Update app.py

app.py

@@ -11,37 +11,58 @@ def load_data(level):
     data = pd.read_csv(csv_file_path)
     return data
 
-def generate_speech(level, x, y, audio_option, pos_filter):
+def generate_speech(t1, level, t2, x, y, t3, audio_option, t4, all_pos, noun, verb, adjective, adverb, preposition):
     data = load_data(level)
     
     x, y = int(x), int(y)
-    if pos_filter != "ALL":  # If a specific POS is selected, filter by it
-        filtered_df = data[(data['SID'] >= x) & (data['SID'] <= y) & (data['POS'].str.lower() == pos_filter.lower())]
-    else:  # Include all words if "ALL" is selected
+
+    if all_pos:  # Include all words if "ALL" is selected
         filtered_df = data[(data['SID'] >= x) & (data['SID'] <= y)]
+    else:
+        # Initialize an empty DataFrame to accumulate results
+        filtered_df = pd.DataFrame()
+
+        # Check each selected POS and append matches to the filtered_df
+        if noun:
+            filtered_df = filtered_df.append(data[(data['SID'] >= x) & (data['SID'] <= y) & (data['POS'].str.lower() == 'noun')])
+        if verb:
+            filtered_df = filtered_df.append(data[(data['SID'] >= x) & (data['SID'] <= y) & (data['POS'].str.lower() == 'verb')])
+        if adjective:
+            filtered_df = filtered_df.append(data[(data['SID'] >= x) & (data['SID'] <= y) & (data['POS'].str.lower() == 'adjective')])
+        if adverb:
+            filtered_df = filtered_df.append(data[(data['SID'] >= x) & (data['SID'] <= y) & (data['POS'].str.lower() == 'adverb')])
+        if preposition:
+            filtered_df = filtered_df.append(data[(data['SID'] >= x) & (data['SID'] <= y) & (data['POS'].str.lower() == 'preposition')])
+    
+    # Ensure the filtered_df is unique in case of overlapping conditions
+    filtered_df = filtered_df.drop_duplicates()
+
 
-    if len(filtered_df) == 0:  # Check if the filtered DataFrame is empty
-        sentence = f"There is no {pos_filter} in the selected range."
+    # Proceed with generating the speech
+    combined_audio = AudioSegment.silent(duration=1000)  # Initial silence
+    if filtered_df.empty:
+        sentence = "No matching words found."
         tts = gTTS(text=sentence, lang='en')
         mp3_fp = io.BytesIO()
         tts.write_to_fp(mp3_fp)
         mp3_fp.seek(0)
-        sentence_audio = AudioSegment.from_file(mp3_fp, format="mp3")
-        combined_audio = sentence_audio
     else:
-        combined_audio = AudioSegment.silent(duration=1000)
         for _, row in filtered_df.iterrows():
-            if audio_option == "Audio with number":
-                sentence = f"Number {row['SID']}. {row['WORD']}!  {row['WORD']} is a {row['POS']}."
-            else:  # "Audio without number"
-                sentence = f"{row['WORD']}!"
+            # Adjust the sentence based on the audio_option
+            if audio_option == "Audio without number":
+                sentence = f"{row['WORD']} is {row['POS']}"
+            elif audio_option == "Audio with number":
+                sentence = f"{row['SID']}. {row['WORD']}. {row['WORD']} is {row['POS']}"
+            else:  # "Word only"
+                sentence = f"{row['WORD']}"
             
             tts = gTTS(text=sentence, lang='en')
             mp3_fp = io.BytesIO()
             tts.write_to_fp(mp3_fp)
             mp3_fp.seek(0)
             sentence_audio = AudioSegment.from_file(mp3_fp, format="mp3")
-            combined_audio += sentence_audio + AudioSegment.silent(duration=1500)
+            # Add a pause after each word
+            combined_audio += sentence_audio + AudioSegment.silent(duration=1000)
 
     mp3_io = io.BytesIO()
     combined_audio.export(mp3_io, format='mp3')
@@ -49,14 +70,24 @@ def generate_speech(level, x, y, audio_option, pos_filter):
     return mp3_io.getvalue()
 
 
+# Adjust inputs for POS with checkboxes
 iface = gr.Interface(
     fn=generate_speech,
     inputs=[
+        gr.Markdown("#### [1] Select Level"),  # Adding a label for POS options
         gr.Dropdown(label="Select Level", choices=['A1', 'A2', 'B1', 'B2', 'C1', '5K']),
-        gr.Number(label="Start Number (x)"),
-        gr.Number(label="End Number (y)"),
-        gr.Radio(label="Audio Option", choices=["Audio with number", "Audio without number"]),
-        gr.Dropdown(label="Select Part of Speech", choices=["ALL", "Noun", "Verb", "Adjective", "Adverb", "Preposition"])  # Updated POS filter input
+        gr.Markdown("#### [2] Select Range"),
+        gr.Number(label= "Start Number (x)"),
+        gr.Number(label = "End Number (y)"),
+        gr.Markdown("#### [3] Audio options (text)"),
+        gr.Radio(label="Audio Option", choices=["Audio with number", "Audio without number", "Word only"]),
+        gr.Markdown("#### Select Part of Speech"),  # Adding a label for POS options
+        gr.Checkbox(label="Any", value=False),
+        gr.Checkbox(label="Noun", value=False),
+        gr.Checkbox(label="Verb", value=False),
+        gr.Checkbox(label="Adjective", value=False),
+        gr.Checkbox(label="Adverb", value=False),
+        gr.Checkbox(label="Preposition", value=False)
     ],
     outputs=gr.Audio(label="Generated Speech"),
     title="Oxford Learner Vocabulary by CEFR levels: Learn with Sound",