Merge pull request #40 from zaibacu/case-sensitive-spacy-rules-bugfix

Cover the case

changes/40.fix.rst

@@ -0,0 +1 @@
+Fix issue regarding Spacy rules with `IN_LIST` and using case-sensitive mode. It was creating Regex pattern which is not valid spacy pattern

rita/__init__.py

@@ -9,7 +9,7 @@
 
 logger = logging.getLogger(__name__)
 
-__version__ = (0, 3, 3, os.getenv("VERSION_PATCH"))
+__version__ = (0, 3, 4, os.getenv("VERSION_PATCH"))
 
 
 def get_version():

rita/engine/translate_spacy.py

@@ -11,7 +11,7 @@ def any_of_parse(lst, config, op=None):
                              for item in lst])
         base = {"LOWER": {"REGEX": r"({0})".format("|".join(normalized))}}
     else:
-        base = {"REGEX": r"({0})".format("|".join(sorted(lst)))}
+        base = {"TEXT": {"REGEX": r"({0})".format("|".join(sorted(lst)))}}
 
     if op:
         base["OP"] = op

rita/engine/translate_standalone.py

@@ -2,6 +2,7 @@
 import re
 
 from functools import partial
+from itertools import groupby
 
 logger = logging.getLogger(__name__)
 
@@ -120,7 +121,7 @@ def compile(self, label, rules):
 
         return re.compile(r"(?P<{0}>{1})".format(label, "".join(rules)), flags)
 
-    def execute(self, text):
+    def _results(self, text):
         for p in self.patterns:
             for match in p.finditer(text):
                 yield {
@@ -130,6 +131,16 @@ def execute(self, text):
                     "label": match.lastgroup,
                 }
 
+    def execute(self, text):
+        results = sorted(list(self._results(text)), key=lambda x: x["start"])
+        for k, g in groupby(results, lambda x: x["start"]):
+            group = list(g)
+            if len(group) == 1:
+                yield group[0]
+            else:
+                data = sorted(group, key=lambda x: -x["end"])
+                yield data[0]
+
 
 def compile_rules(rules, config):
     logger.info("Using standalone rule implementation")

tests/test_examples.py

@@ -104,6 +104,29 @@ def test_exclude_word(engine):
     assert len(r2) == 0
 
 
+@pytest.mark.parametrize('engine', [spacy_engine, standalone_engine])
+def test_case_sensitive(engine):
+    parser = engine(
+        """
+        !CONFIG("ignore_case", "N")
+
+        variants = {"Bitcoin", "BTC", "Bitcoin Cash"}
+
+        {IN_LIST(variants)}->MARK("CRYPTO")
+        """
+    )
+
+    text = """
+    A bitcoin mining magnate has proposed a new development fund for Bitcoin Cash.
+    According to BTC.TOP CEO Jiang Zhuoer, the scheme will 'tax' Bitcoin Cash mining rewards
+    in an effort to increase funding for Bitcoin Cash infrastructure.
+    """
+
+    results = parser(text)
+    print(results)
+    assert results[0] == ("Bitcoin Cash", "CRYPTO")
+
+
 @pytest.mark.parametrize('engine', [spacy_engine, standalone_engine])
 def test_benchmark(benchmark, engine, bench_text):
     """