added manifest

run

@@ -4,14 +4,16 @@ import argparse
 import datetime
 import logging
 import os
+import sys
 
 from transcribe import aws, google, whisper
 
 parser = argparse.ArgumentParser(
     prog="run", description="Run transcription generation for sample data"
 )
 
-parser.add_argument("--output_dir", help="Path to a directory to write results")
+parser.add_argument("--output-dir", help="Path to a directory to write results")
+parser.add_argument("--manifest", default="data.csv", help="Path to data manifest CSV")
 parser.add_argument(
     "--only",
     choices=["whisper", "preprocessing", "aws", "google"],
@@ -27,6 +29,10 @@ if output_dir is None:
 if not os.path.isdir(output_dir):
     os.makedirs(output_dir)
 
+# ensure manifest CSV exists
+if not os.path.isfile(args.manifest):
+    sys.exit(f"manifest file {args.manifest} doesn't exist")
+
 logging.basicConfig(
     filename=os.path.join(output_dir, "transcribe.log"),
     filemode="a",
@@ -35,18 +41,20 @@ logging.basicConfig(
     level=logging.INFO,
 )
 
-
 # run one of the transcription types individually or run them all
 if args.only == "whisper":
-    whisper.run(output_dir)
+    whisper.run(output_dir, args.manifest)
 elif args.only == "preprocessing":
-    whisper.run_preprocessing(output_dir)
+    whisper.run_preprocessing(output_dir, args.manifest)
 elif args.only == "aws":
-    aws.run(output_dir)
+    aws.run(output_dir, args.manifest)
 elif args.only == "google":
-    google.run(output_dir)
+    google.run(output_dir, args.manifest)
 else:
-    whisper.run(output_dir)
-    whisper.run_preprocessing(output_dir)
-    aws.run(output_dir)
-    google.run(output_dir)
+    whisper.run(output_dir, args.manifest)
+    print()
+    whisper.run_preprocessing(output_dir, args.manifest)
+    print()
+    aws.run(output_dir, args.manifest)
+    print()
+    google.run(output_dir, args.manifest)

transcribe/aws.py

@@ -18,9 +18,9 @@
 dotenv.load_dotenv()
 
 
-def run(output_dir):
+def run(output_dir, manifest):
     results = []
-    for file_metadata in tqdm.tqdm(utils.get_data_files(), desc="aws".ljust(10)):
+    for file_metadata in tqdm.tqdm(utils.get_data_files(manifest), desc="aws".ljust(10)):
         file_metadata["run_count"] = len(results) + 1
         file = file_metadata["media_filename"]
 

transcribe/google.py

@@ -13,9 +13,9 @@
 from . import utils
 
 
-def run(output_dir):
+def run(output_dir, manifest):
     results = []
-    for file_metadata in tqdm.tqdm(utils.get_data_files(), desc="google".ljust(10)):
+    for file_metadata in tqdm.tqdm(utils.get_data_files(manifest), desc="google".ljust(10)):
         file_metadata["run_count"] = len(results) + 1
         file = file_metadata["media_filename"]
 

transcribe/utils.py

@@ -28,10 +28,9 @@
 ]
 
 
-def get_data_files():
+def get_data_files(manifest):
     rows = []
-    data_csv = Path(__file__).parent.parent / "data.csv"
-    for row in csv.DictReader(open(data_csv)):
+    for row in csv.DictReader(open(manifest)):
         rows.append(row)
     return rows
 

transcribe/whisper.py

@@ -42,9 +42,9 @@
 ]
 
 
-def run(output_dir):
+def run(output_dir, manifest):
     combinations = list(whisper_option_combinations())
-    files = utils.get_data_files()
+    files = utils.get_data_files(manifest)
     total = len(combinations) * len(files)
     progress = tqdm.tqdm(total=total, desc="whisper".ljust(10))
 
@@ -60,9 +60,9 @@ def run(output_dir):
     utils.write_report(results, csv_filename, extra_cols=["options"])
 
 
-def run_preprocessing(output_dir):
+def run_preprocessing(output_dir, manifest):
     results = []
-    files = utils.get_data_files()
+    files = utils.get_data_files(manifest)
     total = len(files) * len(preprocessing_combinations)
     progress = tqdm.tqdm(total=total, desc="preprocess".ljust(10))