batchalign 0.7.19.post3__tar.gz → 0.7.19.post5__tar.gz

{batchalign-0.7.19.post3/batchalign.egg-info → batchalign-0.7.19.post5}/PKG-INFO

@@ -1,6 +1,6 @@
-Metadata-Version: 2.4
+Metadata-Version: 2.2
 Name: batchalign
-Version: 0.7.19.post3
+Version: 0.7.19.post5
 Summary: Python Speech Language Sample Analysis
 Author: Brian MacWhinney, Houjun Liu
 Author-email: macw@cmu.edu, houjun@cmu.edu
@@ -25,7 +25,7 @@ Requires-Dist: rich~=13.6
 Requires-Dist: click~=8.1
 Requires-Dist: matplotlib<4.0.0,>=3.8.0
 Requires-Dist: pyfiglet==1.0.2
-Requires-Dist: setuptools
+Requires-Dist: setuptools>=78.1.1
 Requires-Dist: soundfile~=0.12.0
 Requires-Dist: rich-click>=1.7.0
 Requires-Dist: typing-extensions
@@ -47,7 +47,6 @@ Dynamic: author-email
 Dynamic: classifier
 Dynamic: description
 Dynamic: description-content-type
-Dynamic: license-file
 Dynamic: provides-extra
 Dynamic: requires-dist
 Dynamic: summary

{batchalign-0.7.19.post3 → batchalign-0.7.19.post5}/batchalign/document.py

@@ -11,6 +11,7 @@ from pathlib import Path
 
 from batchalign.errors import *
 from batchalign.constants import *
+from batchalign.utils.abbrev import abbrev
 
 import re
 
@@ -223,6 +224,11 @@ class Utterance(BaseModel):
         t = re.sub(r",", " , ", t.strip()).strip()
         t = re.sub(r" +", " ", t.strip()).strip()
         t = t.replace("+ ,", "+,").strip()
+
+        abbrevs =  [" " .join(list(i)) for i in abbrev]
+        for i in abbrevs:
+            t = t.replace(i, i.replace(" ", ""))
+
         return t
 
     def __repr__(self):

{batchalign-0.7.19.post3 → batchalign-0.7.19.post5}/batchalign/formats/chat/generator.py

@@ -101,7 +101,7 @@ def generate_chat_utterance(utterance: Utterance, special_mor=False, write_wor=T
         result.append("%wor:\t"+" ".join(wor_elems))
     if has_coref:
         result.append("%coref:\t"+(", ".join(coref_elems)))
-    if utterance.translation != None:
+    if utterance.translation != None and utterance.translation.strip() not in ["", ".", "!", "?"]:
         result.append("%xtra:\t"+utterance.translation)
 
 

{batchalign-0.7.19.post3 → batchalign-0.7.19.post5}/batchalign/pipelines/morphosyntax/ud.py

@@ -711,6 +711,7 @@ def adlist_postprocessor(i, lang, adlist):
 
 ######
 def morphoanalyze(doc: Document, retokenize:bool, skipmultilang:bool, status_hook:callable = None, **kwargs):
+
     L.debug("Starting Stanza...")
     inputs = []
 

batchalign-0.7.19.post5/batchalign/utils/abbrev.py

@@ -0,0 +1,182 @@
+abbrev = [
+    "FBI",
+    "CIA",
+    "NSA",
+    "NATO",
+    "UN",
+    "WHO",
+    "NASA",
+    "CDC",
+    "IRS",
+    "EPA",
+    "HTTP",
+    "URL",
+    "HTML",
+    "CSS",
+    "API",
+    "IP",
+    "DNS",
+    "SQL",
+    "USB",
+    "VPN",
+    "ATT",
+    "AT&T",
+    "CEO",
+    "CFO",
+    "COO",
+    "IPO",
+    "ROI",
+    "GDP",
+    "LLC",
+    "HR",
+    "M&",
+    "KPI",
+    "GPA",
+    "SAT",
+    "ACT",
+    "MBA",
+    "PhD",
+    "BA",
+    "MA",
+    "STEM",
+    "ESL",
+    "GED",
+    "AWOL",
+    "MIA",
+    "POW",
+    "IED",
+    "UAV",
+    "RPG",
+    "NATO",
+    "SEAL",
+    "JAG",
+    "ROTC",
+    "CERN",
+    "GMO",
+    "H2O",
+    "CO2",
+    "UV",
+    "IR",
+    "AI",
+    "VR",
+    "AR",
+    "NPR",
+    "BBC",
+    "MTV",
+    "CNN",
+    "HBO",
+    "ESPN",
+    "TMZ",
+    "AMC",
+    "IMAX",
+    "WWE",
+    "ASAP",
+    "DIY",
+    "ETA",
+    "RSVP",
+    "FYI",
+    "LOL",
+    "BRB",
+    "IDK",
+    "BTW",
+    "TMI",
+    "PBJ",
+    "AIDS",
+    "HIV",
+    "ADHD",
+    "COPD",
+    "PTSD",
+    "CHF",
+    "CAD",
+    "TB",
+    "UTI",
+    "GERD",
+    "MRI",
+    "CT",
+    "ECG",
+    "EEG",
+    "CBC",
+    "BMP",
+    "ABG",
+    "PFT",
+    "FOBT",
+    "ENT",
+    "OB",
+    "PCP",
+    "ICU",
+    "NICU",
+    "ER",
+    "OR",
+    "PT",
+    "OT",
+    "EM",
+    "OTC",
+    "NSAID",
+    "IV",
+    "IM",
+    "SC",
+    "PRN",
+    "BID",
+    "TID",
+    "QID",
+    "NPO",
+    "CNS",
+    "PNS",
+    "GI",
+    "GU",
+    "CV",
+    "MSK",
+    "ENT",
+    "BMI",
+    "BMR",
+    "BP",
+    "WBC",
+    "RBC",
+    "HGB",
+    "HCT",
+    "PLT",
+    "ESR",
+    "CRP",
+    "LFT",
+    "TFT",
+    "INR",
+    "MMR",
+    "DPT",
+    "HPV",
+    "Tdap",
+    "BCG",
+    "IPV",
+    "HBV",
+    "HAV",
+    "HCV",
+    "RSV",
+    "SOAP",
+    "DNR",
+    "AMA",
+    "LOS",
+    "EHR",
+    "EMR",
+    "ICD",
+    "CPT",
+    "HIPAA",
+    "HR",
+    "RR",
+    "SpO2",
+    "MAP",
+    "GFR",
+    "A1C",
+    "LDL",
+    "HDL",
+    "TG",
+    "BUN",
+    "SIDS",
+    "DVT",
+    "PE",
+    "ARDS",
+    "SLE",
+    "RA",
+    "TIA",
+    "CVA",
+    "ALS",
+    "MS",
+]

batchalign-0.7.19.post5/batchalign/version

@@ -0,0 +1,3 @@
+0.7.19-post.5
+May 23th, 2025
+abbreviations

{batchalign-0.7.19.post3 → batchalign-0.7.19.post5/batchalign.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
-Metadata-Version: 2.4
+Metadata-Version: 2.2
 Name: batchalign
-Version: 0.7.19.post3
+Version: 0.7.19.post5
 Summary: Python Speech Language Sample Analysis
 Author: Brian MacWhinney, Houjun Liu
 Author-email: macw@cmu.edu, houjun@cmu.edu
@@ -25,7 +25,7 @@ Requires-Dist: rich~=13.6
 Requires-Dist: click~=8.1
 Requires-Dist: matplotlib<4.0.0,>=3.8.0
 Requires-Dist: pyfiglet==1.0.2
-Requires-Dist: setuptools
+Requires-Dist: setuptools>=78.1.1
 Requires-Dist: soundfile~=0.12.0
 Requires-Dist: rich-click>=1.7.0
 Requires-Dist: typing-extensions
@@ -47,7 +47,6 @@ Dynamic: author-email
 Dynamic: classifier
 Dynamic: description
 Dynamic: description-content-type
-Dynamic: license-file
 Dynamic: provides-extra
 Dynamic: requires-dist
 Dynamic: summary

{batchalign-0.7.19.post3 → batchalign-0.7.19.post5}/batchalign.egg-info/SOURCES.txt

@@ -114,6 +114,7 @@ batchalign/tests/pipelines/cleanup/test_disfluency.py
 batchalign/tests/pipelines/cleanup/test_parse_support.py
 batchalign/tests/pipelines/fa/test_fa_pipeline.py
 batchalign/utils/__init__.py
+batchalign/utils/abbrev.py
 batchalign/utils/config.py
 batchalign/utils/dp.py
 batchalign/utils/names.py

{batchalign-0.7.19.post3 → batchalign-0.7.19.post5}/batchalign.egg-info/requires.txt

@@ -15,7 +15,7 @@ rich~=13.6
 click~=8.1
 matplotlib<4.0.0,>=3.8.0
 pyfiglet==1.0.2
-setuptools
+setuptools>=78.1.1
 soundfile~=0.12.0
 rich-click>=1.7.0
 typing-extensions

{batchalign-0.7.19.post3 → batchalign-0.7.19.post5}/setup.py

@@ -47,7 +47,7 @@ setup(
         "click~=8.1",
         "matplotlib>=3.8.0,<4.0.0",
         "pyfiglet==1.0.2",
-        "setuptools",
+        "setuptools>=78.1.1",
         "soundfile~=0.12.0",
         "rich-click>=1.7.0",
         "typing-extensions",

batchalign-0.7.19.post3/batchalign/version

@@ -1,3 +0,0 @@
-0.7.19-post.3
-May 12th, 2025
-names