[feat] add some more or less uninformative expressions from academic genres to...

[feat] add some more or less uninformative expressions from academic genres to the base stoplist (bug-17)

gargantext/util/toolchain/list_stop.py

@@ -27,7 +27,6 @@ def is_stop_word(ngram, stop_words=None):
             # , "(.*)(\.)(.*)"         trop fort (enlève les sigles !)
             , "(.*)(\,)(.*)"
             , "(.*)(< ?/?p ?>)(.*)"       # marques de paragraphes
-            , "(.*)(study|elsevier)(.*)"
             , "(.*)\b(xx|xi|xv)\b(.*)"
             , "(.*)(result)(.*)"
             , "(.*)(year|année|nombre|moitié)(.*)"
@@ -37,6 +36,87 @@ def is_stop_word(ngram, stop_words=None):
             , "(.*)(terme)(.*)"
             , "(.*)(différent)(.*)"
             , "(.*)(travers)(.*)"
+            # academic stamps
+            , ".*\belsevier\b.*"
+            , ".*\bwiley\b.*)"
+            , ".*\bspringer\b.*"
+            , ".*university press\b.*"
+            , ".*\bstudy\b.*"
+            # academic terms when alone ~~> usually not informative
+            , "hypothes[ie]s$"
+            , "analys[ie]s$"
+            , "bas[ie]s$"
+            , "online$"
+            , "importance$"
+            , "uses?$"
+            , "cases?$"
+            , "effects?$"
+            , "times?$"
+            , "methods?$"
+            , "types?$"
+            , "evidences?$"
+            , "findings?$"
+            , "relations?$"
+            , "terms?$"
+            , "procedures?$"
+            , "factors?$"
+            , "reports?$"
+            , "changes?$"
+            , "facts?$"
+            , "others?$"
+            , "applications?$"
+            , "periods?$"
+            , "investigations?$"
+            , "orders?$"
+            , "forms?$"
+            , "conditions?$"
+            , "situations?$"
+            , "papers?$"
+            , "relationships?$"
+            , "values?$"
+            , "areas?$"
+            , "techniques?$"
+            , "means?$"
+            , "conclusions?$"
+            , "comparisons?$"
+            , "parts?$"
+            , "amounts?$"
+            , "aims?$"
+            , "lacks?$"
+            , "issues?$"
+            , "ways?$"
+            , "ranges?$"
+            , "models?$"
+            , "articles?$"
+            , "series?$"
+            , "totals?$"
+            , "influences?$"
+            , "journals?$"
+            , "rules?$"
+            , "persons?$"
+            , "abstracts?$"
+            , "(?:book)? reviews?$"
+            , "process(?:es)?$"
+            , "approach(?:es)?$"
+            , "theor(?:y|ies)?$"
+            , "methodolog(?:y|ies)?$"
+            , "similarit(?:y|ies)?$"
+            , "possibilit(?:y|ies)?$"
+            , "stud(?:y|ies)?$"
+            # non-thematic or non-NP expressions
+            , "none$"
+            , "other(?: hand)?$"
+            , "whereas$"
+            , "usually$"
+            , "and$"
+            # , "vol$"
+            , "eds?$"
+            , "ltd$"
+            , "copyright$"
+            , "e-?mails?$"
+            , ".*="
+            , "=.*"
+            , "further(?:more)?$"
             , "(.*)(:|\|)(.*)"
             ] :
         compiled_regexes.append(compile(regex))