Update appImage.py

appImage.py

@@ -43,25 +43,35 @@ async def caption_from_frontend(file: UploadFile = File(...)):
 @app.get("/")
 def home():
     return RedirectResponse(url="/")"""
-# app_image_logic.py
-from transformers import AutoProcessor, AutoModelForCausalLM, pipeline
+# appImage.py
+from transformers import pipeline
+import tempfile, os
 from PIL import Image
-import torch
+from gtts import gTTS
 
-try:
-    processor = AutoProcessor.from_pretrained("microsoft/git-large-coco")
-    model = AutoModelForCausalLM.from_pretrained("microsoft/git-large-coco")
-    USE_GIT = True
-except:
-    captioner = pipeline("image-to-text", model="nlpconnect/vit-gpt2-image-captioning")
-    USE_GIT = False
+captioner = pipeline("image-to-text", model="nlpconnect/vit-gpt2-image-captioning")
 
-def generate_caption(image_path):
-    if USE_GIT:
-        image = Image.open(image_path)
-        inputs = processor(images=image, return_tensors="pt")
-        outputs = model.generate(**inputs, max_length=50)
-        return processor.batch_decode(outputs, skip_special_tokens=True)[0]
-    else:
-        result = captioner(image_path)
-        return result[0]['generated_text']
+async def caption_image(file):
+    contents = await file.read()
+    with tempfile.NamedTemporaryFile(delete=False, suffix=".png") as tmp:
+        tmp.write(contents)
+        image_path = tmp.name
+
+    captions = captioner(image_path)
+    caption = captions[0]['generated_text'] if captions else "No caption generated."
+
+    audio_path = text_to_speech(caption)
+
+    result = {"caption": caption}
+    if audio_path:
+        result["audioUrl"] = f"/files/{os.path.basename(audio_path)}"
+    return result
+
+def text_to_speech(text: str):
+    try:
+        tts = gTTS(text)
+        temp_audio = tempfile.NamedTemporaryFile(delete=False, suffix=".mp3")
+        tts.save(temp_audio.name)
+        return temp_audio.name
+    except:
+        return ""