Update app.py

app.py

@@ -1,6 +1,6 @@
 # app.py
 # Whisper transcription app - HYBRID conversion (pydub + small ffmpeg fallback)
-# Paste chunks 1/4 -> 2/4 -> 3/4 -> 4/4 in order into /app/app.py
+# Cleaned, debugged, and Spaces-ready.
 
 import os
 import sys
@@ -32,6 +32,8 @@ except Exception as e:
 
 print("DEBUG: imports OK", flush=True)
 
+
+
 # ---------- Config ----------
 MEMORY_FILE = "memory.json"
 MEMORY_LOCK = threading.Lock()
@@ -70,6 +72,7 @@ memory = load_memory()
 print("DEBUG: memory loaded (words=%d phrases=%d)" % (len(memory.get("words", {})), len(memory.get("phrases", {}))), flush=True)
 
 
+
 # ---------- Postprocessing ----------
 MEDICAL_ABBREVIATIONS = {
     "pt": "patient",
@@ -106,6 +109,11 @@ def expand_abbreviations(text):
             out.append(t)
     return ''.join(out)
 
+
+
+
+
+
 def normalize_drugs(text):
     for k, v in DRUG_NORMALIZATION.items():
         text = re.sub(rf'\b{k}\b', v, text, flags=re.IGNORECASE)
@@ -146,6 +154,11 @@ def postprocess_transcript(text, format_soap=False):
         return soap
     return t
 
+
+
+
+
+
 # ---------- Memory utilities ----------
 def extract_words_and_phrases(text):
     # basic tokenization for words; phrases = sentences
@@ -179,6 +192,10 @@ def update_memory_with_transcript(transcript):
             except Exception:
                 pass
 
+
+
+
+
 def memory_correct_text(text, min_ratio=0.85):
     if not text or (not memory.get("words") and not memory.get("phrases")):
         return text
@@ -225,6 +242,7 @@ def save_as_word(text, filename=None):
 
 
 
+
 # ---------- Hybrid conversion: pydub + small ffmpeg fallback ----------
 def _ffmpeg_convert(input_path, out_path, fmt, sr, ch):
     cmd = [
@@ -277,6 +295,10 @@ def convert_to_wav_if_needed(input_path):
         except Exception:
             pass
 
+
+
+
+
     diag_dir = tempfile.mkdtemp(prefix="dct_diag_")
     diag_log = os.path.join(diag_dir, "conversion_diagnostics.txt")
     diagnostics = []
@@ -412,8 +434,6 @@ def transcribe_multiple(audio_files, model_name, advanced_options, merge_checkbo
             yield "\n\n".join(log), "\n\n".join(transcripts), None, 100
             return
 
-
-
     # collect audio file paths
     paths = []
     if extracted_audio_paths:
@@ -458,62 +478,11 @@ def transcribe_multiple(audio_files, model_name, advanced_options, merge_checkbo
             yield "\n\n".join(log), "\n\n".join(transcripts), None, int(5 + idx * 80 / max(1, total))
             continue
 
-        try:
-            whisper_opts = {}
-            if isinstance(advanced_options, dict):
-                whisper_opts.update(advanced_options)
-
-            result = model.transcribe(wav, **whisper_opts)
-            text = result.get("text", "").strip()
-            log.append(f"Transcribed: {len(text)} chars")
-
-            if enable_memory:
-                text = memory_correct_text(text)
-            text = postprocess_transcript(text)
-            transcripts.append(f"FILE: {os.path.basename(p)}\n{text}\n")
-
-            if enable_memory:
-                try:
-                    update_memory_with_transcript(text)
-                    log.append("Memory updated.")
-                except Exception:
-                    pass
-
-            yield "\n\n".join(log), "\n\n".join(transcripts), None, int(10 + idx * 85 / max(1, total))
-        except Exception as e:
-            log.append(f"Transcription failed for {p}: {e}")
-            transcripts.append(f"FILE: {os.path.basename(p)}\nERROR: Transcription failed: {e}")
-            yield "\n\n".join(log), "\n\n".join(transcripts), None, int(10 + idx * 85 / max(1, total))
-            continue
-        finally:
-            try:
-                if wav and os.path.exists(wav):
-                    tmpdir = tempfile.gettempdir()
-                    try:
-                        if os.path.commonpath([tmpdir, os.path.abspath(wav)]) == tmpdir and (not p.lower().endswith(".wav")):
-                            os.unlink(wav)
-                    except Exception:
-                        pass
-            except Exception:
-                pass
+    # continue... (UI + launch in next message)
+
 
-    if merge_checkbox:
-        try:
-            merged_text = "\n\n".join(transcripts)
-            word_file_path = save_as_word(merged_text)
-            log.append(f"Merged transcript saved: {word_file_path}")
-        except Exception as e:
-            log.append(f"Failed to save merged file: {e}")
-            word_file_path = None
 
-    yield "\n\n".join(log), "\n\n".join(transcripts), word_file_path, 100
 
-    try:
-        if os.path.exists(temp_extract_dir):
-            shutil.rmtree(temp_extract_dir)
-            log.append("Cleaned temporary extraction dir.")
-    except Exception:
-        pass
 
 # ----------------------- Gradio UI -----------------------
 def run_transcription_wrapper(files, model_name, merge, zip_file, zip_password, enable_memory, advanced_options_state):