Avoid replacing DOIs with shorter ones

[lfoppiano]

When we search for a DOI in the page, the regex may truncate DOIs that are split by a breakline, so this PR proposes a simple fix that is to substitute the DOI only when the one found in the page is larger than the one extracted by the header parser

[coveralls]

coverage: 40.786% (-0.001%) from 40.787%
when pulling 7f1d15d on bugfix/fix-doi-search
into 694f0ed on master.

[coveralls]

coverage: 40.396% (+0.002%) from 40.394%
when pulling 7f5b55e on bugfix/fix-doi-search
into 01fe109 on master.

[coveralls]

Pull Request Test Coverage Report for Build 21495896701

Details

• 0 of 0 changed or added relevant lines in 0 files are covered.
• 1028 unchanged lines in 4 files lost coverage.
• Overall coverage increased (+0.02%) to 38.23%

---

Files with Coverage Reduction	New Missed Lines	%
org/grobid/core/document/Document.java	100	72.97%
org/grobid/core/engines/HeaderParser.java	112	56.47%
org/grobid/core/utilities/TextUtilities.java	199	53.1%
org/grobid/core/data/BiblioItem.java	617	56.95%

Totals
Change from base Build 21386310485:	0.02%
Covered Lines:	17237
Relevant Lines:	42511

---

💛 - Coveralls

[lfoppiano]

This fix, does not solves the main issue described in #1126 when the page are concatenated to the DOI. it sems naturally disappearing with #1142 after adding more training data.