PyPI - batchalign - Versions diffs - 0.7.6a13__tar.gz → 0.7.6a15__tar.gz - Mend

@@ -213,11 +213,15 @@ def handler__NOUN(word, lang=None):
     if word.deprel == "obj" and case.strip() == "":
         case = "Acc"
+    ger = ""
+    if word.text.endswith("ing") and lang == "en":
+        ger += "-Ger"
     # clear defaults
     if gender_str == "-Com,Neut" or gender_str == "-Com" or gender_str == "-ComNeut": gender_str=""
     if number_str == "-Sing": number_str=""
-    return handler(word, lang)+gender_str+number_str+stringify_feats(case, type)
+    return handler(word, lang)+gender_str+number_str+stringify_feats(case, type)+ger
 def handler__PROPN(word, lang=None):
     # code as noun
@@ -872,6 +876,8 @@ def morphoanalyze(doc: Document, retokenize:bool, status_hook:callable = None, *
                 for i,j in enumerate(ut):
                     for k in j.text:
                         ud_chars.append(ReferenceTarget(k, payload=i))
+                creaky = False
+                collected = ""
                 # brrr
                 aligned = align(chunks_chars, ud_chars, tqdm=False)
                 for i in aligned:
@@ -879,8 +885,14 @@ def morphoanalyze(doc: Document, retokenize:bool, status_hook:callable = None, *
                         if i.reference_payload not in chunks_backplate[i.payload]:
                             chunks_backplate[i.payload].append(i.reference_payload)
                     elif isinstance(i, Extra) and i.extra_type == ExtraType.PAYLOAD:
-                        # just put it back
-                        chunks_backplate[i.payload].append(i.key)
+                        if i.key == "*":
+                            creaky = not creaky
+                            chunks_backplate[i.payload].append("*"+collected+"*")
+                            collected = ""
+                        elif creaky:
+                            collected += i.key
+                        elif not creaky:
+                            chunks_backplate[i.payload].append(i.key)
                     # we want to replace the morphology of forms that are not actually
                     # supposed to be analyzed
                     elif isinstance(i, Extra) and i.extra_type == ExtraType.REFERENCE:
@@ -915,6 +927,8 @@ def morphoanalyze(doc: Document, retokenize:bool, status_hook:callable = None, *
                 retokenized_ut = retokenized_ut.replace(" ↑", "↑")
                 retokenized_ut = re.sub(r"@ ?w ?p", "@wp", retokenized_ut)
                 retokenized_ut = retokenized_ut.replace(" @", "@")
+                retokenized_ut = re.sub(r"\*[* ]*", "*", retokenized_ut)
+                retokenized_ut = re.sub(r"\*(.*?)\*", r"*\1* ", retokenized_ut)
                 # pray to everyone that it works---this will simply crash and ignore
                 # the utterance if it didn't work, so we are doing this as a sanity
                 # check rather than needing the parsed result

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: batchalign
-Version: 0.7.6a13
+Version: 0.7.6a15
 Summary: Python Speech Language Sample Analysis
 Author: Brian MacWhinney, Houjun Liu
 Author-email: macw@cmu.edu, houjun@cmu.edu

@@ -0,0 +1,3 @@
+0.7.6-alpha.15
+October 16, 2024
+gerund support

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: batchalign
-Version: 0.7.6a13
+Version: 0.7.6a15
 Summary: Python Speech Language Sample Analysis
 Author: Brian MacWhinney, Houjun Liu
 Author-email: macw@cmu.edu, houjun@cmu.edu

@@ -1,3 +0,0 @@
-0.7.6-alpha.13
-October 13, 2024
-patch bug regarding comma structure

batchalign 0.7.6a13__tar.gz → 0.7.6a15__tar.gz

batchalign 0.7.6a13tar.gz → 0.7.6a15tar.gz