suyuan

app.py

@@ -12,7 +12,7 @@ import re
 def create_interval_data_dict(xmin, xmax, sentence):
     return {'xmin': float(xmin), 'xmax': float(xmax), 'text': sentence}
 
-def write_textgrid_file(intervals, output_file_path, total_xmax, tier_name):
+def write_textgrid_file(intervals, output_file_path, total_xmax):
     with open(output_file_path, 'w') as f:
         f.write('File type = "ooTextFile"\n')
         f.write('Object class = "TextGrid"\n\n')
@@ -23,16 +23,22 @@ def write_textgrid_file(intervals, output_file_path, total_xmax, tier_name):
         f.write('item []:\n')
         f.write('    item [1]:\n')
         f.write('        class = "IntervalTier"\n')
-        f.write(f'        name = "{tier_name}"\n')
+        f.write('        name = "Intonational Unit"\n')
         f.write('        xmin = 0\n')
         f.write(f'        xmax = {str(float(total_xmax))}\n')
-        f.write(f'        intervals: size = {len(intervals)}\n')
+        f.write(f'        intervals: size = {len(intervals) + 1}\n')
 
         for idx, interval in enumerate(intervals):
             f.write(f'        intervals [{idx + 1}]:\n')
             f.write(f'            xmin = {interval["xmin"]}\n')
             f.write(f'            xmax = {interval["xmax"]}\n')
             f.write(f'            text = "{interval["text"]}"\n')
+
+        if len(intervals) > 0:
+            f.write(f'        intervals [{len(intervals) + 1}]:\n')
+            f.write(f'            xmin = {intervals[-1]["xmax"]}\n')
+            f.write(f'            xmax = {intervals[-1]["xmax"]}\n')
+            f.write(f'            text = ""\n')
             
 def validate_csv_format(header):
     expected_headers = ['', 'file_name', 'xmin', 'xmax', 'text', 'is_unit_start_pred']
@@ -63,7 +69,7 @@ def validate_row(row):
         return False, f"Value error: {e}"
         
 # ==== Gradio Interface Function ====
-def csv_to_textgrid(file, tier_name="generated_tier"):
+def csv_to_textgrid(file):
     try:
         # Create temporary directory
         temp_dir = tempfile.mkdtemp()
@@ -142,13 +148,9 @@ def csv_to_textgrid(file, tier_name="generated_tier"):
                         if words:
                             intervals.append(create_interval_data_dict(iu_xmin, iu_xmax, ' '.join(words)))
                         
-                        if intervals:
-                            last_xmax = intervals[-1]['xmax']
-                            intervals.append(create_interval_data_dict(last_xmax, last_xmax + 0.001, ''))  # New interval
-
                         if intervals:
                             textgrid_path = os.path.join(output_directory, f"{prev_filename}.TextGrid")
-                            write_textgrid_file(intervals, textgrid_path, intervals[-1]['xmax'], tier_name)
+                            write_textgrid_file(intervals, textgrid_path, intervals[-1]['xmax'])
                             processed_files.append(prev_filename)
                             print(f"Wrote file: {prev_filename}.TextGrid with {len(intervals)} intervals")
                         
@@ -177,20 +179,10 @@ def csv_to_textgrid(file, tier_name="generated_tier"):
                 if not current_file_processed and prev_filename:
                     if words:
                         intervals.append(create_interval_data_dict(iu_xmin, iu_xmax, ' '.join(words)))
-
-                    # Add the new interval with xmin as last xmax and xmax as last xmax + 0.001
-                    if intervals:
-                        last_xmax = intervals[-1]['xmax']
-                        new_xmin = last_xmax
-                        new_xmax = last_xmax + 0.001
-
-                        # Only add the new interval if it's not a duplicate
-                        if new_xmin < new_xmax:  # Ensure they are not the same
-                            intervals.append(create_interval_data_dict(new_xmin, new_xmax, ''))  # New interval
-
+                    
                     if intervals:
                         textgrid_path = os.path.join(output_directory, f"{prev_filename}.TextGrid")
-                        write_textgrid_file(intervals, textgrid_path, intervals[-1]['xmax'], tier_name)
+                        write_textgrid_file(intervals, textgrid_path, intervals[-1]['xmax'])
                         processed_files.append(prev_filename)
                         print(f"Wrote last file: {prev_filename}.TextGrid with {len(intervals)} intervals")
         
@@ -218,40 +210,29 @@ def csv_to_textgrid(file, tier_name="generated_tier"):
         return None, f"Error: {str(e)}"
 
 # ==== Gradio Interface Setup ====
-csv_format_instruction = """
-**Expected CSV Format:**
-
-Please ensure that the CSV file adheres to the following format:\n
-
-- The first row must contain headers: `, file_name, xmin, xmax, text, is_unit_start_pred`.
-- Each subsequent row should contain the following columns for every word or segment in the audio file:
-  - `file_name`: Identifier for the audio file, used to group intervals.
-  - `xmin`: Start time of the segment (in seconds).
-  - `xmax`: End time of the segment (in seconds).
-  - `text`: The actual spoken word or phrase.
-  - `is_unit_start_pred`: Marks the beginning of a new unit (TRUE/FALSE).
-  
-**Please note: We currently only accept CSVs with an index.**
-
-**Tier Name:**  
-Please enter the tier name according to your preference or as deemed appropriate for the data.
-
-**Example CSV:**
-
-|   | file_name  | xmin   | xmax   | text  | is_unit_start_pred |
-|---|------------|--------|--------|-------|--------------------|
-| 0 | example1   | 20.42  | 20.74  | mhmm  | TRUE               |
-| 1 | example1   | 20.74  | 20.81  | hello | TRUE               |
-| 2 | example1   | 20.81  | 20.92  | world | FALSE              |
+csv_format_instruction = """**Expected CSV format:**
+The first row is the header. Each subsequent row should contain:\n
+`file_name, xmin, xmax, text, is_unit_start_pred`
+
+Each row represents a word or segment in an audio file.
+- `file_name`: Identifier for the audio file (used to group intervals).
+- `xmin`: Start time of the segment (in seconds).
+- `xmax`: End time of the segment (in seconds).
+- `text`: The actual spoken word or phrase.
+- `is_unit_start_pred`: Marks the beginning of a new unit (TRUE/FALSE).
+
+Example:\n
+| file_name | xmin   | xmax   | text  | is_unit_start_pred |
+|-----------|--------|--------|-------|---------------------|
+| example1  | 20.42  | 20.74  | mhmm  | TRUE                |
+| example1  | 20.74  | 20.81  | hello | TRUE                |
+| example1  | 20.81  | 20.92  | world | FALSE               |
 
 """
 
 iface = gr.Interface(
     fn=csv_to_textgrid,
-    inputs=[
-        gr.File(label="📁 Upload CSV File", file_types=[".csv"]),
-        gr.Textbox(label="📝 Enter Tier Name", placeholder="Enter the name of the tier")  # New input for tier name
-    ],
+    inputs=gr.File(label="📁 Upload CSV File", file_types=[".csv"]),
     outputs=[
         gr.File(label="📦 Download TextGrid ZIP"),
         gr.Textbox(label="✅ Status")