From 7b818c0d39e09294301991a4f11f60dd3f1f560c Mon Sep 17 00:00:00 2001
From: Anthony Wu <462072+anthonywu@users.noreply.github.com>
Date: Thu, 3 Oct 2024 15:45:48 -1000
Subject: [PATCH] fix bugs, add test coverage

---
 whisper/mlx_whisper/cli.py |  1 -
 whisper/test_cli.sh        | 39 +++++++++++++++++++++++++++++++++++---
 2 files changed, 36 insertions(+), 4 deletions(-)

diff --git a/whisper/mlx_whisper/cli.py b/whisper/mlx_whisper/cli.py
index f5304e5d..956d83a0 100644
--- a/whisper/mlx_whisper/cli.py
+++ b/whisper/mlx_whisper/cli.py
@@ -232,7 +232,6 @@ def main():
         if audio_obj == "-":
             # receive the contents from stdin rather than read a file
             audio_obj = audio.load_audio(from_stdin=True)
-            output_name_template = "content"
 
         try:
             result = transcribe(
diff --git a/whisper/test_cli.sh b/whisper/test_cli.sh
index f3b9b7c9..4e34a217 100755
--- a/whisper/test_cli.sh
+++ b/whisper/test_cli.sh
@@ -7,16 +7,44 @@ TEST_OUTPUT_DIR=$(mktemp -d -t mlx_whisper_cli_test)
 
 # the control output - cli called with audio position arg
 # expected output file name is ls_test.json
+TEST_OUTPUT_NAME_FOR_ALL="--output-name arg is used for all output formats"
 mlx_whisper "$TEST_AUDIO" \
     --output-dir "$TEST_OUTPUT_DIR" \
     --output-format all \
     --output-name '{basename}_transcribed' \
     --temperature 0 \
     --verbose=False
-/bin/ls ${TEST_OUTPUT_DIR}/ls_test_transcribed.{json,srt,tsv,txt,vtt} | sort
+if /bin/ls ${TEST_OUTPUT_DIR}/ls_test_transcribed.{json,srt,tsv,txt,vtt} > /dev/null; then
+    echo "[PASS] $TEST_OUTPUT_NAME_FOR_ALL"
+else
+    echo "[FAIL] $TEST_OUTPUT_NAME_FOR_ALL"
+fi
+
+
+TEST_OUTPUT_NAME_TEMPLATE="testing the output name template usage scenario"
+for test_val in $(seq 10 10 60); do
+    mlx_whisper "$TEST_AUDIO" \
+        --output-name "{basename}_mwpl_${test_val}" \
+        --output-dir "$TEST_OUTPUT_DIR" \
+        --output-format srt \
+        --max-words-per-line $test_val \
+        --word-timestamps True \
+        --verbose=False
+    TEST_DESC="testing output name template while varying --max-words-per-line=${test_val}"
+    if /bin/ls $TEST_OUTPUT_DIR/ls_test_mwpl_${test_val}.srt > /dev/null; then
+        echo "[PASS] $TEST_DESC"
+    else
+        echo "[FAIL] $TEST_DESC"
+    fi
+done
+
 
 TEST_STDIN_1="mlx_whisper produces identical output whether provided audio arg or stdin of same content"
-/bin/cat "$TEST_AUDIO" | mlx_whisper - --output-dir "$TEST_OUTPUT_DIR" --output-format json --temperature 0 --verbose=False
+/bin/cat "$TEST_AUDIO" | mlx_whisper - \
+    --output-dir "$TEST_OUTPUT_DIR" \
+    --output-format json \
+    --temperature 0 \
+    --verbose=False
 if diff "${TEST_OUTPUT_DIR}/content.json" "${TEST_OUTPUT_DIR}/ls_test_transcribed.json"; then
     echo "[PASS] $TEST_STDIN_1"
 else
@@ -25,7 +53,12 @@ else
 fi
 
 TEST_STDIN_2="mlx_whisper produces identical output when stdin comes via: cmd < file"、
-mlx_whisper - --output-name '{basename}_transcribed' --output-dir "$TEST_OUTPUT_DIR" --output-format tsv --temperature 0 --verbose=False < "$TEST_AUDIO"
+mlx_whisper - \
+    --output-name '{basename}_transcribed' \
+    --output-dir "$TEST_OUTPUT_DIR" \
+    --output-format tsv \
+    --temperature 0 \
+    --verbose=False < "$TEST_AUDIO"
 if diff "${TEST_OUTPUT_DIR}/content_transcribed.tsv" "${TEST_OUTPUT_DIR}/ls_test_transcribed.tsv"; then
     echo "[PASS] $TEST_STDIN_2"
 else