.fr> Doesn't apply cleanly to main Created attachment 177508 [details] [review] Bug 38142: Choose language to report from authority to bibliographic record. Currently, when reporting authority field to the bibliographic records, the merge function uses the first matching authority field. However, authorities can provide multiple versions of the same field, with different languages, identified by a code in the $7 field. French reference for languages codes for example, but any code system could be used: https://documentation.abes.fr/sudoc/formats/unmb/DonneesCodees/CodesZone104.htm#$d This patch introduces the LanguageToReportOnMerge system preference, which allows to specify a language code. When set, if there is an authority field with the given language code, it will be used to report to the bibliographic record. If the system preference is not set, or there is no matching field, the default behavior will apply: the first field will be used. The $7 in the authority field can be in a short or long form: - the short form is 2 characters long. Example: ba - the long form is 8 characters long, with the short form in position 5 and 6 (starting from 1). Example: ba0yba0y Test plan: ---------- 1) Edit the 'Personal Name' authority type framework to make the 100 field repeatable 2) Add an authority with several languages, like this (the order is important): 100 _7ba0yda0y _a伊東豊雄 100 _7ba0yba0y _aItō 3) Add a bibliographic record linked to that authority. (Click on the "tag editor" icon next to the 100$a field in the bibliographic record to search for the authority, then Search main heading ($a only): contains Itō) Note that you can choose which version you want when searching for the authority, under the "Get it!" column For this example, choose the latin version: Itō You should now have this in your bibliographic record: 100 _aItō _92646 4) Edit the authority, you don't have to change anything, just edit and save. Note that your bibliographic record now looks like this: 100 _7ba0yda0y _a伊東豊雄 _92646 The first 100 field of the authority has been reported back to the bibliographic record, regardless of what you chose when selecting the authority. 5) Apply the patch, set the LanguageToReportOnMerge system preference to: 'ba' 6) Edit the authority again, and check that your bibliographic record now looks like this, according to the system preference: 100 _7ba0yba0y _aItō _92646 7) Edit the authority again, but set the $7 of the latin field to its short form: 100 _7ba0yda0y _a伊東豊雄 100 _7ba _aItō 8) Check that the bibliographic record still has the latin field reported: 100 _7ba _aItō _92646 9) prove koha/t/db_dependent/Authority/Merge.t 10) Sign-off :) Sponsored-by: ENSA Works smoothly, both with UNIMARC/MARC21 Signed-off-by: Frédéric Demians <f.demians@tamil.fr> Thanks Lucas, the patch has been rebased on main. Created attachment 180178 [details] [review] Bug 38142: Choose language to report from authority to bibliographic record. Currently, when reporting authority field to the bibliographic records, the merge function uses the first matching authority field. However, authorities can provide multiple versions of the same field, with different languages, identified by a code in the $7 field. French reference for languages codes for example, but any code system could be used: https://documentation.abes.fr/sudoc/formats/unmb/DonneesCodees/CodesZone104.htm#$d This patch introduces the LanguageToReportOnMerge system preference, which allows to specify a language code. When set, if there is an authority field with the given language code, it will be used to report to the bibliographic record. If the system preference is not set, or there is no matching field, the default behavior will apply: the first field will be used. The $7 in the authority field can be in a short or long form: - the short form is 2 characters long. Example: ba - the long form is 8 characters long, with the short form in position 5 and 6 (starting from 1). Example: ba0yba0y Test plan: ---------- 1) Edit the 'Personal Name' authority type framework to make the 100 field repeatable 2) Add an authority with several languages, like this (the order is important): 100 _7ba0yda0y _a伊東豊雄 100 _7ba0yba0y _aItō 3) Add a bibliographic record linked to that authority. (Click on the "tag editor" icon next to the 100$a field in the bibliographic record to search for the authority, then Search main heading ($a only): contains Itō) Note that you can choose which version you want when searching for the authority, under the "Get it!" column For this example, choose the latin version: Itō You should now have this in your bibliographic record: 100 _aItō _92646 4) Edit the authority, you don't have to change anything, just edit and save. Note that your bibliographic record now looks like this: 100 _7ba0yda0y _a伊東豊雄 _92646 The first 100 field of the authority has been reported back to the bibliographic record, regardless of what you chose when selecting the authority. 5) Apply the patch, set the LanguageToReportOnMerge system preference to: 'ba' 6) Edit the authority again, and check that your bibliographic record now looks like this, according to the system preference: 100 _7ba0yba0y _aItō _92646 7) Edit the authority again, but set the $7 of the latin field to its short form: 100 _7ba0yda0y _a伊東豊雄 100 _7ba _aItō 8) Check that the bibliographic record still has the latin field reported: 100 _7ba _aItō _92646 9) prove koha/t/db_dependent/Authority/Merge.t 10) Sign-off :) Sponsored-by: ENSA Works smoothly, both with UNIMARC/MARC21 Signed-off-by: Frédéric Demians <f.demians@tamil.fr> Created attachment 180222 [details] [review] Bug 38142: Choose language to report from authority to bibliographic record. Currently, when reporting authority field to the bibliographic records, the merge function uses the first matching authority field. However, authorities can provide multiple versions of the same field, with different languages, identified by a code in the $7 field. French reference for languages codes for example, but any code system could be used: https://documentation.abes.fr/sudoc/formats/unmb/DonneesCodees/CodesZone104.htm#$d This patch introduces the LanguageToReportOnMerge system preference, which allows to specify a language code. When set, if there is an authority field with the given language code, it will be used to report to the bibliographic record. If the system preference is not set, or there is no matching field, the default behavior will apply: the first field will be used. The $7 in the authority field can be in a short or long form: - the short form is 2 characters long. Example: ba - the long form is 8 characters long, with the short form in position 5 and 6 (starting from 1). Example: ba0yba0y Test plan: ---------- 1) Edit the 'Personal Name' authority type framework to make the 100 field repeatable 2) Add an authority with several languages, like this (the order is important): 100 _7ba0yda0y _a伊東豊雄 100 _7ba0yba0y _aItō 3) Add a bibliographic record linked to that authority. (Click on the "tag editor" icon next to the 100$a field in the bibliographic record to search for the authority, then Search main heading ($a only): contains Itō) Note that you can choose which version you want when searching for the authority, under the "Get it!" column For this example, choose the latin version: Itō You should now have this in your bibliographic record: 100 _aItō _92646 4) Edit the authority, you don't have to change anything, just edit and save. Note that your bibliographic record now looks like this: 100 _7ba0yda0y _a伊東豊雄 _92646 The first 100 field of the authority has been reported back to the bibliographic record, regardless of what you chose when selecting the authority. 5) Apply the patch, set the LanguageToReportOnMerge system preference to: 'ba' 6) Edit the authority again, and check that your bibliographic record now looks like this, according to the system preference: 100 _7ba0yba0y _aItō _92646 7) Edit the authority again, but set the $7 of the latin field to its short form: 100 _7ba0yda0y _a伊東豊雄 100 _7ba _aItō 8) Check that the bibliographic record still has the latin field reported: 100 _7ba _aItō _92646 9) prove koha/t/db_dependent/Authority/Merge.t 10) Sign-off :) Sponsored-by: ENSA Works smoothly, both with UNIMARC/MARC21 Signed-off-by: Frédéric Demians <f.demians@tamil.fr> Signed-off-by: Martin Renvoize <martin.renvoize@openfifth.co.uk> This might work for MARC21, but I don't think it should: In MARC21 the 1xx fields are all marked as non-repeatable. And Koha's frameworks for MARC21 reflect that: https://www.loc.gov/marc/authority/ad1xx3xx.html And the $7 is used otherwise: $7 - Data provenance (R) As we usually expect the 1xx to not be repeated I am not sure what kind of problems could arise if it is. It looks like in UNIMARC authority this is OK: https://repository.ifla.org/server/api/core/bitstreams/3a3df76b-dad8-4a41-95ee-5198813784fa/content 1) I feel we should mark this as a UNIMARC feature in the system preference text. (please argue strongly or follow-up) 2) I am not sure if "reported" is clear here as the verb in both the system preference name and description. It's more like it will be the one that is used/copied to the bibliographic record if I understand correctly? (question to native speakers). (not a blocker, but might confuse translators?) 3) Please use say_success in your database updates... fixed before pushing (done) Pushed for 25.05! Well done everyone, thank you! What is this $7 format? I don't see anything like it described on https://www.loc.gov/marc/authority/adapndxh.html I don't think it exists in MARC21, see my previous comment. That's why I was asking to mark this as a UNIMARC only feature (no follow-up from developer yet...) I have to agree with the "reported" sounding weird. Perhaps it should be something like "used to select the authority record language" I nudged Matts on Mattermost. I think to be more precise it would be something like: ... the language code of the main heading to use Created attachment 182652 [details] [review] Bug 38142, Follow-up: Choose language to report from authority to bibliographic record. - Rename LanguageToReportOnMerge system preference to LanguageToUseOnMerge - Use "use" instead of "report" in system preference description - Mark this feature as UNIMARC-only. Matts, can you please double check? Your follow-up patch (only applied the second) doesn't apply cleanly on main. Also the commit title will fail QA tools ;) Created attachment 182671 [details] [review] Bug 38142, Follow-up: Choose language to report from authority to bibliographic record. - Rename LanguageToReportOnMerge system preference to LanguageToUseOnMerge - Use "use" instead of "report" in system preference description - Mark this feature as UNIMARC-only. Fixed the commit title (Bug 3814, ;)) and added a follow-up for sysprefs.sql. |