batchalign 0.7.22.post5__tar.gz → 0.7.22.post8__tar.gz

{batchalign-0.7.22.post5/batchalign.egg-info → batchalign-0.7.22.post8}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: batchalign
-Version: 0.7.22.post5
+Version: 0.7.22.post8
 Summary: Python Speech Language Sample Analysis
 Author: Brian MacWhinney, Houjun Liu
 Author-email: macw@cmu.edu, houjun@cmu.edu

{batchalign-0.7.22.post5 → batchalign-0.7.22.post8}/batchalign/cli/cli.py

@@ -179,7 +179,7 @@ def transcribe(ctx, in_dir, out_dir, lang, num_speakers, **kwargs):
 
     def writer(doc, output):
         doc.content.insert(0, CustomLine(id="Comment", type=CustomLineType.INDEPENDENT,
-                                         content=f"Batchalign {VERSION_NUMBER.strip()}, ASR Engine {asr}. Unchecked output of ASR model; do not use."))
+                                         content=f"Batchalign {VERSION_NUMBER.strip()}, ASR Engine {asr}. Unchecked output of ASR model."))
         CHATFile(doc=doc).write(output
                                 .replace(".wav", ".cha")
                                 .replace(".WAV", ".cha")
@@ -319,6 +319,8 @@ def utseg(ctx, in_dir, out_dir, lang, num_speakers, **kwargs):
 @common_options
 @click.option("--whisper/--rev",
               default=False, help="Use OpenAI Whisper (ASR) instead of Rev.AI (default).")
+@click.option("--whisper_oai/--rev",
+              default=False, help="Use the OpenAI's Whisper implementation instead of Rev.AI (default).")
 @click.option("--lang",
               help="sample language in three-letter ISO 3166-1 alpha-3 code",
               show_default=True,
@@ -328,7 +330,7 @@ def utseg(ctx, in_dir, out_dir, lang, num_speakers, **kwargs):
 @click.option("--wor/--nowor",
               default=False, help="Should we write word level alignment line? Default to no.")
 @click.pass_context
-def benchmark(ctx, in_dir, out_dir, lang, num_speakers, whisper, **kwargs):
+def benchmark(ctx, in_dir, out_dir, lang, num_speakers, whisper, whisper_oai, **kwargs):
     """Benchmark ASR utilities for their word accuracy"""
     def loader(file):
         # try to find a .cha in the same directory
@@ -355,7 +357,8 @@ def benchmark(ctx, in_dir, out_dir, lang, num_speakers, whisper, **kwargs):
     _dispatch("benchmark", lang, num_speakers, ["mp3", "mp4", "wav"], ctx,
               in_dir, out_dir,
               loader, writer, C,
-              asr="whisper" if whisper else "rev", **kwargs)
+              asr="whisper" if whisper else ("whisper_oai " if whisper_oai else "rev"),
+              **kwargs)
     
 
 #################### AVQI ################################

{batchalign-0.7.22.post5 → batchalign-0.7.22.post8}/batchalign/pipelines/analysis/eval.py

@@ -25,7 +25,7 @@ def conform(x):
     result = []
     for i in x:
         if i.strip().lower() in joined_compounds:
-            for k in compounds[joined_compounds.index(i.strip())]:
+            for k in compounds[joined_compounds.index(i.strip().lower())]:
                 result.append(k)
         elif i.strip() in lowered_abbrev:
             for j in i.strip():

batchalign-0.7.22.post8/batchalign/version

@@ -0,0 +1,3 @@
+0.7.22-post.8
+October 21st, 2025
+oai whisper for benchmakring

{batchalign-0.7.22.post5 → batchalign-0.7.22.post8/batchalign.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: batchalign
-Version: 0.7.22.post5
+Version: 0.7.22.post8
 Summary: Python Speech Language Sample Analysis
 Author: Brian MacWhinney, Houjun Liu
 Author-email: macw@cmu.edu, houjun@cmu.edu

batchalign-0.7.22.post5/batchalign/version

@@ -1,3 +0,0 @@
-0.7.22-post.5
-October 19th, 2025
-Bumped AVQI calculation method