use EarleyLM in tests/test_inference.py

genparse/cfglm.py

@@ -7,6 +7,7 @@
 from genparse.cfg import CFG, _gen_nt
 from genparse.lm import LM
 from genparse.semiring import Boolean, Float
+from genparse.experimental.cky import IncrementalCKY
 
 # EOS = '$EOS'
 # EOS = '🛑'
@@ -105,9 +106,6 @@ def __init__(self, cfg, **kwargs):
 
         self.cfg = cfg
         self.pfg = self.cfg.cnf.prefix_grammar.cnf
-
-        from genparse.experimental.cky import IncrementalCKY
-
         self.model = IncrementalCKY(self.pfg, **kwargs)
 
         super().__init__(V=cfg.V, eos=EOS)

genparse/experimental/earley.py

@@ -6,29 +6,35 @@
 # from arsenal.datastructures.pdict import pdict
 from arsenal.datastructures.heap import LocatorMaxHeap
 
-from genparse.cfglm import EOS, add_EOS
+from genparse.cfglm import EOS, add_EOS, locally_normalize, CFG
 from genparse.linear import WeightedGraph
 from genparse.lm import LM
 from genparse.semiring import Boolean
+from genparse import Float
 
 
 class EarleyLM(LM):
     def __init__(self, cfg):
         if EOS not in cfg.V:
             cfg = add_EOS(cfg)
+        self.cfg = cfg  # Note: <- cfg before prefix transform & normalization!
         self.model = Earley(cfg.prefix_grammar)
         super().__init__(V=cfg.V, eos=EOS)
 
     def p_next(self, context):
         return self.model.p_next(context)
 
-    #    def __call__(self, context):
-    #        assert context[-1] == EOS
-    #        return self.p_next(context[:-1])[EOS]
+    def __call__(self, context):
+        assert context[-1] == EOS
+        return self.model(context)
 
     def clear_cache(self):
         self.model.clear_cache()
 
+    @classmethod
+    def from_string(cls, x, semiring=Float, **kwargs):
+        return cls(locally_normalize(CFG.from_string(x, semiring), **kwargs))
+
 
 class Column:
     __slots__ = ('k', 'i_chart', 'c_chart', 'waiting_for', 'Q')

tests/test_inference.py

@@ -2,7 +2,8 @@
 import pandas as pd
 from arsenal import colors
 
-from genparse.cfglm import CFGLM
+# from genparse.cfglm import CFGLM
+from genparse.experimental.earley import EarleyLM as CFGLM
 from genparse.semiring import Float
 from genparse.steer import BruteForceGlobalProductOfExperts, run