<< | Index | >>
Why have the following FI have too long lines?
I consider this problem was resolved when I fixed the problem with the missing spaces before the Wirkstoff, as there was an inbalanced </span> present. After running the import on my own VM the page displayed correctly, see
Running an import for jobs/update_textinfo_swissmedicinfo no-download --target=fi --reparse Crucell
gives the following output
textinfo_swissmedicinfo_company_index returns 467,704. See /var/www/oddb.org/data/html/Crucell_index.html import_swissmedicinfo_by_iksnrs [467, 704] import_swissmedicinfo_by_iksnrs iksnr 467 {} parse_and_update: calls parse_fachinfo, /var/www/oddb.org/data/html/fachinfo/de/Accupaque__swissmedicinfo.html, name Accupaque™ de title Accupaque™, styles p{margin-top:0pt;margin-right:0pt;margin-bottom:0pt;
Looking at the XML-file wie expect the following IDs
00467 1 Vivotif, Kapseln CRUCELL SWITZERLAND AG 08.08. J07AP01 00572 1 Epaxal, Injektionssuspension CRUCELL SWITZERLAND AG 08.08. J07BC02 00614 12 Inflexal V, Injektionssuspension CRUCELL SWITZERLAND AG 08.08. J07BB02 00668 1 Gamunex 10%, Lösung zur intravenösen Injektion CRUCELL SWITZERLAND AG 08.09. J06BA02 00697 1 Prolastin, Trockensubstanz zur Herstellung einer Infusionslösung CRUCELL SWITZERLAND AG 06.01.1. B02AB02 00704 2 Dukoral, orale Suspension CRUCELL SWITZERLAND AG 08.08. J07AE01
sprintf("%05d", id)
instead of id
Still having the problem that calling jobs/update_textinfo_swissmedicinfo no-download --target=both --reparse 467
returns wrong patinfo 56079
import_swissmedicinfo_by_iksnrs ["00467"] target both import_swissmedicinfo_by_iksnrs iksnr "00467" {} parse_and_update: calls parse_fachinfo, /var/www/oddb.org/data/html/fachinfo/de/Vivotif__swissmedicinfo.html, name Vivotif® de title Vivotif®, styles p{margin-top:0pt;margin-right:0pt;margin-bottom:0pt;margin-left:0pt; parse_and_update: update_fachinfo name Vivotif® returned ["00467"] parse_and_update: calls parse_fachinfo, /var/www/oddb.org/data/html/fachinfo/fr/Vivotif__swissmedicinfo.html, name Vivotif® fr title Vivotif®, styles p{margin-top:0pt;margin-right:0pt;margin-bottom:0pt;margin-left:0pt; parse_and_update: update_fachinfo name Vivotif® returned ["00467"] parse_and_update: calls parse_patinfo, /var/www/oddb.org/data/html/patinfo/de/Arcoxia__swissmedicinfo.html, name Arcoxia® de title Arcoxia®, styles p{margin-top:0pt;margin-right:0pt;margin-bottom:0pt;margin-left:0pt; parse_and_update: update_patinfo name Arcoxia® returned ["56079"] parse_and_update: calls parse_patinfo, /var/www/oddb.org/data/html/patinfo/fr/Arcoxia__swissmedicinfo.html, name Arcoxia® fr title Arcoxia®, styles p{margin-top:0pt;margin-right:0pt;margin-bottom:0pt;margin-left:0pt; parse_and_update: update_patinfo name Arcoxia® returned ["56079"]
Wrote a simple test script to show that the procedure extract_matched_name
in text_info.rb does not work correctly. There are no test-cases present.
Added the test to test/test_plugin/text_info.rb
.
The import of Bayer seems to be correct now, but I got another time the error @|@Message: ERROR: prepared statement "ruby-dbi:Pg:5950206401373984273.0981805" already exists@@
Pushed the commit Fix import swissmedic by companies which make the following changes
Following curious situation found. There are 4 iksnrs in the XML consisting of a single '-', e.g for Sérocytol®
. On the swissmedic-HTML you find 00332, 00328, 00278, 00279, 00277, 00282, 00284, 00286, 00288, 00294, 00295, 00296, 00300, 00306, 00309, 00313, 60026, 00314, 00315, 00316, 00299, 00330, 00334, 00335, 00338, 00339 (Swissmedic).
It might be more consistent to get the iksnr of each company using also the AipsDownload_latest.xml instead of using mechanize to get the information from http://www.swissmedicinfo.ch.