Update euctr regex

[maia-sh]

Use lookarounds to avoid doi false positives
See https://s-quest.bihealth.org/ScreenITQA/#trialidentifier-results

[maia-sh]

@ontogenerator: could you please point me to/send the false positives you found in the ScreenITQA project?

[ontogenerator]

@maia-sh:

I have a list of DOIs that contain dubious trial numbers detected as discrepant between the two different versions (preprint pipeline vs screening only the pdf (without the first page of text)). The scripts ran on the text files converted via the pipeline, not on tei.xml files. Some of these false positives were grant numbers or accession numbers given in supplemental tables, while the majority were EUDRA numbers falsely detected from dois in the reference section (see attached image):

10.1101/2022.04.06.22273535
10.1101/2022.04.07.487460
10.1101/2022.04.08.22273602
10.1101/2022.04.13.22273832
10.1101/2022.04.14.22273903
10.1101/2022.04.19.488843
10.1101/2022.04.29.22274481
10.1101/2022.04.29.22274486
10.1101/2022.05.06.22274771
10.1101/2022.05.26.22275631
10.1101/2022.06.01.494373
10.1101/2022.06.02.494567
10.1101/2022.06.06.494494
10.1101/2022.06.07.22276117

[maia-sh]

Another example of false positive euctr preceeded by Chinese characters translating to funding:

doi	trn_detected
10.3760/cma.j.cn121430-20200528-00485	201100310400
10.3760/cma.j.cn121430-20200528-00485	201100310500
10.3760/cma.j.cn121430-20210514-00714	202008040003

10 3760/cma j cn121430-20200528-00485

10 3760/cma j cn121430-20210514-00714