BatchalignHK 0.7.19__tar.gz → 0.7.19.post2__tar.gz

{batchalignhk-0.7.19 → batchalignhk-0.7.19.post2}/BatchalignHK.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: BatchalignHK
-Version: 0.7.19
+Version: 0.7.19.post2
 Summary: Python Speech Language Sample Analysis
 Author: Brian MacWhinney, Houjun Liu
 Author-email: macw@cmu.edu, houjun@cmu.edu
@@ -35,12 +35,16 @@ Requires-Dist: tiktoken
 Requires-Dist: blobfile
 Requires-Dist: sentencepiece
 Requires-Dist: tencentcloud-sdk-python-common
-Requires-Dist: tencentcloud-sdk-python-asrgoogletrans
+Requires-Dist: tencentcloud-sdk-python-asr
+Requires-Dist: googletrans
 Requires-Dist: openai-whisper>=20240930
 Provides-Extra: dev
 Requires-Dist: pytest; extra == "dev"
 Provides-Extra: train
 Requires-Dist: accelerate~=0.27; extra == "train"
+Provides-Extra: docs
+Requires-Dist: mkdocs-material; extra == "docs"
+Requires-Dist: mkdocs-click; extra == "docs"
 Dynamic: author
 Dynamic: author-email
 Dynamic: classifier

{batchalignhk-0.7.19 → batchalignhk-0.7.19.post2}/BatchalignHK.egg-info/SOURCES.txt

@@ -117,4 +117,5 @@ batchalign/tests/pipelines/fa/test_fa_pipeline.py
 batchalign/utils/__init__.py
 batchalign/utils/config.py
 batchalign/utils/dp.py
+batchalign/utils/names.py
 batchalign/utils/utils.py

{batchalignhk-0.7.19 → batchalignhk-0.7.19.post2}/BatchalignHK.egg-info/requires.txt

@@ -25,11 +25,16 @@ tiktoken
 blobfile
 sentencepiece
 tencentcloud-sdk-python-common
-tencentcloud-sdk-python-asrgoogletrans
+tencentcloud-sdk-python-asr
+googletrans
 openai-whisper>=20240930
 
 [dev]
 pytest
 
+[docs]
+mkdocs-material
+mkdocs-click
+
 [train]
 accelerate~=0.27

{batchalignhk-0.7.19 → batchalignhk-0.7.19.post2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: BatchalignHK
-Version: 0.7.19
+Version: 0.7.19.post2
 Summary: Python Speech Language Sample Analysis
 Author: Brian MacWhinney, Houjun Liu
 Author-email: macw@cmu.edu, houjun@cmu.edu
@@ -35,12 +35,16 @@ Requires-Dist: tiktoken
 Requires-Dist: blobfile
 Requires-Dist: sentencepiece
 Requires-Dist: tencentcloud-sdk-python-common
-Requires-Dist: tencentcloud-sdk-python-asrgoogletrans
+Requires-Dist: tencentcloud-sdk-python-asr
+Requires-Dist: googletrans
 Requires-Dist: openai-whisper>=20240930
 Provides-Extra: dev
 Requires-Dist: pytest; extra == "dev"
 Provides-Extra: train
 Requires-Dist: accelerate~=0.27; extra == "train"
+Provides-Extra: docs
+Requires-Dist: mkdocs-material; extra == "docs"
+Requires-Dist: mkdocs-click; extra == "docs"
 Dynamic: author
 Dynamic: author-email
 Dynamic: classifier

{batchalignhk-0.7.19 → batchalignhk-0.7.19.post2}/batchalign/cli/cli.py

@@ -79,6 +79,7 @@ def handle_verbosity(verbosity):
 
 @click.group()
 @click.pass_context
+@click.version_option(VERSION_NUMBER)
 @click.option("-v", "--verbose", type=int, count=True, default=0, help="How loquacious Batchalign should be.")
 def batchalign(ctx, verbose):
     """process .cha and/or audio files in IN_DIR and dumps them to OUT_DIR using recipe COMMAND"""

{batchalignhk-0.7.19 → batchalignhk-0.7.19.post2}/batchalign/pipelines/analysis/eval.py

@@ -10,10 +10,53 @@ from batchalign.pipelines.asr.utils import *
 from batchalign.utils.config import config_read
 
 from batchalign.utils.dp import align, ExtraType, Extra, Match
+from batchalign.utils.names import names
 
 import logging
 L = logging.getLogger("batchalign")
 
+def conform(x):
+    result = []
+    for i in x:
+        if "'s" in i.strip():
+            result.append(i.split("'")[0])
+            result.append("is")
+        elif "americanstyle" == i.strip():
+            result.append("american")
+            result.append("style")
+        elif "postwar" == i.strip():
+            result.append("post")
+            result.append("war")
+        elif "postwar" == i.strip():
+            result.append("post")
+            result.append("war")
+        elif i.strip() in names:
+            result.append("name")
+        elif "dunno" == i.strip():
+            result.append("don't")
+            result.append("know")
+        elif "wanna" == i.strip():
+            result.append("want")
+            result.append("to")
+        elif "gotta" == i.strip():
+            result.append("got")
+            result.append("to")
+        elif "farmhouse" == i.strip():
+            result.append("farm")
+            result.append("house")
+        elif "aa" == i.strip():
+            result.append("a")
+            result.append("a")
+        elif "hmm" == i.strip():
+            result.append("hm")
+        elif "_" in i.strip():
+            for j in i.strip().split("_"):
+                result.append(j)
+        else:
+            result.append(i)
+
+    return result
+
 class EvaluationEngine(BatchalignEngine):
     tasks = [ Task.WER ]
 
@@ -64,6 +107,9 @@ class EvaluationEngine(BatchalignEngine):
             gold_final = gold_forms
             forms_final = forms_finished
 
+        gold_final = conform(gold_final)
+        forms_final = conform(forms_final)
+
         # dp!
         alignment = align(forms_final, gold_final, False)